TEST A TRADING IDEA
Turn a hunch into rules you can test.
Say it your way, or type it exactly. Each phrase becomes a rule you can read, then it runs on real history.
Example
Say it your way
Buy good tech companies when they dip, get out if the trend breaks.
1 · Say it
“when they dip”
Pulls back to its 20‑day average, still above its 50‑day
“the trend breaks”
Closes below its 50‑day average
“good tech companies”
Return on capital above 15% — here, MSFT
2 · See what it understood
MSFT · Oct 2021 – Oct 2026Fell less than just holding
−19.5% worst drop · these rules−37.1% worst drop · just holding MSFT
The trade-off: it made less — +22.8% vs +84.8% for just holding.
Real test on MSFT daily closes, no fees · 21 trades.
3 · Tested against just holding
See exactly what it understood.
Every phrase you used is matched to the rule it became. Anything it couldn't turn into a rule is said out loud — with a one-click fix.
What you said, and the rule it became
“when they dip”Pulls back to its 20-day average
“in an uptrend”Only while above its 200-day average
“not too pricey”Not in the rules yet
Add a valuation limitChange one number.
Test again.
Every number in your rules is editable in place. Results always say which version of your rules they came from.
Sell when SPY closes below its average of 50 days.
Every setting fell far less than just holding — and made less. Pick the trade-off you can live with.
- Worst drop · Your rules
- −10.4%Just holding SPY −33.7%
- Return
- Return +9.6% · Just holding SPY +317.9%
Or start from one that's known.
All 56 systems →How we test
- Real prices
- Daily prices from Financial Modeling Prep. Financials and news from licensed data providers.
- No look-ahead
- Each day of a test only sees what was known by that day's close. A signal trades at the next session's open.
- Costs included
- Backtests on your board charge 0.05% commission and 0.05% slippage on every buy and sell. The quick replays on this site sell at the close and leave fees out.
- Limits, said out loud
- Past results don't predict future ones. A test with only a few trades is mostly noise — the result says so.
Your first system is free.
No code. Real prices. Your system watches and emails you the moment it signals.