PnLSonar beta VI

Methodology (score v0.3.6)

Data sources

Every metric is computed from the Hyperliquid Info API: userFillsByTime, portfolio, userNonFundingLedgerUpdates, userFunding, clearinghouseState. The candidate list comes from Hyperliquid's public leaderboard (not part of the official API documentation), but none of its figures are used.

ROI

Time-weighted return: the window is split at the points of the PnL series, each segment's return = PnL generated / account value at the start of the segment, then compounded. Deposits and withdrawals change account value but not PnL, so they are never counted as profit or loss. Max drawdown is computed on this index.

Trade (episode)

A trade starts when the position in a coin leaves 0 and ends when it returns to 0; a flip counts as closing the old trade and opening a new one. Trade PnL = Σ closedPnl − Σ fees within the trade. Trades opened before the start of the data window are not counted as wins or losses.

Estimated leverage

Position notional after each fill divided by the most recent prior account value. This is an estimate.

The "Passed minimum checks" gates (not a recommendation to copy; no wallet is recommended before paper results exist) Score (0–100)

Return 25%, drawdown 25%, consistency 15%, profit factor 15%, sample size 10%, dispersion of wins 10%. The weights are a hypothesis, not yet validated against real copier results. The test: 30 days of paper copy, comparing copier results after costs for the high-score group and the Top PnL group. Each group has only 11 wallets over one month, so if the high-score group does not come out ahead, that means there is no evidence yet that the score is better in that run; it is not enough to conclude the score is wrong.

Paper copy assumptions

Source price ± 5 bp adverse (stands in for ~1500 ms latency), taker fee 0.045%, builder fee 0.05%, funding booked hourly (− position × 1-hour candle open price × public funding rate).

Beta. Past performance of source wallets does not guarantee future results. Source traders are public wallets; they are not partners and do not endorse this product. Paper results are simulated, not actual profit.