Historical Signal Backtesting
Replay any date range through the WHaiLE conviction engine. Every signal that fired, graded Big Win → Win → Flat → Loss on where the price closed 7 days later, and against holding ETH over the same days.
Overall performance across all signal families and the selected date range. Hit rate = signals whose price was up more than 1% at the 7-day mark (Big Win = up more than 10%). Median 7d = middle value of all 7-day returns — less skewed by outliers than the average. Same rule as the Picks scoreboard.
By signal family
— · —How each WHaiLE conviction signal type performed in the selected window. N = total fires. Hit rate = % of signals up more than 1% at the 7-day mark. Median 7d / Avg 7d = where the price closed 7 days later. Beat ETH = share of fires that did better than holding ETH over the same 7 days (hover for the median gap). BW · W · F · L = count breakdown by verdict. Best = the single best 7-day result in the family.
| Signal | Thesis | N | Hit rate | Median 7d | Avg 7d | Beat ETH | BW · W · F · L | Best |
|---|
Audit trail · individual fires
—Every individual signal that fired in the selected window, one row per fire. Entry = median derived price in the ±2h around the signal. Day 7 = median price in the ±12h around day 7. Δ% = (day 7 ÷ entry − 1) × 100.
| Time | Signal | Token | Entry | Day 7 | Δ% | Verdict |
|---|
