🧪 Walk-Forward Backtest Results
Out-of-sample evaluation across all DSE tickers — train 150 bars, test 50 bars, rolling forward.
Purpose: disprove edge, not confirm it. Positive results are hypotheses — not trading signals.
Last run: 2026-06-10 17:33 UTC
· 396 tickers · 2068 trades
⚠ Important:
A strategy that looks profitable in backtest loses money live more often than not (look-ahead bias, overfitting, market impact).
These results are shown to challenge assumptions, not to guide trade decisions.
All trades simulate entry at next bar's open after signal fires, with a round-trip cost of 1.6% (0.5% brokerage + 0.3% slippage, each way).
Rules with fewer than 10 out-of-sample trades are marked unreliable.
Backtest running…
This may take a few minutes. Page will update automatically.
| Rule |
Reliability |
Trades |
Hit Rate |
Avg Net Return |
Median Net |
Sharpe |
Max Drawdown |
Best / Worst |
Illiquid % |
Per-Window (net%) |
|