The Honest Backtest Census: How Often Famous Strategies Beat Buy & Hold
How to cite this release
The headline is the stocks-&-ETFs cut (composition-stable). Crypto beats buy & hold far more often on this set, so any all-asset figure runs higher — an overfitting artifact, not an edge.
“As of September 25, 2026, 25 famous, named trading strategies were each backtested on 36 liquid US stocks & ETFs — 900 strategy-asset combinations, of which 839 had enough trades to judge (61 were too thin and excluded). Of those judged, about 86% failed to beat simply buying and holding the same asset over the same 5-year window — net of costs and slippage, with no look-ahead. This is a survivorship-biased set of well-known assets, measured over the full period with no out-of-sample split, against each asset's own buy & hold, and is not significance-tested.”
Permanent DOI: 10.5281/zenodo.22974116 — archived on Zenodo, CC BY 4.0. Cite this release by the DOI or the URL above.
Download the data — free, CC BY 4.0
Reuse freely with attribution. Aggregate rates only — no per-trade records, equity curves, or strategy parameters (see the README for what is and isn't here).
By asset class
| Class | Judged | Beat | Failed |
|---|---|---|---|
| Stocks & ETFs | 839 | 14% | 86% |
| Crypto (context) | 257 | 64% | 36% |
| All (context) | 1,096 | 26% | 74% |
The stocks & ETFs row is the headline. Crypto and “all” are context only — a small, volatile, survivorship-biased sample; an overfitting artifact, not a demonstrated edge, and not significance-tested.
By strategy — all 25, on stocks & ETFs
Every one of the 25 famous strategies has a negative median excess — the typical run of each lagged simply holding. The median is a cross-asset central tendency, not a forecast; differences between strategies are within sampling noise on a small, correlated, survivorship-biased set.
| Strategy | Judged | Failed | Median excess |
|---|---|---|---|
| 200-Day MA Regime | 36 | 69% | -22% |
| 52-Week High Breakout | 30 | 93% | -77% |
| ADX / DMI Trend | 36 | 83% | -81% |
| Aroon | 36 | 86% | -49% |
| Awesome Oscillator | 36 | 86% | -54% |
| Bollinger Band Breakout | 36 | 86% | -69% |
| Bollinger Band Reversion | 36 | 92% | -70% |
| Commodity Channel Index | 36 | 89% | -66% |
| Chaikin Money Flow | 36 | 92% | -80% |
| Connors RSI-2 | 36 | 89% | -73% |
| Donchian Breakout | 36 | 86% | -54% |
| Heikin-Ashi Trend | 36 | 92% | -84% |
| Ichimoku Cloud | 36 | 89% | -64% |
| Keltner Channel Breakout | 36 | 83% | -75% |
| MACD | 36 | 86% | -75% |
| Money Flow Index | 30 | 80% | -46% |
| Time-Series Momentum | 36 | 75% | -36% |
| Moving Average Crossover | 23 | 91% | -42% |
| On-Balance Volume | 36 | 81% | -69% |
| Parabolic SAR | 36 | 89% | -84% |
| RSI Reversion | 36 | 86% | -77% |
| Stochastic RSI | 36 | 86% | -47% |
| Stochastic Oscillator | 36 | 83% | -65% |
| SuperTrendsmall sample | 0 | — | — |
| Vortex Indicator | 36 | 92% | -73% |
Drawdowns — strategy vs buy & hold
Strategies often drew down less than holding — but that is mostly sitting in cash(avoidance, not skill), and on stocks & ETFs they still failed 86% of the time. Median maximum drawdown across judged combinations (a negative fraction of equity):
Overfitting artifact on a small survivorship-biased sample, not an edge.
Overfitting artifact on a small survivorship-biased sample, not an edge.
About this release
- As of
- September 25, 2026
- DOI
- 10.5281/zenodo.22974116
- Window
- 2021-09-26 → 2026-09-25 (5 years)
- Sample
- 25 strategies × 36 liquid US stocks & ETFs = 900 combinations; 839 judged, 61 too thin
- Benchmark
- each asset's own buy & hold
- Costs
- 0.05% per side, next-bar fills, costs and slippage on, no look-ahead
- Licence
- CC BY 4.0
- Release / version
- 2026-09-25 / v1