01
Sample adequacy
Is 41 trades evidence, or is it noise?
Nearly every retail backtest is decided on too few trades. We work out how many you actually need for your win rate, and tell you how far short you are.
For TradingView strategy builders
It shows you a profit factor of 3.4 and says nothing about the fact that it came from 41 trades, ignores your spread, and only worked in 2021. Paste your results. Get a grade, and the reasons why.
Strategy Tester → List of Trades → the download icon. That's the file.
Your recent reports
Kept in this browser only — not an account, and never sent to us.
The same export, read two ways
Both sides are generated from one TradingView export — the strategy in the example below. Nothing here is illustrative.
Strategy Tester shows you
All true, and no answer to the only question that matters.
Assay tells you
D Sample adequacy is the problem. Everything else stands up, but that one failure is enough to stop you trading this until it is understood. Be aware that the five best trades account for 51.1% of gross profit — remove them and very little remains.
A verdict, and the number behind every part of it.
Example — what comes back
A real assay of a real strategy, not a mock-up. Every figure below came out of its TradingView export.
SPY · 68 trades · 18 Aug 2025 – 13 Aug 2026
Sample adequacy is the problem. Everything else stands up, but that one failure is enough to stop you trading this until it is understood. Be aware that the five best trades account for 51.1% of gross profit — remove them and very little remains.
Below 100 trades, results are dominated by luck rather than edge. Whatever this strategy's profit factor looks like, 68 trades cannot tell you whether it is real. Widen the date range or move to a lower timeframe until you have a few hundred.
At a reward-to-risk of 2.83, this strategy needs to win 26.1% of the time simply to break even. It wins 38.2%. That is a genuine cushion.
A strategy winning 38.2% of the time, over 68 trades, should throw up a worst run of about 7 losses through chance alone. This one reached 6, which is within what ordinary bad luck produces. It ran from 17 Sep 2025 to 01 Oct 2025. It is not evidence of anything breaking.
The P&L in this file does not reconcile exactly with price movement times size, which means your strategy is modelling commission or slippage itself. The results above are therefore already net of costs. Beyond what you have modelled, the edge absorbs a further 0.73 USD per trade before profit factor reaches 1.00.
Tap any line for the reasoning behind it.
68
Trades
38.2%
Win rate
1.75
Profit factor
2.83
Payoff
0.73
Expectancy (USD)
6
Worst run
A report also gives you a permanent link and a scorecard image to share, and says plainly if anything was left out of the sample.
The four checks
01
Is 41 trades evidence, or is it noise?
Nearly every retail backtest is decided on too few trades. We work out how many you actually need for your win rate, and tell you how far short you are.
02
Your win rate is 38%. Good or catastrophic?
It depends entirely on your reward-to-risk, and almost nobody works it out. We compare your actual win rate against the rate you need just to break even.
03
You lost 13 in a row. Normal, or broken?
A long losing run can be perfectly ordinary for your win rate, or it can be the strategy failing. We tell you which, and when in the data it happened.
04
At what spread does this edge hit zero?
Most small edges die on costs. We find the exact spread and commission at which your profit factor drops to 1.00 — then you compare it to what your broker actually charges.