Which check matters most?

Costs and the best-days test catch the most strategies. The luck test adjusted for versions tried catches the rest.

Can a strategy pass all ten and still fail?

Yes. Passing improves the odds a lot but cannot guarantee the future. That is why the last step is always a 30-trade paper test.

Can I use this checklist for signals or courses I'm sold?

Yes - ask the seller for the full trade list and the number of versions they tried. If they won't share it, that tells you something.

Do you check the strategy for me?

Yes - send the trade list and you get all ten checks plus a trade plan in a plain-English PDF.

Checklist

Backtest checklist: 10 checks before real money

Run these ten checks on any backtest - yours, a course's, or a seller's. If it fails two or more, paper-trade before you risk money.

Ask Tess about this

Enough trades. 100+ is good, 30 is the minimum to say anything.
Profit factor after costs. Money won ÷ money lost, after real fees. 1.3+ is healthy.
Cost stress. Add 0.2% per trade. Still above about 1.1?
Without the best 5 trades. Still profitable?
Without the best 5 days. Still profitable? (Catches strategies carried by crash or news days.)
Both halves of time. Profitable in each half?
Month by month. How many losing months in a row could you live through?
Luck test. Could random wins and losses of the same size do this well? Adjust for how many versions you tried.
Worst fall and losing streak. Size so that twice the worst fall would not stop you.
Just the market? Did simply holding the same market do as well?

Why bother: on thousands of past strategies, the ones that passed checks like these kept making money in data they had never seen about 7 times in 10, versus about 3.5 in 10 for the average backtest (our own tests; hypothetical, past data). Good odds - not a promise.

Special cases

Indicator strategies that are always in the market (stop-and-reverse, moving-average crosses) are hard to judge from a backtest at all - in our study their statistics barely beat a coin flip at predicting the next two years. Crypto backtests also predicted poorly on 57 coins we tested. For both, paper trading is the real test.

Frequently asked questions

Which check matters most?: Costs and the best-days test catch the most strategies. The luck test adjusted for versions tried catches the rest.
Can a strategy pass all ten and still fail?: Yes. Passing improves the odds a lot but cannot guarantee the future. That is why the last step is always a 30-trade paper test.
Can I use this checklist for signals or courses I'm sold?: Yes - ask the seller for the full trade list and the number of versions they tried. If they won't share it, that tells you something.
Do you check the strategy for me?: Yes - send the trade list and you get all ten checks plus a trade plan in a plain-English PDF.

Want your own strategy checked?

Send your trade list. You get a plain-English PDF: verdict, 10 checks, a position size for your chosen risk, a daily stop and a 30-trade paper-test plan. Free in the beta: run it yourself in the Workspace.

Run a strategy review

Keep reading

Why your trading strategy isn't workingThe 6 usual reasons a strategy that looked great stops making money.How to backtest on TradingView properlySettings that make the Strategy Tester honest: costs, fills, repainting, sample size.Is my backtest overfitted?Five simple tests that separate a rule that held up from tuned noise.Prop firm challenge: choosing a position size for your risk limitHow to size so one bad day doesn't end the challenge.

All guides →