Does the Three-bar mean reversion strategy work?
We coded the three-bar mean reversion to exact mechanical rules and ran it on june 1, 2025 – june 1, 2026 (12 months) of real data with fees on. No discretion, losers published as losers. Here is the verified result.
Verified Result
Weak: the mechanical version barely cleared, or failed to clear, the fee hurdle.
| Market | TF | Trades | Win | PF | Max DD | Net |
|---|---|---|---|---|---|---|
| ETH | 1h | 620 | 41.0% | 0.91 | 44.0% | -31.6% |
| BTC | 1h | 608 | 39.1% | 0.77 | 62.2% | -60.5% |
| Gold (PAXG) | 1h | 558 | 37.8% | 0.58 | 82.8% | -82.3% |
| EUR (EURUSDT) | 1h | 441 | 35.6% | 0.35 | 96.0% | -96.0% |
How the SVS 26 breaks down ▾
12 months of real 1-minute data, fees on (0.05%/side), $10k start, 1% risk. How the score works →
The exact rules tested
- Buy after 3 consecutive red candles, sell short after 3 green; 1.5x ATR stop; 1.5R target
How to read this
This is the mechanical floor — what the rules alone produced after costs, with zero discretion. A low score doesn’t mean the idea is worthless; it’s the baseline your filtering and judgement have to beat. The SVS methodology explains exactly how the score is built.
Test your own version of this
Add a trend filter, a session window, or a confluence and see if you can beat the floor. Describe it in plain English and the AI backtester runs it on real data.
Open the AI backtester →Questions
Is the Three-bar mean reversion strategy profitable?
Weak: the mechanical version barely cleared, or failed to clear, the fee hurdle. Across 4 markets and 2,227 trades with fees on, it had a profit factor of 0.72 and an average net return of -67.6%.
What is the win rate of the Three-bar mean reversion strategy?
38.6% across all tested markets. Note that win rate alone is misleading — with a fixed reward-to-risk target, a mechanically negative strategy can still win 30–45% of the time.
How was the Three-bar mean reversion strategy tested?
Coded to exact mechanical rules and run on June 1, 2025 – June 1, 2026 (12 months) of real 1-minute data, fees on (0.05% per side), $10,000 start, 1% risk per trade, on Secuora's deterministic engine. No discretion.