Quick test
Most precise13months
of real tick data, the most recent.
Filled at the real bid and ask, at whatever latency matches your setup.
Describe your strategy in plain English. Claude turns it into a frozen spec you approve, and our engine tests it on a year of real tick data or 16 years of history, with a held-out split and a luck test, and gives you an honest verdict. No code. No guru. No trust-me.
Free account. Tests from $9.99. Fair warning: most ideas don't survive. That's the point.
Three example ideas, replayed to show the stages. The numbers in this demo are illustrative, not live results. Real runs happen in the app, and most ideas fail.
Illustration · click an idea to replay it
13months
of real tick data, the most recent.
Filled at the real bid and ask, at whatever latency matches your setup.
16years
2010 to mid-2026. Pick any range.
Priced conservatively, so results lean to the safe side.
Spread and commission on every trade.
It must still work on days it never saw.
Your entry times on 200 other days. Beat 95%.
With reasons. Most ideas fail.
Every day after the close, the desk replays your frozen strategy on that day's data, posts the result, and flags it if it drifts outside the backtest's cone. The figures above are illustrative.
FEE MATH: $85–249 PER EVAL ATTEMPT — TOPSTEP $85/MO · MFFU $157/MO · APEX $249 ONE-TIME (50K LIST PRICES, READ FROM THE FIRMS' OWN PAGES 15 AUG 2026) · RESETS EXCLUDED: THEIR PUBLISHED SOURCES DISAGREE · SLIDE TO YOUR HONEST NUMBER
Run the numbers yourself, free
Set your own target and drawdown, or start from the Apex, TopStep and MFFU 50K rules, and replay 10,000 simulated evaluation attempts. Signed out it uses one fixed example series; sign in to test your own finished runs. No account or email needed to try it.
A few minutes of conversation. A year of tick data or 16 years of history. An honest answer, and the evidence behind it, so the next decision is yours to make.
Test your idea Compare plans — from $9.99
Runs automatically on our own server · verdicts usually in minutes · most ideas fail
Make a free account first. You pay only when you test an idea.
Both plansDescribe your idea in plain English and approve the spec. Our engine runs it on our server, usually in minutes.
Real CME dataA year of ticks or 16 years of 1-minute bars Real costsSpread and commission on every trade Held-out checkIt must work on days it never saw Luck testYour entries on 200 other days Both start with a free account. You pay when you test an idea. Hypothetical results, not investment advice. Most ideas fail.