Know Before You Risk.
Backtest any strategy against months of real market data — with accurate spread, slippage, and commission modeling. No hand-wavy results. Just honest numbers.
Run Your First BacktestHow Backtesting Works
ForgeAlpha backtests simulate real trading conditions — not the ideal ones.
Data Fetch
Historical OHLC candles are fetched and cached. Specify any symbol, timeframe, and date range — data reused across runs for speed.
4-Tick Simulation
Each candle replays as 4 ticks: Open → Low/High → High/Low → Close. SL/TP checks on every tick prevent false results from single-tick evaluation.
Results + Analysis
Equity curve, win rate, profit factor, max drawdown, total trades, and friction breakdown — instantly. Cost warnings flag strategies that don't survive real spreads.
3-Way Backtest Compare
Stop guessing which version is better. Run all three side-by-side — original, AI-improved, and your own edits.
| Metric | Original | AI Suggested | Your Version |
|---|---|---|---|
| Win Rate | 52% | 61% | 58% |
| Profit Factor | 1.31 | 1.87 | 1.64 |
| Max Drawdown | 18.4% | 9.2% | 12.1% |
| Total Trades | 489 | 312 | 371 |
| Net Return | +18.3% | +43.2% | +31.7% |
Real Costs. Real Results.
Most backtests ignore what kills real accounts. ForgeAlpha models every cost layer.
Spread
Bid/ask spread applied on every open and close. Configurable per symbol — match your broker's actual spread.
Commission
Per-lot commission charged on open. Set to match your broker (e.g. $7/lot round-trip) — no hidden assumptions.
Slippage
Fixed or random slippage on market orders. Random mode adds noise to prevent curve-fitting on exact entry points.
Margin Guard
Margin checked on every trade open with a 50% stop-out level — no trades open if margin is insufficient.
Stop guessing. Start knowing.
Run a backtest in minutes. See exactly how your strategy performs — before you risk anything.
Run Your First Backtest