Key numbers
| Measure | Value |
|---|---|
| Backtests | 17,167 |
| Distinct strategies | 5,917 |
| Symbols | 1 |
| Median Sharpe | −0.94 |
| 10th to 90th percentile | −2.51 to +0.05 |
| Beat cash (Sharpe above 0) | 11.7% |
| Sharpe above 1 | 0.16% |
| Beat buy-and-hold | 1.7% |
| Median trades | 306 |
| Of those with positive returns: still positive after 2023 | 53% |
By timeframe and rule
| Timeframe | Rule family | Backtests | Median Sharpe | Beat cash | Beat buy-and-hold |
|---|---|---|---|---|---|
| 15-minute | Donchian breakout rules | 1715 | −1.29 | 0% | 0% |
| 15-minute | MACD crossover rules | 1627 | −2.20 | 0% | 0% |
| 15-minute | RSI zone rules | 1723 | −2.10 | 0% | 0% |
| 15-minute | Moving-average crossover rules | 1664 | −0.43 | 18% | 1% |
| 15-minute | Bollinger squeeze rules | 1679 | −1.79 | 0% | 0% |
| 1-hour | Donchian breakout rules | 1703 | −0.42 | 13% | 0% |
| 1-hour | MACD crossover rules | 1631 | −0.13 | 32% | 1% |
| 1-hour | RSI zone rules | 1815 | −1.07 | 0% | 0% |
| 1-hour | Moving-average crossover rules | 1609 | −0.09 | 42% | 12% |
| 1-hour | Bollinger squeeze rules | 1672 | −0.62 | 6% | 0% |
| 4-hour | Donchian breakout rules | 40 | +0.68 | 95% | 57% |
| 4-hour | MACD crossover rules | 38 | −0.32 | 5% | 0% |
| 4-hour | RSI zone rules | 49 | −0.41 | 10% | 0% |
| 4-hour | Moving-average crossover rules | 33 | +0.43 | 79% | 36% |
| 4-hour | Bollinger squeeze rules | 39 | +0.14 | 77% | 0% |
| daily | Donchian breakout rules | 42 | +0.76 | 93% | 69% |
| daily | MACD crossover rules | 36 | +0.20 | 75% | 11% |
| daily | RSI zone rules | 38 | −0.01 | 50% | 0% |
| daily | Bollinger squeeze rules | 13 | +0.37 | 100% | 8% |
By filter family
| Filter family | Backtests | Median Sharpe | Beat cash |
|---|---|---|---|
| liquidity (money supply, balance sheets) | 2,286 | −0.78 | 11.2% |
| market data (VIX, indices, commodities) | 4,950 | −0.87 | 15.8% |
| crypto market data | 1,320 | −0.92 | 12.4% |
| interest rates and yield curve | 2,647 | −0.93 | 12.2% |
| sentiment and surveys | 2,883 | −1.06 | 8.3% |
| macro releases (CPI, PMI, payrolls ...) | 3,081 | −1.09 | 8.2% |
How to use this
Before you backtest a technical rule on BTCUSD, look up the median for that rule and timeframe. A result far above it after many tries is more likely luck than an edge. Your agent can run the same check on its own result with the check_backtest tool.
Caveats
In-sample, not validated on new data: full-history backtests after retail costs, Sharpe minus a 4% cash rate. "Still positive after 2023" only counts variants with positive returns over the full history, which includes 2023–2026, and 2023–2026 were rising markets, so read it as an upper bound. Monthly macro filters were applied 1–3 weeks before publication in these backtests, which flatters them. Some price files had data errors.
Terms used: in-sample upper bound, trial count, dead end.
Let your AI check how ideas like yours did before it runs a single test. In our pilot, agents that checked first spent 1% of their backtests on known dead ends instead of 15%, and every backtest not wasted is time and tokens you keep. Cite this entry as ER-LIB-BTCUSD, as of 2026-11-01.
Related evidence
Questions
Do technical indicators work on BTCUSD?
BTCUSD: 17,167 backtests of RSI, MACD, Donchian, squeeze and moving-average rules with macro filters, full history after costs. 11.7% beat cash and 1.7% beat buy-and-hold; median Sharpe −0.94. The strongest combination by median was Donchian breakout rules on daily bars (+0.76).
Is this investment advice?
No. Research statistics only.