How JoinEm ranks Hyperliquid traders
Data
Everything comes from Hyperliquid's public API: the leaderboard, each wallet's fills, funding payments and positions. We find wallets on the leaderboard and in the public trade stream, download up to a year of fills, and rebuild closed trades from them. Each day we also read every listed account's value and perp profit history and its open positions, so losses a trader hasn't taken yet are counted. A trader is listed only if they pass the copyable rules and their perp trading is up over 90 days. Rankings refresh hourly from the latest daily recalculation. Free pages show positions from the daily snapshot; Pro shows them live.
Who is left out
Market makers and high-frequency bots (hundreds of fills a day or mostly maker orders) can't be copied and never appear. Wallets with too few trades to judge, or whose profit came mostly from one or two trades, are typed Lucky and kept off the rankings.
Metrics
Does the ranking predict anything?
We check every ranking a month later: was the top half up on perp trading more often than the bottom half? The first live result appears here once a ranking is a month old. On past data (eight dates, two groups of traders) the top half was up about as often as the rest, so read the score as a description of a record, not a forecast.
Trader types
Each trader also gets a holding horizon, a market focus and a size class:
Names
Wallets are shown by address. A few well-known wallets carry a name that the trader uses publicly or that is widely attributed to them; we don't guess identities, and anyone can ask for a name to be removed.