Direct Answer
Sample size determines how confidently any conclusion can be drawn from data. In gambling, sample size is the difference between knowing whether you have an edge and assuming you do based on noise.
Key Takeaways
- Hundreds of bets minimum for a directional read.
- CLV reaches confidence faster than win-rate.
- Small-sample conclusions are almost always wrong.
How much is enough
For binary outcomes (win/loss), a few hundred bets is the minimum for a directional read; 1,000+ for confident magnitude. CLV cuts this dramatically because each bet contributes signal proportional to the line move, not just a binary.
The cost of small samples
Decisions made on 20–50 bets are essentially decisions made on guesses. Stake sizing, market selection, even continuing to bet at all — none of these should be guided by sub-sample results.
Frequently asked questions
Can I evaluate a model on backtested data?+
Only with strict out-of-sample and walk-forward validation. In-sample fit is meaningless.
Educational only. Not wagering, financial, or legal advice. See our editorial policy.