Example
A real test,
start to finish.
This is real output from the live engine, unedited. It isn't a winning strategy, and that's the point: you see what an idea actually did after costs.
You
"Backtest a 30-minute opening range breakout on NQ, 5-minute bars, 60-tick stop, 120-tick target, one trade a day, 2020 to now."
Your AI · ran it in 0.9s on 471,310 bars
trades 815
win rate 35.5%
profit factor 1.04
net P&L +$6,655
max drawdown -$12,932
sharpe 0.21
exits 524 stop · 282 target · 9 time
verdict barely breakeven after costs — the stop is too tight for NQ's noise
saved vault · run #614f · 1 credit
Then iterate
From here you'd ask your AI to change one thing — a wider stop, a time filter, a different session — and compare the versions side by side. Every version stays in your vault, so "what was my best Sharpe this month" has an answer.
Reading the numbers
| Metric | What it tells you |
|---|---|
| Trades | How many times the rules fired. More trades means a more reliable statistic. |
| Win rate | Share of trades that made money. Low is fine if winners are much bigger than losers. |
| Profit factor | Gross profit divided by gross loss. 1.0 is breakeven; 1.04 is barely above it. |
| Net P&L | Total result after slippage and commission. |
| Max drawdown | The deepest drop from a peak in the equity curve — what you'd have had to sit through. |
| Sharpe | Return per unit of risk. Under 1 is weak. |
Hypothetical results: a simulation on historical data, not real trading. Past performance does not predict future results.
Every plan starts with 5 free backtests. $0 today, plans from $19 a month, cancel anytime.
Try 5 backtests free