Question · data as of 2026-10-30

Is this Hyperliquid trader skilled or just lucky?

Paste the address below. On 2026-10-30, of the 98 wallets in Hyperliquid’s all-time PnL top 100 with enough history to score, 9 would pass the usual test (t ≥ 2) if you had picked them blind, and 0 pass the bar that applies when you found them by sorting a board of 48,407 accounts (t ≥ 4.90). For the 30-day ROI top 100: 11 and 0 of 76.

What the check does

It takes the wallet’s public portfolio history from Hyperliquid, turns it into weekly returns (change in cumulative PnL over the account value at the start of the week, so deposits and withdrawals are not returns), and asks how strong that record is: t = annualised Sharpe × √years, adjusted for short and irregular records.

Which bar applies

If someone sent you this one wallet and you had no other candidates, t ≥ 2 is the usual bar. If you found it on a leaderboard, it is the best of 48,407, and someone is always on a streak in a crowd that size. The bar for the best of 48,407 is t ≥ 4.90 (Bonferroni). Most people find wallets on leaderboards. That bar counts every account as independent; accounts that trade alike are fewer independent bets, which would lower it to about 2.6 by our rough estimate (an order of magnitude, not a measurement).

What the verdict cannot see

A strategy that sells insurance can show no losing week for a long time. The page flags a record with no losing week, one where a single week made most of the gains, and one where most of the PnL was booked while the account held under $1,000 and cannot be turned into a return. In the PnL top 10 on 2026-10-30, 1 carry that last flag and 0 have too little measurable history to score at all.

From an agent: the MCP tool hyperliquid_wallet, or GET https://api.evidencedeflator.com/api/wallet?a=0x…, or the skill in the repository.

Is this Hyperliquid trader skilled or just lucky?

What the check does

Which bar applies

What the verdict cannot see

Related questions