A public wallet can show what an account did, but it does not automatically show how well a strategy performed. The record must first be organized into positions, cash flows, and returns. Even then, the address may reveal only one part of a larger portfolio.
The most useful starting question is specific: did this account increase a position, reduce it, or transfer capital? A feed of large transactions can make all three look like the same kind of signal.
From fills to a position
Suppose an account is short 100 units and buys 40. It is now short 60. Another purchase of 70 closes the remaining short and establishes a long of 10. The purchases total 110 units, but only the last ten become new long exposure.
Orders, fills, and account states answer different questions. An order is an instruction that might be canceled. A fill changes a position. A snapshot shows the resulting exposure. Hyperliquid’s public API separates these records and warns that incomplete pagination or querying an agent address instead of the account address can give an incomplete picture.
Keep quantity separate from dollar value. If an unchanged 100-unit position rises from $10 to $12 per unit, its displayed exposure grows from $1,000 to $1,200 without any purchase. An account can look more committed simply because its earlier position appreciated.
Deposits and investment results
An account starts with $10,000, receives a $5,000 deposit, and ends with $16,000. The net increase attributable to everything other than that deposit is $1,000, before resolving any omitted flows or charges. Calling the entire $6,000 increase trading profit would be wrong.
A percentage return also depends on when the deposit arrived. Imagine the account first grows from $10,000 to $11,000, then receives $5,000, then loses 10%. It ends at $14,400. Its net dollar result is $14,400 − $10,000 − $5,000 = −$600. The two trading periods returned +10% and −10%, which compound to −1%.
Those numbers answer different questions. The time-weighted result links the periods without allowing a deposit to masquerade as performance. The dollar result reflects how much capital was present for each move. The CFA Institute’s performance-evaluation discussion explains why the choice matters when comparing managers or accounts.
Capital behind the return
A $50,000 gain could come from a large, modestly fluctuating portfolio or a small account that repeatedly approached liquidation. Peak exposure, drawdown, margin additions, funding, and unrealized losses help distinguish those paths. Closed-trade win rates are especially incomplete if losing positions remain open.
Look at comparable episodes. Did the trader cut exposure when evidence changed, or repeatedly add collateral until prices recovered? How much capital would a follower have needed during the worst interval? A later recovery does not erase the funding obligation that came before it.
Selection matters too. A leaderboard shows the addresses that meet its ranking rules. The winner of thousands of high-risk attempts is interesting, but choosing that winner afterward is different from identifying it before the gains occurred. Keep the observation start date and selection rule explicit when following an account over time.
What the address cannot reveal
An unseen hedge can change the meaning of a visible short. The controller’s identity and information advantage are usually unknown. A defensible note is therefore narrow: “This address reduced its short after the announcement; its other holdings are unknown.”
Use the activity to generate a question about a market, then check that question against independent evidence. Your available entry price, costs, and collateral differ from those of the address you observed. The address can show that someone changed a position. Whether the same trade is attractive at your available price remains a separate question.