Short answer: first choose the right main or subaccount address and check that both sides of the trades you want to judge are available. Then separate the result from price moves, actual trading fees or rebates, and funding paid or received. Only after a few closed trades match the account record should you compare setups or ask AI Coach what keeps working.
A profitable-looking dashboard can be misleading if an older opening fill has fallen out of the accessible history, a different subaccount made the trade, or a current open position is mixed with realized results. No attractive chart repairs a missing start to the trade.
Hyperliquid's entry-price and P&L documentation notes that displayed entry price and closed P&L are frontend conveniences; the underlying accounting rests on balances and trades. Use displayed numbers as a cross-check, but keep the events behind a conclusion.
First check whose trades and how much history you have
Begin with the public address of the account whose trades you want to review. Hyperliquid's own API guide warns that using an agent-wallet address instead of the master or subaccount address can return an empty result. A main account and its subaccounts can tell different stories; a testnet account should never be mixed with mainnet performance. Decide whether you are reviewing perps, Spot or both, and note the first and last trades you can actually see. The same official guide describes a retained fill window, so an older opening trade may no longer be available even when its closing trade is visible. In that case, “all-time return” is not a defensible label.
A quick completeness test is to choose one round trip you know well. Can you find the opening and closing executions, the costs that belong to them, and any funding paid or received while the position was open? If not, keep the result marked incomplete and find the missing account or earlier records before comparing strategies. A long list of recent trades is not proof of a complete history.
TSB's Hyperliquid route starts from a public address; it should not ask for a private key, trading permission or wallet signature just to read that history. Do not paste those credentials into a journal. A public address can still reveal sensitive trading behavior, so choose deliberately which account to review.
Perp result: do not count fees twice or forget funding
For a perpetual trade, keep price result, actual trading fees or rebates, and funding paid or received separate. Hyperliquid's funding documentation says payments occur hourly; either side can pay or receive depending on the rate and position. Funding is therefore not always a cost. Its fee documentation distinguishes spot and perp schedules and maker/taker or rebate conditions. Today's fee-tier badge is not a substitute for charges on your actual fills.
Here is illustrative arithmetic, not a live Hyperliquid or TSB account. Suppose one closed perp cycle has $45 gross price profit, -$0.90 total signed execution fees across opening and closing fills, and -$2.00 signed funding while held. The illustrated realized result is $42.10. If the funding was +$2.00 received instead, the analogous result is $46.10. The amount, direction, collateral currency and scope must come from actual records.
Before subtracting costs from a displayed P&L number, check what that number already includes. Hyperliquid's P&L explanation includes a closing fee in its displayed closed-trade figure; subtracting the same fee again makes the result look worse than it was. Funding is a separate payment, and it can be either a cost or a credit. Reconcile one familiar round trip with the actual executions and balance changes before applying the method to many trades.
If the opening trade is older than the available history, you may still see a closing P&L figure. Treat it as an observed partial result, not a fully verified after-cost round trip: the entry fee or earlier funding may be missing. You can still use the visible amount to investigate a discrepancy, but not to declare that one setup or holding period beats another.
Spot is a different accounting problem
Spot buys build an inventory basis and later sales realize a result; a deposit or transfer can introduce quantity without a buy inside the visible fill window. Hyperliquid's P&L documentation describes special display treatments for transfers and pre-existing balances. Those frontend conventions do not guarantee that a journal can recover your actual acquisition price from recent fills.
For a useful Spot review, match a sale to the quantity and cost of the coins you actually acquired. If some coins arrived by transfer or were bought before the available history begins, the sale price alone cannot tell you the profit. Fees paid in another token also need a reliable value in the currency you are comparing. Under portfolio margin, borrowing interest can add another cost; do not call a Spot result fully net of costs if that interest is missing.
If the visible sales exceed the buys you can account for, stop at the known portion. Do not invent a zero purchase price for the rest. Ask where the earlier acquisition or transfer record lives before making a performance claim.
| Question | Minimum evidence | Stop condition |
|---|---|---|
| Did this perp cycle make money after recorded costs? | Opening and closing fills, signed fees/rebates, funding during the held interval, same collateral currency | Missing opening fill, funding surface or fee sign |
| Did this Spot sale make money? | Complete acquisition basis for sold quantity, sale fills, fee token and quote conversion | Missing old buy/transfer basis or unresolved third-token fee |
| Is one setup better than another? | Reconciled closed cycles, setup tags created consistently, comparable market and period | Mixed networks, incomplete cycles or too few comparable examples |
| What happened across accounts? | Each correct master/subaccount address and environment mapped separately | A combined address label hiding different accounts |
What should a journal add to a venue history?
The venue can show the fills and position state. The journal's distinct job is to preserve a repeatable decision loop: group the economically complete trades, tag the plan actually followed, identify a recurring cost or behavior pattern, then test one change on later comparable trades. For example, if a long-held perp setup looks worse than short-lived entries, separate the price outcome from funding paid/received and from a changed market regime. “Avoid holding overnight” is not a conclusion until those alternatives are checked; Hyperliquid funding is hourly, not one universal overnight fee.
Use a few closed trades as a reality check before trusting a dashboard of hundreds. Include a trade with several entries or a partial close and, for perps, one that crossed a funding payment. If a required opening or cost is missing, the honest result is not verified, however convincing the chart looks.
TSB's supported-source directory lists Hyperliquid public-address sync and file upload; its product docs explain the address connection. After connecting, check that the chosen address and visible date range include a trade whose outcome you already know. The connection alone does not prove that older trades, every account or every cost is present.
Once the history is reconciled, TSB AI Coach can examine a selected portion of your own recorded trades for recurring setup or session patterns. It requires Full Access and should show the scope of the evidence behind an answer. A question such as “did longer holds lose their edge after funding?” is answerable only when the selected, comparable trades include the actual funding paid or received and complete opening/closing context. If those records are missing, the after-funding comparison is not verified—even if the price-only pattern looks convincing.
Ask the personal question only after the ledger is whole
“Why is my Hyperliquid P&L lower than I expected?” can mean four different things: a losing price move, fee drag, funding paid while a perp was open, or a dashboard/account-scope mismatch. It can also be impossible to answer from the retained provider window. A useful AI Coach review starts by declaring the exact address, network, market, dates and closed-cycle count. It then separates those recorded components rather than assigning the difference to “bad timing” by intuition. If the question is about a repeated setup, the trader's own setup label and planned holding rule must be attached to the matched cycles first.
An appropriate next question is: “On my verified closed perps in this period, did longer holds still outperform shorter holds after the funding actually paid or received?” Compare like-for-like market and setup groups, show how many complete cycles exist in each, and reference the fills and funding events supporting the answer. Do not use today's displayed funding rate as a substitute for the history. Do not turn a missing opening fill into a zero-cost entry, or use a Spot sale with unknown acquisition basis as evidence that the trader's setup works. If one group is tiny or one collateral token differs, the Coach should explain the limit and suggest what new comparable records would make the test possible.
That repeatable loop is the reason to use a persistent trading journal instead of asking a general chatbot to read a screenshot. The answer can be revisited when more complete trades arrive, with the earlier scope and caveats intact. A public-address connection is not proof that every historical transaction is available.
The decision at the end is not “Hyperliquid is good” or “this strategy is good.” It is narrower: given this verified address and retained period, did the specific trade pattern survive its observed execution, fees and funding—and what should be tested next?
Common Hyperliquid P&L questions
Can a public address show all of my Hyperliquid trading? It can identify the selected account's available public history, not automatically every master, subaccount, network or older fill. Check the first and last visible trades against records you already know. Never provide a private key to a journal for a read-only review.
Why is my closed P&L different from my balance change? The displayed closed figure and the account balance do not necessarily have the same scope. Funding, trading fees, open positions, deposits, withdrawals and transfers need separate treatment. Reconcile one closed cycle and its account events before forcing the totals to match.
Can I compare Spot and perp results in one chart? Only after each has a valid basis and cost record in a common reporting currency. A Spot sale without the older acquisition cost is not a verified profit, and a perp close without its earlier opening fill and funding is not a verified full-cycle result.
