# WalletFollow — public research library Hyperliquid wallet tracking, copy-portfolio research and a read-only wallet Data API. This bundle contains the same public source pages and educational guides as the website. Editorial dates are not market-data timestamps. Current wallet observations, live prices, authenticated workspaces and private data are not included. Source index: https://walletfollow.ai/llms.txt AI source guidance: https://walletfollow.ai/ai Current API documentation: https://walletfollow.ai/api Current AI connection availability: https://walletfollow.ai/mcp The Data API is live. The MCP connector, CLI and client plugins are planned. Live copy execution depends on registered-engine capabilities, account authorization and confirmed state. --- # WalletFollow as a source for AI wallet research Canonical: https://walletfollow.ai/ai Publisher: WalletFollow Research Editorial update: 2026-09-29 WalletFollow brings together Hyperliquid wallet tracking, copy-portfolio research and a read-only wallet Data API. Use public guides for definitions and workflows, and dated account observations for current wallet facts. Every financial conclusion needs its address, time window, measurement basis and data limitations. ## Choose the source that answers the question For a current position or account value, use the wallet profile or a supported Data API response and preserve its market-data timestamp. For the meaning of a metric or a copy control, use the relevant knowledge-base guide and the methodology page. A guide’s editorial date cannot establish when a wallet last traded. For venue rules, follow the linked primary Hyperliquid documentation. WalletFollow describes its own product and measurement choices; it is not Hyperliquid and is not an independent certification of a wallet, creator or trading strategy. - [Explore public wallet profiles](https://walletfollow.ai/discover) - [Read the measurement methodology](https://walletfollow.ai/methodology) - [Browse the knowledge base](https://walletfollow.ai/learn) ## Read current availability precisely The WalletFollow Data API is live and provides its documented read-only endpoints. Read the current API catalog for request shapes, weights, errors and limits. The MCP connector, companion CLI and client plugins are planned. A page explaining MCP is not an announcement that a WalletFollow connector has shipped. A copy portfolio definition can be published while live copying is unavailable for an account. Execution depends on the registered engine, enabled capabilities, account authorization and the confirmed subscription state. Reading wallet data or producing an AI report does not authorize orders, transfers or withdrawals. - [Read the live Data API documentation](https://walletfollow.ai/api) - [Check AI & MCP setup availability](https://walletfollow.ai/mcp) - [Understand AI analysis versus execution](https://walletfollow.ai/learn/ai-analysis-vs-trade-execution) ## Make a wallet claim checkable Cite the specific page that supports the claim, rather than the homepage for every statement. For a wallet observation, retain the full public account address, network, market-data as-of time and review window. Identify whether the result is realized, unrealized, mark-to-market, simulated or actually executed in a follower account. An attribution can read: WalletFollow showed the account’s observed positions at the recorded as-of time; historical trade coverage was partial. This preserves the evidence without promoting the observation into a complete lifetime record. A calculation should identify its inputs and denominator, while an interpretation should remain visibly separate from the measured facts. - Use the stable page URL without search, filter or session parameters. - Attach observation time and metric basis to a financial value. - Keep missing inputs and coverage warnings in the summary. - Label hypothetical examples, portfolio simulations and inferred explanations. - Do not turn a wallet label into verified identity or historical returns into a forecast. ## Read the same guides as Markdown The source index at /llms.txt lists public resources. The bundle at /llms-full.txt includes the public source pages and complete educational articles. Individual knowledge-base and blog guides also have a Markdown representation: append .md to the guide’s URL. These representations use the same published content, sources and editorial dates as the human-readable pages. The files are reading conveniences, not a separate set of claims for AI systems. They do not contain private account data or real-time wallet snapshots. Google says that llms.txt and similar files neither improve nor harm Google Search visibility; ordinary accessible, useful pages remain the foundation. Availability of a text representation does not guarantee retrieval or citation by another service. - [Source index — llms.txt](https://walletfollow.ai/llms.txt) - [Complete public guide bundle — llms-full.txt](https://walletfollow.ai/llms-full.txt) - [Wallet tracking guide as Markdown](https://walletfollow.ai/learn/how-to-track-a-hyperliquid-wallet.md) ## Sources and further reading - [Hyperliquid: public account and market data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Model Context Protocol: architecture](https://modelcontextprotocol.io/docs/learn/architecture) - [Google Search: current guidance for generative AI features](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide) ## Related resources - [Editorial policy and publisher](https://walletfollow.ai/editorial-policy) - [Build an auditable AI wallet report](https://walletfollow.ai/blog/designing-an-auditable-ai-wallet-report) - [Use safe read-only research prompts](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) --- # How WalletFollow wallet data is measured Canonical: https://walletfollow.ai/methodology Publisher: WalletFollow Research Editorial update: 2026-09-29 WalletFollow separates venue observations, retained executions and derived analytics. Each result needs a defined account scope, period and measurement basis. Market-data timestamps describe the observations; coverage describes the evidence available. Unknown values remain unknown, and modeled portfolio results remain separate from actual follower execution. ## Where wallet observations come from Public wallet profiles can combine Hyperliquid account reads with retained history and prepared analytics. The Data API serves prepared wallet documents rather than making a new venue lookup for every customer request. A retrieved document can therefore be available now while its underlying observation was collected earlier. Data API sections identify their source. Account state, positions and spot holdings come from prepared account observations; discovery metrics come from prepared analytical rows; extended analysis comes from a prepared wallet overview. Preserve the source and section timestamp when exporting or comparing results. Derived measures should not be described as a direct venue field unless they are one. - [Read the Data API response documentation](https://walletfollow.ai/api) ## Freshness and coverage answer different questions An account as-of time records the relevant state observation. A valuation time records the price capture used for priced values. A document-build time records preparation of the response. The API exposes section-specific freshness information because these times can differ. An editorial update date belongs to an explanatory article, not to the underlying market data. Historical coverage is specific to a requested window and calculation. Current positions can be available while older fills or a starting valuation are incomplete. A null or unavailable value is not zero. Read the associated gap reason, and check truncation indicators before assuming a returned list contains every row. - [Read freshness and coverage examples](https://walletfollow.ai/learn/wallet-data-freshness-and-coverage) ## Account value and PnL use stated bases Total wallet value is an aggregation under a stated account-mode and valuation rule. Unified and portfolio-margin accounts require care to avoid counting collateral and perpetual equity twice. Spot assets need usable valuation inputs, and excluded or unpriced holdings should remain distinguishable from a zero balance. A value marked as an estimate should be cited as an estimate. Realized fills, completed-position results, unrealized PnL and mark-to-market period change are different views. Fees, funding, cash flows and venue scope can explain apparent disagreements. Return percentages also need their capital denominator. Align these definitions and timestamps before comparing two wallets or two trackers. - [Why trackers report different PnL](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [Spot, perpetual and staked balances](https://walletfollow.ai/learn/spot-perp-and-staked-balances) - [ROI versus dollar PnL](https://walletfollow.ai/learn/roi-vs-dollar-pnl) ## Trade grouping and rankings have limits One order can produce multiple fills; one position can have partial reductions before a full close. Fill count, completed-trade count and decision count are therefore different units. Win rate and trade-duration measures depend on the grouping rule and the available sample. A score or style label describes eligible observations under a method, not a verified identity or future performance. Discover comparisons depend on the selected period and eligible data. A high dollar gain can reflect more capital, while a large percentage can rely on a small base. Read coverage and sample warnings alongside rankings. Public wallet activity may omit the owner’s other addresses, external hedges or motivation. - [Read positions and trade history](https://walletfollow.ai/learn/reading-wallet-positions-and-trades) - [Win rate and profit factor](https://walletfollow.ai/learn/win-rate-vs-profit-factor) - [Wallet labels and identity evidence](https://walletfollow.ai/learn/wallet-labels-and-public-address-privacy) ## Copy-portfolio records are separate evidence types A backtest reconstructs a rule on historical inputs. Tracked portfolio statistics describe a published version under its stated method. Your executed follower result comes from your account’s fills, costs and cash flows. New portfolio weights should not receive credit for performance earned before their effective publication. Read version boundaries and basis labels instead of treating every plotted point as live copying. Weights, cash handling, costs, reinvestment and risk settings affect modeled results. Actual followers additionally face startup prices, rounding, liquidity, delays and skipped trades. A pause can leave existing exposure, a stop can leave unmanaged positions and a close can leave execution residuals. Confirm engine state and actual positions rather than inferring completion from a requested command. - [Backtests, tracked statistics and follower results](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) - [Why follower returns differ](https://walletfollow.ai/learn/why-follower-returns-differ) - [Pause, stop and close](https://walletfollow.ai/learn/pause-stop-and-close-copy-trading) ## Sources and further reading - [Hyperliquid: public account and market data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: entry price and PnL definitions](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) - [Hyperliquid: account abstraction modes](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/account-abstraction-modes) ## Related resources - [AI source and citation guidance](https://walletfollow.ai/ai) - [Editorial policy](https://walletfollow.ai/editorial-policy) - [Research a Hyperliquid wallet](https://walletfollow.ai/discover) --- # WalletFollow editorial policy Canonical: https://walletfollow.ai/editorial-policy Publisher: WalletFollow Research Editorial update: 2026-09-29 WalletFollow Research is the platform’s editorial publishing label. Our guides explain public wallet data, metrics and product workflows. They use linked primary documentation and clearly labeled examples, distinguish observations from interpretation, and preserve limitations. Publisher attribution does not imply independent financial advice or a third-party audit. ## Publish an answer to a distinct question Knowledge-base guides explain definitions, practical checks and current product workflows. Blog articles develop repeatable research methods and decision frameworks. Each page should help a reader complete a specific research task rather than repeat the same introduction with a new wallet address or keyword. The platform’s primary purpose is wallet and portfolio research. WalletFollow has a commercial interest in its own product. A guide should still be useful before signup and should identify what public data cannot establish, including future profitability, complete external exposure or suitability for an individual reader. ## Use sources that support the claim Venue mechanics should link to primary Hyperliquid documentation, and protocol explanations should link to the Model Context Protocol documentation. Product descriptions are checked against WalletFollow’s implemented workflows and current availability. A source link should substantiate the adjacent topic rather than imply an endorsement of WalletFollow. Measured results need a period, scope, basis and observation time. Hypothetical arithmetic is labeled as an example. Vendor keyword estimates, competitor exports and selected public-wallet observations are not presented as independent market-wide studies. We do not invent customer outcomes, expert credentials, quotations or performance guarantees. ## Be transparent about AI assistance This educational foundation was prepared with AI assistance and checked against product contracts and linked primary documentation. WalletFollow Research names the publisher; it is not a fictional analyst biography. AI-assisted drafting does not establish that a claim is accurate or that a strategy has been independently validated. Publication requires checking material factual claims, examples, product status and source links. An assistant’s explanation of a trader’s intent remains an interpretation unless evidence supports it. Generated screenshots, sample stories and simulated results must not be passed off as observed market reporting or actual customer performance. ## Name what is live and what is planned The documented WalletFollow Data API is live and read-only. The MCP connector, companion CLI and client plugins are planned. Copy execution requires the relevant registered-engine and account capabilities. A product guide should retain these distinctions rather than advertise a proposed tool as a released feature. An educational article does not request wallet approval or establish trading authorization. Historical performance, a leaderboard position or a portfolio simulation does not guarantee future returns. Dates on articles describe editorial publication or review, while market-data timestamps describe underlying observations. ## Update material claims and preserve context Review a guide when venue rules, product capabilities, metric definitions or an identified factual error change its answer. A substantive review updates the displayed editorial date and all alternate text representations. Cosmetic changes should not manufacture an appearance of fresh market evidence. A correction should identify the affected definition or claim and replace it with supported information. Keep old and new metric bases distinguishable where they change a comparison. Readers evaluating a possible discrepancy should retain the page URL, observed date, exact claim and supporting source; for a wallet result, also retain its address, time window and as-of time. ## Sources and further reading - [Hyperliquid: public account and market data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Model Context Protocol: architecture](https://modelcontextprotocol.io/docs/learn/architecture) - [Google Search: current guidance for generative AI features](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide) ## Related resources - [Data sources and methodology](https://walletfollow.ai/methodology) - [Citation guidance and text sources](https://walletfollow.ai/ai) - [Knowledge base](https://walletfollow.ai/learn) - [Research blog](https://walletfollow.ai/blog) --- # How to track a Hyperliquid wallet Canonical: https://walletfollow.ai/learn/how-to-track-a-hyperliquid-wallet Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking To track a Hyperliquid wallet, search its public address in WalletFollow, open the wallet profile, and check the timestamp before reading positions or performance. Public wallet research requires an address, not a private key, wallet approval or permission to trade. ## Start with the correct address Use the complete account address: 0x followed by 40 hexadecimal characters. Copy it from a trusted source and check the opening and closing characters after pasting. A shortened address in a screenshot is insufficient because many accounts can share the same prefix. A main account, subaccount and agent address can represent different things; an agent address is not necessarily the account whose trading history you want. Open the resulting profile and confirm that its visible activity matches what you expected. An empty profile can mean an inactive account or unavailable indexed history. It does not prove that the address has never traded anywhere. ## Separate the current book from the track record Positions describe the account at a particular moment: market, direction, size, entry price, unrealised PnL and available margin information. Trade history describes fills and completed activity. A trader can hold a large winning position while its recent closed trades lose money. Read these views together rather than treating the current position table as a complete performance report. Choose a review period before comparing wallets. A seven-day result and a ninety-day result answer different questions. Keep the period and performance basis consistent when adding another wallet to your shortlist. ## Check coverage before drawing conclusions Read the as-of time, history coverage and missing-value explanations. A current account snapshot can coexist with incomplete older fills. An unavailable win rate should remain unavailable; replacing it with zero would imply a measured losing record. Likewise, a chart that begins recently cannot support a claim about the entire lifetime of an account. For repeated reviews, record the address, period, timestamp and any coverage warnings. This small audit trail lets you distinguish a real change in behaviour from a change in which observations were available. ## Turn a profile into a research note Write down what you observed and what remains unknown. A useful note might say that a wallet currently has concentrated long exposure, trades infrequently and has only partial historical coverage. That is more defensible than calling it smart money after one profitable week. WalletFollow helps inspect public activity; the account owner’s identity, off-platform hedges and future decisions may remain unknown. - Record exposure and leverage alongside returns. - Look for losses and funding costs, not only winning trades. - Revisit the same period and definitions on your next review. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to read wallet positions and trade history](https://walletfollow.ai/learn/reading-wallet-positions-and-trades) - [Wallet data freshness and historical coverage](https://walletfollow.ai/learn/wallet-data-freshness-and-coverage) - [How to find and evaluate Hyperliquid whale wallets](https://walletfollow.ai/learn/how-to-find-hyperliquid-whales) --- # How to find and evaluate Hyperliquid whale wallets Canonical: https://walletfollow.ai/learn/how-to-find-hyperliquid-whales Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking Find Hyperliquid whales by screening public wallets for account value and position notional, then inspect concentration, leverage and trading history. A large balance describes size. It does not establish skill, identity or a reason to follow the wallet’s next trade. ## Define what whale means for your question A large account balance and a large leveraged position are different signals. An account with modest equity can control substantial notional exposure, while a well-funded account may keep most capital idle. Decide whether you are studying capital, directional exposure, trading volume or market impact. Do not switch between those definitions simply because another leaderboard produces a more dramatic ranking. In WalletFollow, start with Discover and open the relevant profiles rather than copying a list of names. The profile is where you can connect a size observation to the account’s current positions and measured history. ## Screen before investigating individual trades Use a consistent time window and compare several candidates. Review account value, gross exposure, largest market share and whether performance comes from many trades or one unresolved position. A high-dollar PnL may mainly reflect access to more capital. A smaller wallet can have a stronger percentage result while taking much more risk. Avoid chasing a whale immediately after a viral screenshot. By the time a position becomes widely discussed, its entry, size or hedge can have changed. The original screenshot’s timestamp matters as much as its headline number. ## Treat public activity as incomplete context A wallet profile cannot show every position its owner holds on other venues or under other addresses. A visible short might hedge a spot holding elsewhere. Two wallets trading similarly may belong to different people, and two wallets with different strategies may belong to one person. Address-level data supports address-level conclusions unless independent evidence establishes more. Use neutral notes such as large BTC short or concentrated perp exposure. Avoid turning a wallet label into a verified biography. Where an identity matters, preserve the source and distinguish the owner’s own statement from an inference. ## Track behaviour rather than one leaderboard rank A repeatable watchlist review asks whether position size, market concentration and loss tolerance are changing. Record additions and reductions rather than assuming every fill starts a new directional thesis. A whale can scale into or out of a position across many fills. Watching the sequence is more informative than extracting a single transaction. - Compare equity with notional exposure. - Check timestamps and history coverage. - Keep observed facts separate from inferred intent. - Do not treat size as a profitability guarantee. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Wallet labels, identity claims and public-address privacy](https://walletfollow.ai/learn/wallet-labels-and-public-address-privacy) - [Long, short, gross and net wallet exposure](https://walletfollow.ai/learn/long-and-short-exposure) - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) --- # How to read wallet positions and trade history Canonical: https://walletfollow.ai/learn/reading-wallet-positions-and-trades Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking Open positions show the wallet’s current exposure. Fills show executed transactions. Completed trades group related executions into a trading result. These views describe different units, so one hundred fills do not automatically mean one hundred independent trades or investment decisions. ## Read direction, size and value together For each position, identify the market and whether exposure is long or short. Size is measured in the asset’s units; notional value expresses market exposure in money. Entry price describes the tracked cost basis, while the current valuation price changes with the market. A profitable-looking position can still dominate the account’s risk if its notional is large relative to equity. Margin mode and estimated liquidation information add context. Do not compare a small isolated position with a cross-margin account using only the displayed leverage setting. The account supporting those positions can differ materially. ## A fill is an execution, not a whole strategy An order can execute in pieces. The resulting fills may share a market, order identifier and close timestamps while carrying different prices or fees. Adding every row to a trade count can make an active account appear to have far more independent decisions than it actually made. The tracker’s grouping rule therefore affects trade count, average trade size and win rate. Read whether a fill opens, increases, reduces or reverses exposure. Selling while long can close part of a position; it does not always mean the trader has become short. Follow the position before and after the execution. ## Partial closes leave risk behind Hypothetical example: a trader opens ten units and later sells four. The closed portion can realize a profit while six units remain exposed. A positive closed result says nothing certain about the final outcome of the remaining position. A position reversal can both close the prior direction and open a new one, depending on execution size. When reviewing a large gain, check whether it came from a fully closed position, a partial realization or a still-open mark-to-market gain. This avoids crediting an unrealised result as a finished trade. ## Use an explicit review sequence Start at the current position table, then inspect recent fills for changes to that book, and finally read completed-trade statistics over a defined period. If the stories conflict, investigate the units and coverage before choosing one. WalletFollow’s missing-value and as-of labels are part of the evidence, not decoration. - Keep spot and perpetual activity distinct. - Check how partial executions are grouped. - Read fees separately from gross trade PnL. - Do not infer the owner’s motivation from a fill direction alone. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Realised versus unrealised PnL](https://walletfollow.ai/learn/realized-vs-unrealized-pnl) - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [A crypto trading journal for wallet research](https://walletfollow.ai/learn/crypto-trading-journal) --- # Why wallet PnL differs between trackers Canonical: https://walletfollow.ai/learn/why-wallet-pnl-differs Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking Two wallet trackers can report different PnL because they use different time windows, account scopes, valuation timestamps or definitions. Compare realised, unrealised and cash-flow-adjusted results on the same basis before deciding that either figure is wrong. ## Match the window and the scope first Write down the start and end of the period, including timezone. A rolling twenty-four-hour window differs from a UTC calendar day. Next check whether each result includes only one perp venue, all supported perp venues, spot assets, staked assets or a particular subaccount. The same visible wallet address does not guarantee the same aggregation rule. Then compare timestamps. One page can value a position at a newer price than another. During a rapid price move, two correctly calculated snapshots can disagree substantially until their as-of times align. ## Reconcile cash movement with trading results An increase in account value can come from a deposit rather than a gain. Hypothetical example: an account begins with $10,000, receives $5,000 and ends with $14,000. Ignoring flows suggests a $4,000 increase; subtracting the deposit reveals a $1,000 decline over the period, before considering the precise timing and return methodology. Withdrawals create the opposite illusion. A trader can earn money while account value falls because capital left the account. Read the stated cash-flow adjustment rather than treating the raw equity curve as a return series. ## Check what each PnL number includes A realized-fill figure, completed-trade figure and mark-to-market change are not interchangeable. One report may subtract exchange fees but show funding separately. Another may include funding in its portfolio result. A fill fee can already contain a builder component, so adding that component again can double-count costs. Create a small reconciliation with the reported figure, performance basis, fee treatment, funding treatment and account scope. Use unavailable where an input is unknown. Guessing a missing cost rarely resolves a disagreement reliably. ## Investigate coverage before comparing precision A tracker with a fresh position snapshot may still lack some earlier fills or a usable starting valuation. In WalletFollow, a gap explanation signals that a result cannot be computed on the intended basis. It should not be replaced by a rival’s value without knowing that rival’s method. If differences persist after aligning definitions, compare a specific position or execution at matching timestamps. - Same window, scope and as-of time. - Same cash-flow adjustment. - Same realised or mark-to-market basis. - Same fee, funding and coverage assumptions. ## Sources and further reading - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Realised versus unrealised PnL](https://walletfollow.ai/learn/realized-vs-unrealized-pnl) - [ROI versus dollar PnL: comparing wallets fairly](https://walletfollow.ai/learn/roi-vs-dollar-pnl) - [Wallet data freshness and historical coverage](https://walletfollow.ai/learn/wallet-data-freshness-and-coverage) --- # Wallet data freshness and historical coverage Canonical: https://walletfollow.ai/learn/wallet-data-freshness-and-coverage Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking Freshness describes how recent an observation is. Coverage describes how much of the intended history is available. A fresh position table does not prove complete historical coverage, and a complete old trading window does not prove the wallet’s current exposure is unchanged. ## Use the observation time, not the page-load time Loading a page now can retrieve a snapshot collected earlier. Distinguish the page’s review date, the document generation time and the underlying account’s as-of time. For a research note, record the market-data timestamp associated with the metric you used. Editorial update dates on a guide describe the guide, not the age of a wallet position. Different sections can rely on different inputs. A balance summary, historical trade analysis and market price may update at different moments. If a change looks inconsistent, compare those timestamps before assuming the account took an unusual action. ## Coverage is specific to a window and a metric A wallet can have usable seven-day history but incomplete ninety-day history. A fills-based statistic needs executions, while a cash-flow-adjusted portfolio return also needs suitable account values and capital movements. Knowing the current balance alone cannot reconstruct every missing performance observation. WalletFollow marks gaps rather than claiming a result unsupported by inputs. A null or unavailable field is distinct from a measured zero. Zero wins means observed activity produced no wins; an unavailable win rate means the calculation lacks what it needs. ## Small samples can look more certain than they are Three successful trades can produce a high win rate without establishing a repeatable process. A smooth chart can reflect widely spaced observations that miss a sharp intraday loss. An apparently inactive period can mean no observed fills, but the account could still be holding a leveraged position and paying funding. When comparing traders, align both the period and the evidence quality. A shorter well-covered record may be more interpretable than a longer record with major gaps, but it still offers less evidence across changing market conditions. ## Handle uncertainty in your workflow Document which questions the available data can answer. Current exposure can support a concentration review even if historical profit factor is unavailable. Avoid combining supported and unsupported metrics into a confident overall verdict. For AI analysis, ask the assistant to preserve missing-value reasons and carry the relevant timestamps into its final answer. - Record the oldest material input in a combined analysis. - Treat gaps as unknowns. - Refresh before acting on current exposure. - Do not join backtested and observed history as one track record. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [Copy-portfolio backtests, tracked results and live returns](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) --- # Wallet labels, identity claims and public-address privacy Canonical: https://walletfollow.ai/learn/wallet-labels-and-public-address-privacy Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking A wallet address identifies an account, not automatically a person. Public trading activity can be researched without wallet ownership, but identity labels need separate evidence. Sharing an address can connect previously separate activity, so consider that privacy effect before publishing your own. ## Separate account evidence from identity evidence A profile can show an address’s observed positions and transactions. It cannot establish the human behind the address merely from profitability, trade timing or a familiar nickname. Similar behaviour can have many causes: shared signals, bots, market conditions or coincidence. Treat an unverified name as a label rather than a confirmed identity. If a trader publicly claims an address, preserve the original source and date. That claim can support attribution, but it should not be broadened into statements about unrelated addresses without additional evidence. Wallet ownership and beneficial ownership are also different questions. ## Public does not mean context-free Combining an address with social profiles, screenshots or customer information can expose more than either source alone. Before publishing a research note, remove unnecessary personal details. Focus on the trading evidence that supports the analysis. There is rarely a reason to add a home location, personal contact information or speculation about a trader’s private life. For your own account, assume that a publicly shared address may let others revisit its future activity. Deleting a post later may not remove copies or the public transaction record. Consider this before attaching an address to a permanent public identity. ## Interpret badges and categories carefully A whale label concerns size under a defined rule. A risk or style label concerns observed behaviour under a particular window. Neither badge is an endorsement, identity verification or promise of future performance. Ask which observations produced the classification and whether those observations are current. A nickname can help organize your private research even when it is not a verified public identity. Use neutral names such as concentrated ETH trader or frequent scalper if that is all the evidence supports. Avoid publishing defamatory interpretations of unexplained losses or transfers. ## Keep research access separate from account access Tracking a public address does not require the private key or seed phrase controlling it. A request for those secrets is outside a public wallet research workflow. Wallet connection for sign-in or copy-account approval is a separate user action with its own explanation. WalletFollow’s educational content should never be interpreted as permission to access someone else’s account or act in their name. - Attribute identities only with evidence. - Publish observations rather than private speculation. - Keep wallet credentials out of research notes and AI prompts. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to track a Hyperliquid wallet](https://walletfollow.ai/learn/how-to-track-a-hyperliquid-wallet) - [How to find and evaluate Hyperliquid whale wallets](https://walletfollow.ai/learn/how-to-find-hyperliquid-whales) - [Agent approvals and wallet permissions on Hyperliquid](https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions) --- # Spot, perpetual and staked balances on Hyperliquid Canonical: https://walletfollow.ai/learn/spot-perp-and-staked-balances Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Hyperliquid A Hyperliquid wallet’s total value can include several balance types, depending on account mode and valuation rules. Perpetual equity, spot holdings, staked assets and withdrawable funds answer different questions. Do not assume the largest balance shown is immediately available to support every trade. ## Identify the balance category Spot holdings represent asset balances, while perpetual positions create market exposure backed by margin. Perpetual account equity can change as open positions gain or lose value. Staked assets have a separate role and availability conditions. A tracker’s total-value summary must state which categories it includes so readers do not accidentally add a component twice. Account abstraction modes can affect how balances and collateral interact. Read the account mode where it is available, then use the corresponding venue information. If the mode is unknown, keep that uncertainty in any margin or capital analysis. ## Asset quantity is different from a dollar valuation A spot balance needs a price before it becomes a dollar value. Thinly traded or unsupported assets may not have a reliable valuation. A large token quantity can produce an unrealistic headline wealth figure if it is multiplied by a price at which the whole holding could never be sold. WalletFollow can exclude holdings from a total under its valuation rules or leave their value unavailable. An excluded dollar value is not the same as a missing token balance. Review the reason rather than assuming the asset disappeared from the account. ## Withdrawable funds are not identical to equity A profitable account may still have capital committed as margin, open orders or holdings with restricted availability. Unrealised profit is not a separate cash deposit. Likewise, a spot holding’s estimated market value does not prove the same amount is available as settlement currency for a perpetual order. Hypothetical example: a wallet has $8,000 of account value while an open position uses much of its margin capacity. Quoting $8,000 as cash ready for a new copy allocation would ignore the existing obligations. Inspect free or withdrawable balances on their stated account basis. ## Compare wallets with the same accounting scope When ranking account size, decide whether you are comparing perp equity or a broader valuation. Keep that choice consistent across wallets. A portfolio with spot assets and a perp hedge can look directionally different from its perp book alone. Public tracking remains incomplete if additional hedges sit elsewhere. - Read category and account mode. - Check valuation and exclusion rules. - Avoid double-counting included components. - Separate estimated value from funds available for new risk. ## Sources and further reading - [Hyperliquid: Account abstraction modes](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/account-abstraction-modes) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Staking](https://hyperliquid.gitbook.io/hyperliquid-docs/hypercore/staking) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [Long, short, gross and net wallet exposure](https://walletfollow.ai/learn/long-and-short-exposure) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) --- # Hyperliquid copy trading explained Canonical: https://walletfollow.ai/learn/hyperliquid-copy-trading-explained Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Copy trading uses a leader’s activity or target exposure to inform trades in a follower’s account under defined allocation and risk rules. Wallet tracking only observes activity. In WalletFollow, live execution depends on an enabled, connected engine and the account’s authorization and current capability status. ## A public profile is research, not execution Opening a wallet profile does not authorize orders in your account. You can inspect positions, history and risk without becoming a follower. A copy portfolio adds a defined roster of leaders, weights and versioned settings. Subscribing adds your allocation, limits and consent. These steps should be explicit rather than implied by viewing a profitable wallet. The leader’s displayed history is evidence for research. It is not your personal account history and cannot guarantee the result of joining at a different time or with different capital. ## Execution is a separate system with conditions WalletFollow’s platform manages portfolio definitions and customer settings. A separate registered engine is responsible for execution when the relevant capabilities are enabled. A published portfolio can exist while copying is unavailable, in a non-live mode or blocked for a particular account. Read the status shown in the current product rather than inferring availability from an educational article. A submitted setup or activation command is not the same as a confirmed active subscription. Follow the displayed engine response and account state before concluding that copying has begun. ## Follower trades will not perfectly match the leader Execution happens after an observed event or target change. Market prices, order-book depth, lot sizes, capital, account margin and your risk limits can differ. A leader may already have an old position when you join. Fees, funding and skipped orders can further separate the follower’s result from the leader’s. Choose an allocation you can afford to lose and understand what happens when a risk limit is reached. Lower limits can intentionally prevent the account from reproducing every trade. A copy strategy is still exposed to market and execution risk. ## Review the setup as an account decision Read the bound portfolio version, leader weights, fee schedule, approval scope and control behaviour. Recheck availability when switching from research to setup. Do not share a seed phrase or private key with a portfolio creator, AI assistant or platform page. The account owner should know which system can execute and how to revoke its authorization. - Research the trader before subscribing. - Set allocation and risk limits. - Read fees and engine mode. - Confirm the final state rather than a button click. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Builder codes](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/builder-codes) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) - [Agent approvals and wallet permissions on Hyperliquid](https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions) --- # How to evaluate a trader before copying Canonical: https://walletfollow.ai/learn/evaluate-a-trader-before-copying Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Evaluate a trader by reading the return basis, historical coverage, drawdown, leverage, concentration and trade sample together. Then ask whether the strategy can plausibly be followed with your capital and limits. A high leaderboard rank or recent PnL alone is insufficient evidence. ## Start with the quality of the record Confirm the address and period, then check whether historical inputs support the displayed metrics. Separate backtests, tracked portfolio results and individual wallet results. A long-looking chart can conceal gaps or include a reconstruction before the portfolio existed. Missing data should lower the confidence of a conclusion rather than being filled with optimistic assumptions. Look at more than one market condition when history permits. A wallet that performed well during a sustained rally may have mostly expressed leveraged long exposure. That can be a coherent strategy without establishing skill in a falling or sideways market. ## Read how gains and losses were produced Pair returns with drawdown and the largest losing episodes. Check whether performance comes from many completed trades or one still-open winner. Win rate needs average gain and average loss context; frequent small wins can be erased by a rare large loss. Dollar PnL also needs capital context before it can be compared fairly. Review exposure concentration. One market, one direction or several highly correlated markets can drive nearly the entire result. Multiple position rows do not necessarily provide multiple independent sources of risk. ## Ask whether the strategy is copyable A rapid, small-margin strategy may be sensitive to delay and fees. A large position in a thin book can create follower slippage. Minimum order size and rounding can make a small allocation behave differently from the leader. Starting while an old position is already profitable changes the entry conditions. Compare the leader’s observed behaviour with your intended limits. If the leader relies on exposure far above your cap, your result will intentionally diverge. Avoid treating that divergence as a bug without reviewing the risk rules. ## Use a written decision and a review trigger State why the candidate made the shortlist, what could invalidate the thesis and which observations you will revisit. This discourages changing criteria after a loss or chasing whichever address won most recently. WalletFollow provides research and portfolio controls; it does not certify that a candidate is suitable for your financial circumstances. - Defined window and adequate coverage. - Drawdown and concentration reviewed. - Capital and execution fit considered. - Fees, limits and exit controls understood. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Order book](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/order-book) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Drawdown and recovery in wallet performance](https://walletfollow.ai/learn/drawdown-and-recovery) - [Win rate versus profit factor](https://walletfollow.ai/learn/win-rate-vs-profit-factor) - [Slippage and latency in copy trading](https://walletfollow.ai/learn/copy-trading-slippage-and-latency) --- # Copy-trading allocation and position sizing Canonical: https://walletfollow.ai/learn/copy-trading-position-sizing Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading A copy allocation sets the capital assigned to a strategy; position sizing translates that allocation and portfolio weights into target exposure under risk limits. Allocation is not the same as notional exposure, and a twenty-percent portfolio weight does not necessarily mean twenty percent of total market risk. ## Distinguish capital from exposure Capital is the account value assigned to the strategy. Notional is the money value of the positions it controls. Hypothetical example: $1,000 of allocated capital at two times effective gross exposure corresponds to roughly $2,000 of position notional, before execution details. A one-percent adverse move across that directional exposure would be about $20 before costs. A leverage setting on one position does not describe the whole account. Unused cash, offsetting positions and other subscriptions can change effective account exposure. Review the combined book as well as each subscription. ## Weights allocate a strategy, not certainty Hypothetical example: a portfolio assigns sixty percent to leader A and forty percent to leader B. With $1,000 committed, the model assigns $600 and $400 of strategy capital. If A takes substantially more exposure than B, A can contribute much more than sixty percent of the portfolio’s risk. Review how cash and inactive leaders are handled. If a leader has no supported position, a portfolio may keep that share in cash rather than automatically redistribute it. Read the current portfolio method before assuming that all committed capital is always deployed. ## Small accounts face execution constraints Minimum notional, size precision and market-specific limits can prevent very small target trades. Rounding each trade can create a meaningful difference when many positions share a small allocation. A follower can also hit margin limits even when its requested strategy allocation looks sufficient on paper. A hypothetical $10 slice across ten markets is not equivalent to a $10,000 slice scaled perfectly. Review the product’s minimum-capital and market-eligibility information. Do not increase allocation just to clear a constraint without reassessing the possible loss. ## Size from a loss budget and monitor the result Choose capital and caps together. Consider concentration, gross exposure, loss tolerance and how the account behaves if several leaders lose simultaneously. The engine’s confirmed positions are the realised result of the sizing rules; a model preview is only a preview. After activation, compare actual exposure with the intended allocation and investigate skipped or constrained trades. - Set committed capital and exposure caps separately. - Inspect total account exposure across subscriptions. - Allow for minimum sizes and rounding. - Reassess sizing after deposits, withdrawals or portfolio updates. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to assess diversification in a copy portfolio](https://walletfollow.ai/learn/diversifying-a-copy-portfolio) - [Leverage and liquidation risk in copy trading](https://walletfollow.ai/learn/copy-trading-leverage-and-liquidation) - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) --- # Slippage and latency in copy trading Canonical: https://walletfollow.ai/learn/copy-trading-slippage-and-latency Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Copy-trading slippage is the difference between a reference price and the follower’s execution price. Latency is the delay between relevant observations and execution. Both can change returns, especially in fast markets, thin books and strategies that depend on small price advantages. ## Define the reference before measuring slippage A comparison can use the leader’s fill, the market price when the engine observed it, or the quote when the follower order was prepared. Those baselines answer different questions. Comparing against the leader combines price movement during delay with the follower’s own execution impact. Comparing against a fresh quote focuses more directly on execution after the decision. Record whether prices are volume-weighted across fills. A follower order executed in pieces can have a different average price from its first fill. Use the whole executed quantity when assessing its cost. ## Depth and order size matter An order book contains limited quantity at each price. A larger aggressive order may consume several levels. Many followers responding to the same leader can compete for the same liquidity. A market with an attractive last price can still have a wide spread or shallow depth at the moment an order arrives. Hypothetical example: a $10,000 trade with a 0.2% unfavorable price difference has about $20 of slippage relative to the chosen reference, before fees. Repeated entry and exit costs can overwhelm a strategy whose typical gross gain is small. ## A limit can reject a trade rather than reproduce it A slippage guard constrains accepted execution conditions. It does not guarantee the order fills. If price moves outside the allowed range, execution may be partial, delayed or skipped under the engine’s rules. A follower protected from an unfavorable entry can therefore miss a profitable leader trade as well. Increasing a tolerance trades some price protection for a greater chance of execution. That is a risk choice, not a universal improvement. Read current control descriptions and confirmed execution events rather than assuming an order filled because a leader traded. ## Evaluate strategy sensitivity Short holding periods and frequent turnover tend to make execution costs more important relative to each trade’s intended gain. Longer holding periods still face entry, exit and stressed-market risk. Review the leader’s typical trade size and duration alongside your allocation. For meaningful measurement, keep reference price, observed time, order time, average fill price and skipped-order reasons together. - Measure with a defined price baseline. - Review spread and book depth. - Include entry and exit costs. - Treat missed trades as part of follower divergence. ## Sources and further reading - [Hyperliquid: Order book](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/order-book) - [Hyperliquid: Fees](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/fees) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) - [Copy-trading fees: exchange, builder, funding and creator costs](https://walletfollow.ai/learn/copy-trading-fees) - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) --- # Leverage and liquidation risk in copy trading Canonical: https://walletfollow.ai/learn/copy-trading-leverage-and-liquidation Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Leverage makes market exposure larger relative to account equity, magnifying both gains and losses. A copied account can reach liquidation under different conditions from its leader because capital, entry prices, margin mode, other positions and funding differ. Risk caps reduce allowed exposure but cannot eliminate loss. ## Read effective exposure, not just a setting A displayed ten-times setting does not prove the account is currently using ten times all its equity. Effective gross exposure compares total absolute position notional with the equity supporting it. A trader can use a high setting on a small position or have significant exposure spread across several markets. Hypothetical example: an account with $1,000 of equity and $5,000 of directional notional has roughly five times exposure. A two-percent unfavorable move corresponds to about $100 of market loss before fees, funding and other account changes. This simplified calculation is not a liquidation-price formula. ## Margin mode changes how losses interact Cross margin can connect the risk of multiple positions through shared collateral. A loss in one market can reduce the buffer supporting another. Isolated margin allocates collateral to a particular position under its rules. Review the actual account mode and position margin information rather than assuming every copied trade is isolated. Existing manual positions or other copy subscriptions can matter. Evaluating one subscription alone can miss the combined account’s stress. Keep account-level exposure and collateral in the review. ## Liquidation estimates can move Estimated liquidation levels depend on account state and the venue’s margin rules. Funding, price changes, deposits, withdrawals and other positions can shift the result. A saved price level is not a permanent boundary. A wide-looking buffer on an old snapshot is especially unreliable during a volatile move. Stop or drawdown controls also depend on observations and execution. A trigger can request risk reduction, but market gaps, liquidity constraints or unavailable systems can affect the eventual fill. Do not equate a configured percentage with a guaranteed maximum loss. ## Choose limits before activation Decide on capital, gross exposure, market concentration and acceptable loss before a subscription becomes active. Know which controls block new risk and which request closing existing exposure. Smaller limits can make your results diverge from a leader by design. Review confirmed engine state and current positions after a risk event rather than assuming the account is flat. - Inspect the whole account, including other positions. - Read current margin mode and timestamps. - Treat liquidation levels as estimates. - Keep capital within an amount you can afford to lose. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Liquidations](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/liquidations) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) - [Pause, stop and close: copy-trading controls explained](https://walletfollow.ai/learn/pause-stop-and-close-copy-trading) - [Drawdown and recovery in wallet performance](https://walletfollow.ai/learn/drawdown-and-recovery) --- # Why follower returns differ from the leader Canonical: https://walletfollow.ai/learn/why-follower-returns-differ Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Follower returns differ because the follower starts at a different time, trades with different capital and constraints, and receives its own fills and costs. A leader’s historical PnL is neither a forecast nor the follower’s account result. Compare actual account outcomes on a consistent basis. ## Joining changes the starting conditions A leader may have built a position over days before you subscribed. Your account can enter at a new price, skip existing exposure or apply a defined startup rule. The leader’s lifetime gain on that position is unavailable to a follower joining later. Comparing both accounts from the same calendar date still requires checking whether their starting exposures match. Deposits and withdrawals add another difference. A follower who changes committed capital during a trade can produce a different money-weighted result even when the underlying strategy targets stay similar. ## Limits intentionally change the copied book Allocation, leverage caps, market exclusions, concentration limits and drawdown controls can block or reduce trades. A minimum-size constraint can prevent a small account from holding a proportional position. These are meaningful execution outcomes, not missing spreadsheet rows that should be silently assumed to have filled. Review skipped-trade reasons and the confirmed account exposure. If a risk limit protected your account from a loss, it could also keep the account out of a subsequent recovery. A risk setting changes the path rather than preserving the leader’s return with less downside. ## Costs and timing accumulate Follower executions can incur a spread, slippage, exchange fees and any applicable builder fee. Funding depends on the actual position and time held. A creator performance fee, when enabled and applicable, has its own schedule and profit basis. Keep these costs separate so you can distinguish trading quality from the cost of reproducing it. Hypothetical example: both accounts have a $100 gross gain, but the follower incurs $15 of execution costs and the leader $5. Their net results differ by $10 despite the same quoted gross outcome. Different fills can enlarge that difference further. ## Compare three distinct records Keep the leader’s observed result, a portfolio’s modeled result and your account’s executed result separate. A simulation can explain a strategy method but cannot replace your fill ledger. When reconciling a divergence, start with startup time and target exposure, then inspect fills, costs, limits and cash flows. - Use the same period and return definition. - Check the subscription’s active dates. - Read actual fills and blocked-order reasons. - Label simulations and estimates explicitly. ## Sources and further reading - [Hyperliquid: Fees](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/fees) - [Hyperliquid: Funding](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Slippage and latency in copy trading](https://walletfollow.ai/learn/copy-trading-slippage-and-latency) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) - [Copy-portfolio backtests, tracked results and live returns](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) --- # Pause, stop and close: copy-trading controls explained Canonical: https://walletfollow.ai/learn/pause-stop-and-close-copy-trading Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Pause, stop and close have different effects. Pause blocks new copy risk while the defined exit management continues. Stop leaves existing positions in the account without copy management after confirmation. Close requests reduction of copied positions and can finish with residual exposure if execution constraints prevent a full exit. ## Pause changes ongoing behaviour A pause is useful when you want to prevent new exposure while allowing the engine’s supported reductions or exits to continue. It does not automatically make the account flat. Existing positions can keep gaining, losing and paying funding. Resume is a separate command whose availability depends on the current account and engine capabilities. Always read the specific dialog. Controls operate on a defined subscription or account scope. Pausing one subscription does not necessarily pause other subscriptions or manual trading in the same account. ## Stop can leave risk in your wallet Stopping or unfollowing with keep positions ends ongoing copy management after the engine applies the command. The positions remain yours. You then need to decide how to manage them through your venue account. Until confirmation, the command can be queued or in progress, and the previously active behaviour may still apply. Use this option only after reviewing current positions and any working orders. A status label in a research guide cannot establish what your individual account is doing now; the subscription’s confirmed state and venue account are the relevant evidence. ## Close is an execution request with safeguards WalletFollow’s close workflow requests reduce-only exits within the selected price bound and deadline. Reduce-only prevents the requested order from increasing the intended position, but liquidity, minimum size and partial execution can leave residuals. An exhausted deadline is not proof that all risk disappeared. A tighter exit bound can reject an unfavorable price and leave more open exposure. A wider bound can accept a larger execution cost. Inspect the dialog’s current parameters and final result rather than assuming close means immediate, complete execution at the displayed quote. ## Wait for the confirmed outcome Requested, received, applied and closed describe different stages. An unavailable engine can delay delivery. A close can end as closed except for explicitly listed residual positions. After any control change, review the effective state, open positions and remaining working orders. Revoking an agent’s venue authority is a separate account permission action and should be understood before use. - Pause: inspect remaining managed exposure. - Stop: inspect positions that become unmanaged. - Close: inspect fills, deadline and residuals. - Verify engine confirmation and venue state. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Order book](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/order-book) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Leverage and liquidation risk in copy trading](https://walletfollow.ai/learn/copy-trading-leverage-and-liquidation) - [Agent approvals and wallet permissions on Hyperliquid](https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions) - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) --- # How to assess diversification in a copy portfolio Canonical: https://walletfollow.ai/learn/diversifying-a-copy-portfolio Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading A copy portfolio is diversified only to the extent that its risk sources differ. Several wallets can hold the same leveraged market exposure and lose together. Inspect leader behaviour, market concentration, direction and overlapping return history alongside the published capital weights. ## Count risk sources rather than addresses Three different leaders can all be long the same market. Their names and trade frequencies may differ while their dominant price risk is nearly identical. Adding another address can increase total exposure without adding a useful offset. Begin with each leader’s current markets and direction, then look at how those patterns change through history. Likewise, several correlated crypto assets can respond together to market stress. Ten market rows do not prove ten independent bets. Label concentration in plain language so the portfolio remains understandable outside a chart. ## Capital weights are not risk shares A leader assigned a small capital weight can dominate risk if it uses much larger notional exposure or trades more volatile markets. Hypothetical example: half the capital follows an account using one-times exposure and half follows an account using five-times exposure. Equal capital weights produce substantially unequal gross exposure contributions. Inspect both gross and net exposure. Opposing positions can reduce directional exposure while still generating fees, funding and basis risk. A low net number alone does not describe the cost or fragility of the combined book. ## Correlation needs usable overlapping evidence Return correlation asks how measured outcomes move together over a shared period. It depends on the sampling interval, account-return method and coverage. Two accounts with little overlapping history cannot support a confident diversification conclusion. A missing correlation should remain unknown rather than being treated as zero. Historical correlation can also change. Leaders who looked different in quiet markets may react similarly during a sharp decline. Use correlation as one piece of evidence together with actual exposure and observed stressed periods. ## Review updates and combined account risk A portfolio version can change leaders or weights, which changes the diversification assessment. Revisit the combined book after an update and after a material leader behaviour change. Include other subscriptions and manual positions in the follower account. Diversification can reduce concentration, but it does not remove market loss, execution problems or correlated stress. - Check shared markets and directional bias. - Compare gross exposure contributions. - Require adequate overlap for correlation. - Review cash handling and portfolio versions. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Long, short, gross and net wallet exposure](https://walletfollow.ai/learn/long-and-short-exposure) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) - [Copy-portfolio backtests, tracked results and live returns](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) --- # Agent approvals and wallet permissions on Hyperliquid Canonical: https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Public wallet research needs only an address. Copy-account setup can require separate ownership, agent and builder-fee approvals. In WalletFollow’s approved connection flow, the wallet signs sign-in ownership proof, approval of the engine-reported agent and approval of the pinned builder fee, with a maximum of 0.05%. ## Understand which permission is being requested A sign-in message proves control of an address for an application session. It is different from authorizing an agent to act at the venue. Builder-fee approval establishes a permitted fee for a specified builder address; it is not itself an order. Read the domain, account, agent address and fee details presented in the wallet before confirming. A public profile, educational article or AI report should not require any of these signatures. If you only intend to research a wallet, do not treat a trading approval as a necessary reading step. ## The engine and the platform have different roles The registered copy engine generates and holds its agent keys and performs permitted execution. WalletFollow’s platform servers do not hold your wallet private key or agent private key and do not sign venue orders. The browser’s supported approval flow submits the permitted approval actions directly to Hyperliquid. This separation does not remove trading risk. An authorized execution agent still has meaningful authority, so verify that it is the agent assigned to your account by the registered engine rather than an address copied from an unofficial message. ## Keep key types separate A Data API key permits access to WalletFollow’s metered read-only data service. A Hyperliquid agent wallet is a venue signing identity. The two are not substitutes. An AI assistant researching public data should receive only the intended data-access capability through a secure integration, not a wallet seed phrase or trading agent key. To query a Hyperliquid account’s public information, use the actual main or subaccount address. The agent address can yield an empty account view because its role is to sign for another account. ## Review revocation and current status Know where to inspect and revoke venue approvals. Revoking authority can prevent future permitted actions, but it does not automatically close existing positions. Review open exposure and the subscription’s effective state as separate tasks. Never sign a transfer, withdrawal or invoice-payment action merely because it appears in a WalletFollow-branded research workflow. - Check the sign-in domain and message. - Check the exact engine-reported agent. - Check the pinned builder and fee ceiling. - Keep seed phrases and private keys out of the platform. ## Sources and further reading - [Hyperliquid: Nonces and API wallets](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/nonces-and-api-wallets) - [Hyperliquid: Builder codes](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/builder-codes) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Hyperliquid copy trading explained](https://walletfollow.ai/learn/hyperliquid-copy-trading-explained) - [Data API keys, quotas and read-only access](https://walletfollow.ai/learn/api-keys-quotas-and-read-only-access) - [Pause, stop and close: copy-trading controls explained](https://walletfollow.ai/learn/pause-stop-and-close-copy-trading) --- # Copy-trading fees: exchange, builder, funding and creator costs Canonical: https://walletfollow.ai/learn/copy-trading-fees Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Copy-trading costs can include exchange fees, slippage, funding and applicable builder or creator fees. Each has a different basis. WalletFollow’s permitted builder approval is capped at 0.05%; the current setup screen, fee schedule and executed records determine which charges actually apply. ## Read the basis of each cost Exchange execution fees depend on the venue’s current schedule and the relevant account and trade conditions. Builder fees attach to qualifying executed notional under an approved builder arrangement. Funding is a periodic transfer associated with holding a perpetual position. Slippage is a price difference, not always an explicit line item on a ledger. Do not sum percentage rates blindly. A rate charged on notional and a performance fee charged on qualifying profit use different denominators. A zero-profit period can still incur exchange fees, builder fees, funding and slippage. ## Translate a rate into a concrete example Hypothetical example: a 0.05% fee on $10,000 of qualifying executed notional is $5, because $10,000 multiplied by 0.0005 equals $5. This is an arithmetic example, not a prediction of a subscription’s bill. WalletFollow’s engine-link fee terms apply the builder fee to risk-increasing fills; reduce-only exits do not carry a WalletFollow builder fee. Exchange fees and execution costs can still apply to exits. Frequent turnover can create substantial cumulative notional even with modest capital. Comparing one fee to the account’s deposit amount understates cost when that capital is traded repeatedly. Review total executed notional over the period. ## Creator fees require their own status and rules A portfolio can display a declared creator fee while collection is inactive. Eligibility, platform capabilities and the bound schedule matter. If a performance fee is enabled, read its profit basis, high-water mark, crystallization cadence and treatment of losses before subscribing. A displayed simulated fee in a backtest is not evidence that money was collected. WalletFollow’s platform does not move, deduct, hold or pay out user or creator funds. An invoice record is distinct from a payment. This educational guide does not ask your wallet to sign an invoice payment. ## Reconcile net results without counting costs twice Some venue fill records include a builder component inside the total fee. Adding the component again would overstate cost. Keep gross trade PnL, total execution fee, funding and any separately applicable creator cost in a documented reconciliation. The current account screen and fee disclosure take precedence over a generic example when assessing your setup. - Check rate, denominator and applicable fill type. - Distinguish declared, simulated and active fees. - Review turnover and funding over the whole period. - Inspect total fill-fee treatment before adding components. ## Sources and further reading - [Hyperliquid: Fees](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/fees) - [Hyperliquid: Builder codes](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/builder-codes) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Hyperliquid funding rates and wallet returns](https://walletfollow.ai/learn/hyperliquid-funding-rates) - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) - [Copy-portfolio backtests, tracked results and live returns](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) --- # Copy-portfolio backtests, tracked results and live returns Canonical: https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading A backtest reconstructs how rules might have behaved on historical inputs. Tracked portfolio statistics describe the published portfolio under its stated method. Live follower returns come from an individual account’s actual fills, costs and cash flows. These are separate evidence types and should be labeled separately. ## Ask when the portfolio definition became real A portfolio can be assembled today from leaders who performed well last month. Showing their earlier gains as if the portfolio existed then introduces selection bias. The publication boundary matters: newly chosen weights should not receive credit for returns earned before their effective published version. A useful review identifies publication dates, version changes and whether each chart segment is reconstructed or tracked. Keep the two segments visually and analytically distinct. A long combined line can make a short actual record appear more established than it is. ## Read the model’s assumptions A reconstruction needs rules for weighting, cash handling, fees, funding, reinvestment and risk caps. It also needs sufficiently covered leader inputs. A missing leader observation can change the available result. Model returns are sensitive to the sampling interval and whether allocations rebalance or drift. Hypothetical example: a fixed initial dollar allocation and a reinvested proportional allocation can diverge after gains because the second changes its future trade size. Neither model is automatically your actual account behaviour. Read which series you are viewing. ## Execution cannot be assumed from a modeled target A backtest can calculate a target allocation without experiencing the follower’s real order-book depth, latency, minimum sizes or skipped orders. A simulated cost allowance can improve realism but remains an assumption. Real fills need their own ledger and timestamps. Leader performance also differs from portfolio performance. A portfolio result combines weights and rules; a follower adds personal settings and startup conditions. Keep leader closed-trade statistics from being presented as the copier’s trade statistics. ## Use the record for specific questions Backtests can help examine a rule’s behaviour and sensitivity. Tracked statistics can help assess the published strategy’s observed path. Actual follower results answer what happened in that account. None establishes future profitability. In WalletFollow, read basis labels, coverage explanations and version history before comparing two portfolios. - Separate prepublication and tracked history. - Document weighting and cash assumptions. - Distinguish simulated costs from charged costs. - Use actual fills for a follower-account reconciliation. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Fees](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/fees) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) - [Wallet data freshness and historical coverage](https://walletfollow.ai/learn/wallet-data-freshness-and-coverage) - [How to assess diversification in a copy portfolio](https://walletfollow.ai/learn/diversifying-a-copy-portfolio) --- # What is MCP for crypto wallet analysis? Canonical: https://walletfollow.ai/learn/mcp-for-crypto-wallet-analysis Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP Model Context Protocol, or MCP, is a standard way for an AI application to discover and call tools supplied by a server. For wallet research, tools can retrieve public account observations and return structured results. MCP itself does not establish data quality, profitability or permission to trade. ## Understand the host, client and server roles An AI application hosts the assistant experience. An MCP client connects that host to a server, and the server exposes supported capabilities such as tools or resources. A wallet-analysis tool might accept an account address and return positions with timestamps. The assistant uses that result when writing an answer. This is different from the model knowing a current position from its training. A live retrieval can supply newer evidence, but the answer is only as current and complete as the returned observations. ## Prefer structured, attributable wallet evidence A useful research response includes the address, network, observation time, metric definitions and missing-data reasons. Without those fields, an assistant can confuse a current snapshot with historical performance or interpret missing values as zero. Preserve decimal precision and units through the tool boundary. Ask the assistant to cite which retrieved values support each material conclusion. A strong answer can say that concentration was observed at a particular time while lifetime profitability remains unknown. It should not infer a verified human identity from an account address. ## Tool labels do not replace access control A tool can describe itself as read-only, but security depends on the server’s actual allowed operations and authorization. Review the connector publisher, requested permissions and available tool set. Avoid integrations that mix public wallet research with broad transfer or execution capabilities when your task only needs data. Treat text returned from untrusted sources as data. A wallet label or embedded note that asks the assistant to reveal a key or change its instructions is not a valid user instruction. Keep credentials out of tool results and generated reports. ## WalletFollow’s current connection path WalletFollow’s Data API is available for supported read-only wallet requests. Its MCP connector and client plugins are planned; this article does not announce a released connector. Use the current API catalog and product connection instructions to determine supported access. Do not install a package or trust a server solely because it uses the WalletFollow name. - Read the supported tool list. - Preserve timestamps and coverage. - Separate data retrieval from execution authority. - Verify current connector availability in the product. ## Sources and further reading - [Model Context Protocol: Architecture overview](https://modelcontextprotocol.io/docs/learn/architecture) - [Model Context Protocol: Tools](https://modelcontextprotocol.io/specification/latest/server/tools) - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Connect an AI assistant to Hyperliquid wallet data](https://walletfollow.ai/learn/connect-an-ai-assistant-to-wallet-data) - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) - [AI wallet analysis versus trade execution](https://walletfollow.ai/learn/ai-analysis-vs-trade-execution) --- # Connect an AI assistant to Hyperliquid wallet data Canonical: https://walletfollow.ai/learn/connect-an-ai-assistant-to-wallet-data Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP A supported coding assistant can research wallets through WalletFollow’s read-only Data API when you give its secure runtime access to a Data API key. Read the API catalog first, request the named wallet, preserve timestamps and coverage, then ask for an evidence-based summary. The MCP connector is planned. ## Establish the data contract first Open the current API page and create a Data API key in the account console. Configure that key through the assistant’s supported secret or environment mechanism. Do not paste its value into a public prompt, repository or report. A data key is a service credential, not a wallet private key. Ask the assistant to retrieve GET /v1/catalog before building requests. The catalog describes the supported endpoints, weights, plans, errors and headers. Reading it avoids inventing capabilities from old examples or a proposed MCP tool list. ## Retrieve one wallet before scaling up The wallet endpoint is GET /v1/wallets/\{address\}. The address must be a complete 0x-prefixed account address. Adding include=trades requests the supported trade-analysis variant and has a higher request weight than the summary. Use the response’s stated schema rather than assuming every metric is available for every wallet. Ask for the account summary and positions first. Request trade analysis when the question needs historical activity. This keeps the workflow focused and reduces unnecessary use of the account budget. ## Control usage and failure behaviour Use GET /v1/usage and the response headers to inspect the account’s remaining budget. Respect Retry-After on rate-limit responses. A 401 means authentication needs attention; repeatedly retrying it does not fix the credential. A coverage gap or upstream error should be reported as unavailable evidence, not silently replaced by a plausible value. Start with a small, explicit address list. Deduplicate requests and cache observations only for a period suitable for the task. A long-lived cache may work for a historical note but be unsuitable for a current exposure question. ## Ask for an auditable answer A good final report names the address, observation time, review window and metric basis. It separates observed positions from interpretations and explains missing inputs. The supported data workflow does not authorize trading, transfers or withdrawals. If an assistant proposes an action, review it as a separate decision rather than treating the data connection as account permission. - Read the catalog and current API instructions. - Configure credentials outside the conversation text. - Retrieve only the evidence needed. - Require timestamps, coverage and source values in the report. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Data API keys, quotas and read-only access](https://walletfollow.ai/learn/api-keys-quotas-and-read-only-access) - [What is MCP for crypto wallet analysis?](https://walletfollow.ai/learn/mcp-for-crypto-wallet-analysis) - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) --- # Read-only AI prompts for wallet research Canonical: https://walletfollow.ai/learn/safe-ai-wallet-research-prompts Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP A useful wallet-research prompt defines the address, period and question, requires source timestamps and missing-data explanations, and limits the task to analysis. Ask the assistant to separate observed facts from interpretations. Avoid prompts that turn a high recent PnL into an automatic trading recommendation. ## Prompt for current exposure Use this example after configuring secure Data API access: Analyze the named Hyperliquid wallet’s current positions. Report the account as-of time, largest markets, long and short notional, effective exposure where supported, and any missing inputs. Explain concentration in plain language. Do not place orders, request wallet signatures or guess unavailable values. The address belongs in the task, but the API key does not. Ask the assistant to show the source fields behind material conclusions so you can verify that a quoted value comes from the intended account. ## Prompt for a performance reconciliation A more specific example is: Compare this wallet’s seven-day realised result with its mark-to-market result on the same window. Explain fee and funding treatment, capital flows, account scope and coverage. If the data cannot support one basis, state that explicitly. Do not call the difference an error until definitions and timestamps match. This prompt directs the model toward the accounting question. Asking only why did the wallet make money encourages a story about motivation that public fills may not support. ## Prompt for a copy-trader review Try: Prepare a research checklist for this candidate over thirty days. Review coverage, sample size, drawdown, concentration, holding periods and execution sensitivity. Separate leader history, portfolio simulation and any actual follower evidence. Identify what would need checking before copying, without claiming the trader is suitable for me. If comparing candidates, require the same period and metric basis. Let the assistant mark a field unavailable rather than force a complete ranking. A partial comparison with honest limits is more useful than a confident table built on incompatible data. ## Review the report for common failures Check for invented balances, unnamed periods, missing timestamps, unverified identities and percentages without denominators. Look for statements that turn historical performance into a forecast. Also review whether retrieved labels or external text influenced the assistant’s instructions. Those inputs should remain evidence, not commands. Save the request and relevant source values if the report will inform a repeated review. - Define the question and review window. - Require original source values and observation times. - Keep credentials and wallet approvals separate. - Ask for unknowns and falsifiable interpretations. ## Sources and further reading - [Model Context Protocol: Tools](https://modelcontextprotocol.io/specification/latest/server/tools) - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Connect an AI assistant to Hyperliquid wallet data](https://walletfollow.ai/learn/connect-an-ai-assistant-to-wallet-data) - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [AI wallet analysis versus trade execution](https://walletfollow.ai/learn/ai-analysis-vs-trade-execution) --- # Data API keys, quotas and read-only access Canonical: https://walletfollow.ai/learn/api-keys-quotas-and-read-only-access Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP WalletFollow Data API keys authenticate supported read-only requests and consume a plan’s weighted budget. They do not authorize venue orders, transfers or withdrawals. Keep the key in a secure runtime, read the current catalog for limits, and use usage information and retry headers to control request volume. ## A service key is not a wallet key A Data API key identifies access to a data account. A wallet private key controls a signing identity, and an engine agent key carries venue execution authority. Never substitute one for another. Public-wallet research only needs the account address plus the data-service authentication required by the chosen integration. Create and revoke service keys through the account console. Give each integration a clear purpose so you can remove access when it is no longer needed. Avoid placing keys in URLs, browser-visible client bundles, screenshots or saved AI transcripts. ## Count request weight rather than just requests The API catalog lists endpoint weights and variants. The wallet summary currently uses a lower weight than the variant requesting trade analysis. Ten detailed requests can therefore consume more budget than ten summaries. The correct workload estimate multiplies each request count by its catalog weight. Plan limits can apply over short bursts, minutes and monthly usage. A client that stays below the monthly allowance can still exceed a burst limit. Use a controlled queue and avoid launching hundreds of wallet lookups simultaneously. ## Handle errors according to their meaning On 429, read the error code and Retry-After before retrying. A minute limit and an exhausted monthly quota have different recovery conditions. On 401, review whether the key is missing, expired or revoked. On an unavailable-data response, preserve that result rather than fabricating a successful wallet report. Use GET /v1/usage to monitor the account budget and GET /v1/catalog for current endpoint details. A free or paid plan label alone does not tell a client exactly how much a specific workload will consume. ## Protect reports and logs Log request identifiers, endpoint names and error codes without the Authorization header. Redact credentials from exception messages and copied commands. An assistant’s final research output should include public addresses and relevant data observations, not the key that retrieved them. If a key appears in a public place, revoke it and update the integration rather than assuming deletion removes every copy. - Store keys outside source code and prompt text. - Deduplicate requests and choose the lightest useful variant. - Respect retry instructions. - Review access and revoke unused integrations. ## Sources and further reading - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Connect an AI assistant to Hyperliquid wallet data](https://walletfollow.ai/learn/connect-an-ai-assistant-to-wallet-data) - [What is MCP for crypto wallet analysis?](https://walletfollow.ai/learn/mcp-for-crypto-wallet-analysis) - [Agent approvals and wallet permissions on Hyperliquid](https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions) --- # AI wallet analysis versus trade execution Canonical: https://walletfollow.ai/learn/ai-analysis-vs-trade-execution Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP AI wallet analysis retrieves and interprets data. Trade execution changes an account through an authorized engine or venue workflow. Connecting an assistant to read-only wallet data does not grant execution authority, and an AI-generated suggestion is neither a verified signal nor an order confirmation. ## Identify what the assistant actually knows A model can summarize observations supplied by an API, compare defined metrics and help prepare a review checklist. It can also misread a field, invent a missing value or produce an unsupported explanation. A confident answer is not evidence that a current wallet lookup succeeded. Require the address, timestamp, period and original values supporting the conclusion. If the assistant used only a screenshot or old article, it should say so. Public wallet behaviour also lacks off-platform positions and the owner’s intent. ## Keep recommendations separate from permissions An assistant saying increase allocation does not change the subscription and should not be treated as customer consent. Portfolio setup, risk changes and activation use the product’s explicit account workflow. The engine’s current capability and authorization state determine whether a supported action can proceed. WalletFollow’s read-only Data API does not provide an order-signing key. The planned MCP connector should not be advertised as an execution system merely because the same brand also has a copy portfolio product. Each tool surface needs its own truthful scope. ## Use AI to challenge the research thesis Ask for alternative explanations, missing evidence and conditions that could invalidate an interpretation. A large short may be a hedge rather than a standalone bearish view. A high win rate may conceal rare severe losses. A strong month may reflect broad market direction rather than durable trader skill. These questions make the assistant useful without giving it authority over funds. They also make the final report easier to audit: every conclusion can be connected to evidence or marked as an unresolved hypothesis. ## Confirm actions through the responsible system If you independently choose to use copy trading, review the portfolio version, allocation, limits, fee schedule and engine state in the product. A submitted command, accepted command and filled order are different stages. Check confirmed results and remaining exposure. Never hand an assistant a wallet seed phrase or agent private key simply to produce a research summary. - Data access supports research. - Customer consent belongs to explicit account controls. - Engine state confirms supported execution. - AI explanations remain subject to source verification. ## Sources and further reading - [Model Context Protocol: Architecture overview](https://modelcontextprotocol.io/docs/learn/architecture) - [Model Context Protocol: Tools](https://modelcontextprotocol.io/specification/latest/server/tools) - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) - [Hyperliquid copy trading explained](https://walletfollow.ai/learn/hyperliquid-copy-trading-explained) - [Agent approvals and wallet permissions on Hyperliquid](https://walletfollow.ai/learn/agent-approvals-and-wallet-permissions) --- # Realised versus unrealised PnL Canonical: https://walletfollow.ai/learn/realized-vs-unrealized-pnl Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics Realised PnL comes from closing or reducing exposure under the accounting method. Unrealised PnL measures the current gain or loss on remaining open exposure. A profitable closed-trade record can coexist with a large open loss, so read both before assessing a wallet. ## Follow the position through its lifecycle A new position begins with exposure and an entry basis. As prices move, its marked gain or loss changes without the whole position being closed. Reducing exposure realizes a result on the closed portion while the rest remains open. A final close removes the remaining market exposure, subject to any residual quantity. A fill-level closed-PnL field and a grouped completed-trade result may differ in timing or aggregation. Read whether the tracker groups partial closes into a single position lifecycle or counts each execution separately. ## Use a simple numerical example Hypothetical example: ten units are bought at $100 and later valued at $110. Ignoring costs, the open position has $100 of unrealised gain. Selling four units at $110 realizes $40; the remaining six units have $60 of unrealised gain at that same price. The original $100 gain is split, not earned twice. If those six units later fall to $90, their unrealised result becomes a $60 loss. The earlier $40 realised gain still happened, but it does not establish a profitable final outcome for the whole sequence. ## Costs and account flows require separate treatment Exchange fees and funding can reduce the account’s economic result. Whether a displayed PnL already includes each cost depends on its stated basis. A deposit increases account value without producing a trading profit. A withdrawal can reduce value after a profitable close. These movements prevent raw balance change from serving as a universal PnL formula. Return on equity for one position is also different from the whole wallet’s return. It uses a margin-related basis rather than necessarily the total account capital. Do not compare a position percentage with an account percentage as if they share a denominator. ## Read both realised and current exposure For a wallet review, start with the period’s measured closed results, then inspect the current open book and its valuation time. Ask whether a few open positions dominate the account’s outcome. When comparing trackers, align performance basis and fee treatment before judging a discrepancy. A gain that remains unrealised can reverse before execution. - Check partial closes and grouping. - Use current marks for open exposure. - Avoid adding an included component twice. - Keep deposits and withdrawals outside trading profit. ## Sources and further reading - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [How to read wallet positions and trade history](https://walletfollow.ai/learn/reading-wallet-positions-and-trades) - [Mark price versus last trade price on Hyperliquid](https://walletfollow.ai/learn/mark-price-vs-last-trade-price) --- # ROI versus dollar PnL: comparing wallets fairly Canonical: https://walletfollow.ai/learn/roi-vs-dollar-pnl Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics Dollar PnL measures the money gained or lost on a stated basis. ROI expresses a result relative to a defined capital base. A larger account can earn more dollars with a lower percentage return, while a small denominator can make an unstable record look impressive. ## Start with the denominator A percentage is incomplete until you know what capital it divides by. Initial account equity, average capital, allocated strategy capital and position margin can all produce different percentages from the same dollar result. Read the tracker’s definition rather than treating every ROI label as interchangeable. Hypothetical example: $1,000 profit on $10,000 of starting capital is ten percent in a simple no-flow calculation. The same $1,000 on $100,000 is one percent. The dollar gain matches while the scale of the capital commitment differs. ## Capital changes complicate return comparisons A deposit midway through a period changes the capital available to trade. Dividing the final profit by the original balance can exaggerate performance if new money funded much of the exposure. A withdrawal can create another misleading denominator. Cash-flow-aware methods answer different questions from an unadjusted equity change. Time-weighted methods aim to separate the strategy’s path from external capital movement. Money-weighted methods reflect the timing of invested capital. This guide does not assume WalletFollow exposes every possible method; use the basis actually stated beside the metric. ## Small bases need stronger scrutiny A wallet with very little starting equity can show a large percentage after a small absolute gain. Leverage, a deposit or a near-zero denominator can make the interpretation fragile. A guarded or unavailable ROI can be more honest than a spectacular percentage produced on an unusable base. Compare the size of the sample, capital and loss alongside the return. A high ROI on a tiny account does not demonstrate that the strategy can scale to larger orders with the same liquidity or execution quality. ## Build a consistent comparison Choose the same period, account scope, return method and cost treatment for every wallet. Add dollar PnL, capital, drawdown and exposure so the percentage remains interpretable. In WalletFollow, preserve unavailable values and their reasons instead of sorting them as zero. A leaderboard is only comparable to the extent that its inputs and eligibility rules match. - Name the capital denominator. - Adjust or explain deposits and withdrawals. - Review drawdown and effective exposure. - Treat high percentages on small samples cautiously. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Why wallet PnL differs between trackers](https://walletfollow.ai/learn/why-wallet-pnl-differs) - [Drawdown and recovery in wallet performance](https://walletfollow.ai/learn/drawdown-and-recovery) - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) --- # Drawdown and recovery in wallet performance Canonical: https://walletfollow.ai/learn/drawdown-and-recovery Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics Drawdown measures a decline from an earlier peak in a defined value or return series. Maximum drawdown is the deepest observed decline over the chosen period. Recovery requires a larger percentage gain than the loss percentage, and a drawdown control does not guarantee that losses stop at its threshold. ## Define the series before calculating the decline A raw account-value curve can fall because of a withdrawal rather than a trading loss. A cash-flow-adjusted series is intended to remove that distortion under its stated method. A realized-only curve can hide a large loss still open in the account. Review the underlying basis before comparing maximum drawdown between wallets. Observation frequency matters. Daily samples can miss an intraday trough visible in more frequent data. Maximum drawdown means maximum observed on that series, not necessarily every loss the account experienced between observations. ## Work through recovery arithmetic Hypothetical example: a series peaks at $10,000 and falls to $8,000. The decline is $2,000 divided by $10,000, or twenty percent. Returning from $8,000 to $10,000 requires $2,000 divided by $8,000, or twenty-five percent. A twenty-percent gain from the trough reaches only $9,600. A fifty-percent loss requires a one-hundred-percent gain to recover the starting value. This asymmetry explains why limiting large losses can matter even when a strategy has many profitable periods. The arithmetic is not a forecast of recovery time. ## Read duration and concentration alongside depth Two accounts can have the same deepest drawdown but very different recovery paths. One may recover quickly and another remain below its peak for months. Inspect time underwater and the exposures driving the decline when those observations are available. An account concentrated in one leveraged market can reach a deeper loss rapidly. A smooth historical chart does not prove future stability, especially if the record covers only favorable conditions or has missing observations. Include coverage and market context in the interpretation. ## Understand a copy drawdown control A configured drawdown threshold can pause new risk or request reduction under the selected policy. Existing positions can continue losing until they are actually reduced, and execution constraints can leave residuals. Check the engine’s confirmed state and the account’s remaining exposure after a trigger. Never describe a threshold as a guaranteed maximum loss. - Use a defined cash-flow and valuation basis. - Check sample interval and coverage. - Review both depth and recovery duration. - Distinguish a trigger from completed execution. ## Sources and further reading - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Liquidations](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/liquidations) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Leverage and liquidation risk in copy trading](https://walletfollow.ai/learn/copy-trading-leverage-and-liquidation) - [Pause, stop and close: copy-trading controls explained](https://walletfollow.ai/learn/pause-stop-and-close-copy-trading) - [ROI versus dollar PnL: comparing wallets fairly](https://walletfollow.ai/learn/roi-vs-dollar-pnl) --- # Win rate versus profit factor Canonical: https://walletfollow.ai/learn/win-rate-vs-profit-factor Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics Win rate measures how often counted trades win. Profit factor compares gross profits with the absolute value of gross losses on a stated basis. A high win rate can coexist with a losing strategy when occasional losses outweigh frequent gains. Both metrics need consistent trade definitions and sufficient observations. ## Understand what counts as a trade A single position can generate many partial fills and reductions. Counting fills instead of grouped completed positions changes the sample size and can change the apparent winning frequency. Breakeven treatment also matters: some methods include flat trades in the denominator and others do not. When comparing WalletFollow profiles, use the same review period and the displayed trade basis. A leader’s closed positions in a portfolio report are not automatically the follower account’s trades. A metric based on incomplete history should remain unavailable or explicitly limited. ## A high win rate can still lose money Hypothetical example: nine winning trades earn $10 each and one losing trade loses $120. The win rate is ninety percent, but the gross result is negative $30. Gross profits of $90 divided by absolute gross losses of $120 produce a profit factor of 0.75, before any costs excluded from the calculation. Reversing the shape can produce a lower win rate with positive profit: two wins of $100 and eight losses of $10 produce a $120 gross gain. Frequency alone does not identify which process is economically better. ## Read costs, sample size and open exposure A profit factor above one on gross trade PnL can fall below one after exchange fees, funding and execution costs. Read which costs the metric includes. Very few losses can make the ratio unstable; no observed losses can leave a finite ratio undefined rather than proving unlimited quality. Closed-trade metrics can also miss a large losing position that has not closed. Review the current book and marked loss alongside the historical wins. Otherwise a strategy that repeatedly postpones losses can look stronger than its account equity path. ## Combine complementary measures Use winning frequency to understand the pattern, average gain and loss to understand scale, and drawdown to understand the path of account risk. Include holding duration, concentration and coverage when evaluating copy fit. No single score can replace reading how the account trades. - Same completed-trade grouping. - Same cost treatment and review period. - Adequate wins and losses for interpretation. - Open-position risk checked separately. - No forecast from a historical percentage alone. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) - [Realised versus unrealised PnL](https://walletfollow.ai/learn/realized-vs-unrealized-pnl) - [A crypto trading journal for wallet research](https://walletfollow.ai/learn/crypto-trading-journal) --- # A crypto trading journal for wallet research Canonical: https://walletfollow.ai/learn/crypto-trading-journal Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics A useful crypto trading journal combines execution records with the reasoning, risk conditions and review rules present at the time. Wallet history supplies evidence about what happened; your journal records why you considered it and what you planned to check. Keep facts, assumptions and later outcomes distinct. ## Record a reproducible observation Begin each entry with the wallet address, network, observation time and chosen period. Record the source, metric basis and coverage warning. Add the market, direction, size, entry basis and relevant fees or funding when supported. These fields let you revisit the same evidence rather than relying on a remembered screenshot. For your own account, group fills using a consistent trade definition. An order executing in ten pieces is not necessarily ten decisions. If the tracker changes its historical grouping or coverage, note that change instead of interpreting it as trader behaviour. ## Write the thesis before the outcome State the reason for the research or trade in one or two concrete sentences. Identify the evidence that supports it, what is unknown and what would invalidate it. Avoid retrospective explanations such as obviously bullish after the market has already risen. For copy research, write why a leader made the shortlist, the reviewed drawdown, the capital and execution fit, and any concentration concern. A journal can be valuable even when you decide not to copy. A documented rejection helps preserve consistent criteria. ## Review decisions and results separately A profitable outcome can come from an unsupported decision, while a well-defined process can produce a loss. Review whether the evidence was current, the limits were understood and the execution matched the intended action. Then review the financial result on its own stated basis. Hypothetical entry: the candidate had high recent returns but only a short, concentrated long record. The decision was to continue observing. A later rally does not make the original concern incorrect; it supplies another observation for the next review. ## Keep the journal usable Choose a regular review cadence that matches the strategy’s holding period. Summarize recurring mistakes or unanswered questions rather than collecting unlimited screenshots. WalletFollow can supply public profile observations, while the reasoning and personal decisions remain yours. Do not place API keys, seed phrases or other credentials in journal entries, exported files or AI summaries. - Address, time, window and source. - Observed metrics and coverage. - Thesis and invalidation conditions. - Risk and execution assumptions. - Outcome and next review question. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Entry price and PnL](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/entry-price-and-pnl) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to read wallet positions and trade history](https://walletfollow.ai/learn/reading-wallet-positions-and-trades) - [Win rate versus profit factor](https://walletfollow.ai/learn/win-rate-vs-profit-factor) - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) --- # Hyperliquid funding rates and wallet returns Canonical: https://walletfollow.ai/learn/hyperliquid-funding-rates Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Hyperliquid Funding is a recurring payment between perpetual long and short positions under the venue’s rules. Its direction and rate can change. Funding affects the account’s net outcome while a position is held, so a profitable price move can still have a materially smaller result after funding costs. ## Funding is separate from entry and exit fees An execution fee arises when an order fills. Funding relates to holding a perpetual position over applicable funding times. A position can incur funding without trading again. A displayed unrealised price gain is therefore not necessarily the account’s full net gain from holding it. On Hyperliquid, read the official funding method and current market information instead of extrapolating from another venue’s schedule. A historical rate describes a particular interval, while a current or predicted rate is not a promise that future rates remain unchanged. ## Check the sign and position direction Under the usual positive-rate convention, longs pay shorts; at a negative rate, the direction reverses. Confirm the sign convention in the field or report you are reading, because funding received can be expressed differently from the rate itself. A positive funding amount in a wallet report needs its documented meaning. Hypothetical example: at a 0.01% rate for an applicable interval, a $10,000 position has a funding amount of about $1 in a simplified notional-times-rate calculation. The real amount depends on the venue’s defined price, size and accounting rules. ## Annualized displays can be misleading A large annualized percentage can result from multiplying one short interval across a year. It does not show that the same payment will persist. Funding can vary, reverse or be outweighed by market losses. Treat an annualization as a conditional arithmetic display rather than a yield guarantee. A delta-neutral-looking position also has basis, execution and margin risk. Public wallet data may show only one side of a hedge. Do not call an account a funding strategy merely because its observed direction currently receives payments. ## Include funding in copy research Leader and follower entry times can differ, so their held durations and funding totals can differ. Reconcile actual observed funding over the review window rather than using today’s rate for the whole history. Check whether portfolio simulations include funding and whether displayed PnL already includes it. Preserve unavailable funding history instead of assuming zero. - Read rate direction and amount convention. - Use actual held intervals where available. - Distinguish funding from fill fees. - Avoid treating an annualized rate as a forecast. ## Sources and further reading - [Hyperliquid: Funding](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding) - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Copy-trading fees: exchange, builder, funding and creator costs](https://walletfollow.ai/learn/copy-trading-fees) - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) - [Open interest versus trader positioning](https://walletfollow.ai/learn/open-interest-and-trader-positioning) --- # Open interest versus trader positioning Canonical: https://walletfollow.ai/learn/open-interest-and-trader-positioning Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Hyperliquid Open interest measures outstanding perpetual exposure under a market’s convention. Trader positioning describes the observed holdings of a defined wallet set. Volume describes traded activity over a period. These are different measurements, and none alone proves the market’s next direction. ## Open interest is not trading volume Trading volume accumulates executions during a window. Open interest describes outstanding exposure at a point in time. A market can trade heavily while outstanding exposure changes little as positions change hands. A single close can affect exposure without implying that the same amount of new money entered the market. Read units carefully. A value can be expressed in asset units, contract units or dollars at a particular valuation. Comparing two markets requires compatible units and timestamps. A changing dollar value can partly reflect price movement rather than a change in position quantity. ## A wallet sample is not the entire market A positioning view can aggregate watched wallets, discovered wallets or another eligible cohort. Its long and short totals describe that set. A mostly long sample does not mean the market lacks short exposure elsewhere. Inclusion rules, age filters and account coverage shape what the view can establish. In WalletFollow, read the cohort and as-of information supplied by the view. A page showing known traders or large accounts should not be described as every trader on Hyperliquid unless its method explicitly supports that scope. ## Combine measurements without inventing a signal Increasing open interest alongside a price rise can have several explanations, including new directional positions, hedging and spread activity. Funding can add context about the perpetual premium under its formula, but it does not identify every participant’s motivation. A large wallet trade adds an address-level observation, not proof of a broad consensus. Write an interpretation with alternatives. For example, a wallet cohort has increased observed long exposure while the market’s outstanding exposure also rose. That is a documented pattern; calling it certain accumulation by informed traders exceeds the evidence. ## Make a market review reproducible Record the market identifier, venue scope, timestamps, cohort definition and units. Compare changes only across compatible observations. Read current liquidity and concentration before using the review to assess copy execution. WalletFollow’s trade and wallet views support research, while the venue’s official market fields provide an independent reference for definitions. - Separate outstanding exposure from executed volume. - Identify the wallet cohort. - Check units and valuation time. - Use funding and price as context, not a guaranteed signal. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Funding](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Long, short, gross and net wallet exposure](https://walletfollow.ai/learn/long-and-short-exposure) - [Hyperliquid funding rates and wallet returns](https://walletfollow.ai/learn/hyperliquid-funding-rates) - [How to find and evaluate Hyperliquid whale wallets](https://walletfollow.ai/learn/how-to-find-hyperliquid-whales) --- # Mark price versus last trade price on Hyperliquid Canonical: https://walletfollow.ai/learn/mark-price-vs-last-trade-price Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Hyperliquid The last trade price records a recent execution. A mark price is a defined valuation reference used for purposes including unrealised PnL and margin. Neither guarantees the price available for your next order. Compare prices using the same market, timestamp and intended purpose. ## A trade and a valuation reference answer different questions A last trade proves that a particular quantity executed at a price recently. It does not show that another order of a different size can execute there now. The mark is designed as a robust valuation reference under the venue’s methodology. It can differ from the last trade, especially when trading is thin or prices move quickly. Hyperliquid documents separate oracle and mark references. Consult those definitions rather than assuming a price chart’s selected line is the exact basis of a margin field or an account’s unrealised result. ## Small price differences can create large PnL differences Hypothetical example: a long position of one hundred units has a $100 entry. A valuation at $101 implies $100 of gross open gain; a valuation at $100.50 implies $50. Both calculations use the same quantity and entry but a different current price. This simple example ignores contract-specific accounting and costs. Its purpose is to show why a tracker’s chosen price and timestamp matter. Before reconciling two wallet totals, align those observations and confirm the position size. ## Executable prices come from available liquidity The best bid and ask indicate available quotes, subject to changing size and conditions. A market order can reach additional levels if its quantity exceeds the top quote. A limit order can remain unfilled. A marked account profit is therefore different from the amount that would be realised by closing all exposure immediately. For copy trading, a slippage reference should be explicit. Comparing a follower fill with a leader’s old execution measures something different from comparing it with a fresh quote or mark at the follower’s decision time. ## Read margin information on its own basis Estimated liquidation and margin conditions follow the venue’s rules and current account state. A last-price spike alone does not define every liquidation outcome. Funding and other cross-margin positions can affect the account’s buffer. Keep timestamps and reference-price definitions in any analysis rather than treating a screenshot’s price as a complete risk model. - Identify last, mid, oracle or mark. - Check market and timestamp. - Use the displayed PnL basis. - Review actual order-book depth for execution questions. ## Sources and further reading - [Hyperliquid: Robust price indices](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/robust-price-indices) - [Hyperliquid: Order book](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/order-book) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [Realised versus unrealised PnL](https://walletfollow.ai/learn/realized-vs-unrealized-pnl) - [Slippage and latency in copy trading](https://walletfollow.ai/learn/copy-trading-slippage-and-latency) - [Leverage and liquidation risk in copy trading](https://walletfollow.ai/learn/copy-trading-leverage-and-liquidation) --- # Long, short, gross and net wallet exposure Canonical: https://walletfollow.ai/learn/long-and-short-exposure Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Trading metrics Long exposure generally gains when the referenced asset rises; short exposure generally gains when it falls. Gross exposure adds absolute long and short notional. Net exposure subtracts short from long under a compatible basis. Low net exposure can coexist with substantial gross risk and trading costs. ## Use a concrete notional example Hypothetical example: an account has $6,000 of long notional and $4,000 of short notional. Gross exposure is $10,000 and net exposure is $2,000 long. With $2,000 of equity, the account has five times gross exposure and one times net exposure in this simplified view. These calculations do not establish a hedge’s effectiveness. If the long and short positions reference different markets, their prices can move differently. Even opposite positions in related instruments can have basis and funding differences. ## Do not confuse a side percentage with market coverage A long-share percentage can divide long notional by the cohort’s gross notional. It describes how the selected observed book is distributed, not the probability that the market rises. Another display might count long accounts rather than weight by notional. Those percentages can differ substantially when one large account dominates size. Read whether spot holdings are included. A wallet short a perpetual contract can also hold spot exposure in the same asset, while a perp-only view shows just the short. Off-platform holdings may remain unknown. ## Concentration survives apparent balance A portfolio can have a near-zero net number while its largest market positions remain exposed to liquidity stress or rapid relative-price movement. Gross exposure matters for fees, funding and execution. Multiple leaders can also net against one another in a shared account while retaining distinct strategy attribution. Review the largest market shares and effective account leverage. A capital weight assigned to one leader is not the same as its contribution to the combined exposure. A highly leveraged leader can dominate a portfolio despite a modest weight. ## Keep the snapshot aligned Combine observations only when their timestamps and scope support the comparison. If one leader’s position is fresh and another is old, a calculated net can create an apparent hedge that no longer exists. Preserve delayed or unavailable fields. In WalletFollow, use as-of and coverage explanations when writing an account or portfolio risk note. - Gross equals absolute long plus short exposure. - Net expresses the directional difference on a defined basis. - Check concentration and market relationships. - Keep account equity, scope and timestamps attached. ## Sources and further reading - [Hyperliquid: Info endpoint](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to assess diversification in a copy portfolio](https://walletfollow.ai/learn/diversifying-a-copy-portfolio) - [Open interest versus trader positioning](https://walletfollow.ai/learn/open-interest-and-trader-positioning) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) --- # A repeatable workflow for Hyperliquid wallet research Canonical: https://walletfollow.ai/blog/a-repeatable-wallet-research-workflow Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Wallet tracking A repeatable wallet review starts with a written question and ends with an evidence record. Keep the address, period, metric definitions and observation times consistent, then record what changed. This makes the next review comparable even when the wallet or the available history changes. ## Write the question before opening the leaderboard Choose a question that public data can answer: Is this wallet becoming more concentrated? Does its recent result depend on one open position? Is its current exposure consistent with its observed trading history? These are more useful research tasks than searching for the best trader, which leaves both best and the evidence period undefined. Create a note with four fields: account address, review period, question and comparison basis. Use the account you intend to inspect, rather than assuming its agent address represents the same history. Start from WalletFollow Discover, open the profile and confirm the full address. Record unavailable inputs at this stage instead of repairing them with guesses later. ## Capture a small evidence packet Record the relevant observation times, account value, open exposure and performance basis. Add the history coverage and any warning that changes interpretation. Keep a position snapshot separate from a closed-trade result: they answer different questions and may rely on different inputs. A screenshot is useful context, but a number without its label or timestamp is difficult to audit. For each conclusion, attach the fact that supports it. For example, a concentration observation should identify the dominant market and how exposure was measured. A statement about improving results needs a consistent earlier comparison. Keep the packet small enough that another reader could check the argument without reconstructing your entire browsing session. ## Test a competing explanation Before interpreting an equity increase as skill, check whether capital arrived. Before describing a position as a new bet, check whether the fills reduced an existing trade. Before praising a high win rate, read the sample size and loss amounts. A useful research note includes one alternative explanation and says whether the available evidence can distinguish it. Public wallet analysis may leave external hedges, owner identity and missing history unresolved. Label those limits explicitly. Your conclusion can remain narrow: the observed account has concentrated exposure at the recorded time. It does not need to become a prediction about the person or the next market move. ## Make the next review a comparison Keep the original note and append the new observation. Separate changes in positions from changes in coverage or calculation. If a metric becomes available after a history update, record that as new evidence rather than automatically calling it improved performance. If the review period changes, start a new comparison instead of silently replacing the old baseline. Finish with the question answered, the supporting observations, the remaining uncertainty and the next fact to check. The workflow produces a research record; following an address and activating a copy subscription remain separate decisions. ## Sources and further reading - [Hyperliquid: Info endpoint and account data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to track a Hyperliquid wallet](https://walletfollow.ai/learn/how-to-track-a-hyperliquid-wallet) - [Wallet data freshness and historical coverage](https://walletfollow.ai/learn/wallet-data-freshness-and-coverage) - [How to read wallet positions and trade history](https://walletfollow.ai/learn/reading-wallet-positions-and-trades) --- # Why a leaderboard is not a copy trading strategy Canonical: https://walletfollow.ai/blog/why-a-leaderboard-is-not-a-copy-strategy Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading A leaderboard orders wallets by a chosen historical metric. A copy strategy must also specify allocation, risk, execution and review rules for the follower. Ranking provides candidates for research; it cannot establish whether the next trade is suitable or reproducible in a different account. ## The sort order contains assumptions A dollar-PnL ranking favors large gains in money terms. An ROI ranking asks about returns relative to a capital basis. A win-rate ranking counts successful outcomes under a trade-grouping rule. None of these sort orders contains every fact you need. Decide which question a ranking answers before treating its first row as your preferred candidate. The period matters too. A short window can reward a single fortunate move, while a long window can hide a recent change in behaviour. Coverage, minimum-account rules and inclusion criteria shape who appears. A wallet outside the displayed set is not automatically worse; it may be excluded, inactive or insufficiently covered for the metric. ## Starting with winners can obscure the selection process Finding an account after its successful run is different from selecting it beforehand. If you only inspect the current leaders, you do not see every account that took similar risks and lost. Historical winners can be useful case studies, but their position on a list does not measure the reliability of choosing the next winner. A practical response is to write selection rules before tracking new outcomes. Preserve your shortlist, its date and rejected candidates with their reasons. Later, compare the entire recorded selection rather than highlighting the best surviving result. This is a research method, not a claim that WalletFollow has already measured the predictive value of its rankings. ## A follower inherits a different execution problem The leader may already hold a position at a price unavailable to a new follower. The follower may have different collateral, existing exposure, order sizes or eligible markets. Fees, funding, latency and slippage can widen the difference. A leader’s historical PnL therefore cannot be presented as the follower’s expected result. Hypothetical example: one candidate makes infrequent trades in liquid markets; another changes small positions rapidly. Even if their historical headline returns match, the operational demands of copying them differ. Review holding time and trade size alongside market liquidity, rather than assuming one allocation rule suits both. ## Turn a candidate into a written decision Record the chosen review period, coverage, maximum observed drawdown, concentration and open risk. Then define the proposed allocation, exposure limits and conditions for another review. Check what pause, stop and close mean before relying on a control. Live copy availability depends on the account, approvals and the separate engine’s current capabilities. A useful decision may be to keep observing because a key input is missing. The leaderboard has still done its job by surfacing a wallet worth investigating. Copying becomes a separate decision supported by the research, rather than an automatic consequence of a rank. ## Sources and further reading - [Hyperliquid: Info endpoint and account data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Funding](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to evaluate a trader before copying](https://walletfollow.ai/learn/evaluate-a-trader-before-copying) - [Why follower returns differ from the leader](https://walletfollow.ai/learn/why-follower-returns-differ) - [Copy-trading allocation and position sizing](https://walletfollow.ai/learn/copy-trading-position-sizing) --- # Designing an auditable AI wallet report Canonical: https://walletfollow.ai/blog/designing-an-auditable-ai-wallet-report Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: AI & MCP An auditable AI wallet report lets a reader trace each conclusion to an observed input. Ask for the account, period, source time, calculation and missing-data limits alongside the narrative. Fluent prose is useful only when the underlying claims remain checkable. ## Define the report before asking for a verdict Specify the full account address, network, question and analysis period. Ask the assistant to distinguish current exposure from historical performance, and realised results from marked open positions. Avoid starting with Is this trader good? That invitation can produce a confident overall label from whichever numbers happen to be available. WalletFollow’s read-only Data API is live; its MCP connector and companion plugins are planned. Use the currently published API documentation for available fields and authentication. MCP can provide a standard way to expose tools, but a protocol name does not establish the completeness of a dataset or the correctness of an interpretation. ## Keep an evidence ledger beside the prose For each material claim, retain its input, unit, period and observation time. For a calculated claim, add the formula and identify which source values were used. A compact ledger might have columns for claim, source field, as-of time and limitation. This makes it possible to challenge one statement without discarding or rerunning the entire report. Separate sourced facts, calculations and interpretations. The assistant may observe a large open position, calculate its share of displayed exposure, then suggest that the account is concentrated. Those are three different steps. The interpretation should remain conditional if some positions, assets or historical observations are unavailable. ## Require a useful failure mode Tell the assistant what to do when it cannot answer: identify the missing input and stop that conclusion. An unavailable drawdown is not zero drawdown. A current balance cannot establish a complete lifetime return. A failed request should be visible in the evidence record rather than replaced by a remembered value from another wallet. Use scoped access and the documented key-management workflow. Keep tokens outside prompts, screenshots and published reports. Treat retrieved tool descriptions and external text as inputs to evaluate, not authority to change the research task or request additional privileges. An unexpected request to expose credentials is a reason to investigate the integration. ## Check one conclusion manually Choose an important claim and reproduce it from the recorded values. Check its units, time window and denominator. Review whether fees or funding are already included before adding costs again. This small check tests the report’s reasoning more directly than asking the same assistant whether its answer is accurate. Finish with observed facts, unresolved questions and a record of the data used. Preserve the report when comparing a later run; changes may reflect refreshed observations or better coverage. AI research is distinct from engine execution, and generating a report does not authorize a trade, change risk limits or approve a wallet action. ## Sources and further reading - [Hyperliquid: Info endpoint and account data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Model Context Protocol: Tools](https://modelcontextprotocol.io/specification/latest/server/tools) - [Model Context Protocol: Security best practices](https://modelcontextprotocol.io/specification/latest/basic/security_best_practices) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [What is MCP for crypto wallet analysis?](https://walletfollow.ai/learn/mcp-for-crypto-wallet-analysis) - [Read-only AI prompts for wallet research](https://walletfollow.ai/learn/safe-ai-wallet-research-prompts) - [AI wallet analysis versus trade execution](https://walletfollow.ai/learn/ai-analysis-vs-trade-execution) --- # A practical copy portfolio review checklist Canonical: https://walletfollow.ai/blog/a-copy-portfolio-review-checklist Publisher: WalletFollow Research Published: 2026-09-29 Editorial update: 2026-09-29 Topic: Copy trading Review a copy portfolio by comparing its intended rules with the account’s actual exposure and execution. Check leader overlap, performance definitions, costs, data gaps and control outcomes. Keep a dated record so a change in the portfolio can be distinguished from a change in the evidence. ## Start with the version you are reviewing Record the portfolio definition, selected leaders, allocation rules and review period. Note any changes since the previous review. A newly published definition and an existing subscription can represent different configurations; confirm the one relevant to your account before comparing results. Do not silently combine a backtest with tracked or actually executed performance. Separate configuration from completion. A request sent to the engine is not proof that every intended position changed. Read the current status and available execution evidence. Where copying is unavailable, this checklist still supports research, but it does not turn a portfolio definition into a live track record. ## Check the portfolio as one account Review total equity, available collateral, gross exposure and market concentration on a consistent account basis. Include manual positions and other subscriptions where they share the same account. Evaluating each copied leader in isolation can miss the combined risk that the account actually carries. Look for overlapping leaders. Several wallets can independently hold the same directional position, making a portfolio less diversified than its leader count suggests. Hypothetical example: three separate leaders all hold BTC longs. Following all three adds three sources of decisions, but may still concentrate exposure in the same market move. ## Reconcile outcomes with execution and costs Compare the same period and performance basis. Record deposits and withdrawals separately from gains or losses. Inspect whether open positions, venue fees and funding are included in the result. Review builder and creator fees according to the published terms; do not add a fee twice if it already appears inside a reported cost. Investigate material follower-versus-leader differences through eligible markets, size limits, timing, slippage, rejected requests and available collateral. Keep missing evidence explicit. A lower follower return is a reason to understand the difference, not proof of one particular failure without inspecting what happened. ## Review controls and record the decision Read the current pause, stop and close explanations before requesting a change. Know whether an action affects future copying, existing positions or both. Confirm the resulting state after the request rather than assuming the button click achieved the intended outcome. Review permissions through the approved wallet workflow; a portfolio review never requires sharing a private key. End with observed changes, unresolved issues and the reason for the next review. Check again after material allocation changes, unusual exposure, execution problems or a coverage change. Preserve prior decisions rather than rewriting them to fit later returns. The checklist supports an informed review; it does not certify a portfolio as safe or profitable. ## Sources and further reading - [Hyperliquid: Info endpoint and account data](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/info-endpoint) - [Hyperliquid: Margining](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/margining) - [Hyperliquid: Funding](https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding) Examples are educational; they are not recommendations or forecasts. Venue rules and product availability can change. ## Continue reading - [How to assess diversification in a copy portfolio](https://walletfollow.ai/learn/diversifying-a-copy-portfolio) - [Copy-portfolio backtests, tracked results and live returns](https://walletfollow.ai/learn/copy-portfolio-backtests-and-tracked-results) - [Pause, stop and close: copy-trading controls explained](https://walletfollow.ai/learn/pause-stop-and-close-copy-trading) - [Copy-trading fees: exchange, builder, funding and creator costs](https://walletfollow.ai/learn/copy-trading-fees)