The bet: ride the trend (breakout).
Flat-to-negative: pooled across all 232 out-of-sample trades the trade t-statistic is -1.50, the average trade loses 0.86%, win rate 28%. No coin clears even its most lenient bar. The only individually significant result is AVAX — and it is a loser (t -3.27); the two positive coins are noise on tiny samples. A uniform positive in-sample-to-out-of-sample gap marks it as the same mild-overfit mirage.
Donchian N-bar-high breakout, long-only, behind a permissive not-chop regime gate, with a chandelier ATR trailing stop — a different entry from trend_follow’s MA-cross but the same persistence bet. It is our second independent momentum construction to come back flat-to-negative, which is what turns "no persistence premium after cost" from a one-off into a confirmed finding.
| Pair | Sharpe | PF | trade t | n |
|---|---|---|---|---|
| ADA | -0.03 | 0.87 | -0.29 | 26 ⚠ |
| XRP | -0.05 | 0.84 | -0.40 | 37 |
| TON | +0.03 | 1.22 | +0.21 | 23 ⚠ |
| SOL | -0.08 | 0.84 | -0.34 | 28 ⚠ |
| AVAX | -0.67 | 0.24 | -3.27 | 28 ⚠ |
| LINK | -0.33 | 0.60 | -1.17 | 33 |
| BTC | -0.36 | 0.49 | -1.58 | 31 |
| ETH | +0.14 | 1.32 | +0.43 | 26 ⚠ |
Pooled across 232 out-of-sample trades: trade t -1.50, mean -0.86%/trade, 28% win, equal-weight mean Sharpe -0.17.
⚠ = underpowered (fewer than 30 out-of-sample trades). Negative values in red.
Flat-to-negative — a second momentum idea, same dead end.
Every trade we place — wins and losses, from real money — is on the live track record. This page is the research that decides what we will and won't run.
See the live track record →