A quiet stretch of NSE charts, a handful of winners, and the mind closes the case: “strategy works.” That is not testing. That is storytelling with a short cast list.
The useful question is not a single magic number. It is which question you are asking—because edge testing and personal fit testing need different samples, different logging, and different humility.
Two tests people keep collapsing into one
Statistical edge testing asks: under fixed rules, across varied regimes, is the observed expectancy distinguishable from noise after costs? That is a research problem. Trade count alone never finishes it. Regime diversity, rule consistency, dependence between trades, and how the sample was collected matter as much as N.
Personal setup-fit testing asks: under uncertainty, do you follow a defined process for this setup—or do you chase, freeze, and rewrite rules mid-trade? That is a practice problem. A few dozen decisions can already expose process failure. A few hundred make patterns harder to dismiss as “bad luck that day.”
Retail traders usually need the second answer before they can honestly use the first. If you cannot execute the rule set, a pretty backtest is decoration.
Why small samples feel so convincing
After 15–40 trades, win rate and “feel” swing wildly. One oversized winner can dress a weak process as genius. A cluster of false breaks in the first hour of the cash session can bury a workable continuation idea that simply needed different conditions.
Online guides often float thresholds—roughly thirty as a crude checkpoint, hundreds as a more serious look. Treat those as reminders that variance is large, not as gospel formulas. Exact requirements depend on payoff distribution, how correlated your trades are, and how confident you need to be before risking live capital. Inventing a single N as proof is marketing, not statistics.
A practical frame: three ladders, not one number
| Ladder | Question | What “enough” roughly means |
|---|---|---|
| Executable? | Can I state entry, invalidation, and skip rules without improvising? | A short block of consistent decisions—enough to see rule chaos |
| Fit for me? | Which setup tag shows cleaner plan-adherence under hidden-outcome practice? | On the order of a focused month of decisions (300-trade practice experiment)—useful for self-observation, not edge proof |
| Edge plausible? | Across regimes, with frozen rules, do results stay coherent after costs? | Larger, regime-diverse samples plus out-of-sample discipline—far beyond a lucky week |
Three hundred practice decisions sit on the middle rung. They help you notice whether breakouts make you chase while pullbacks feel manageable—or the reverse. They do not certify that a discretionary style has a durable edge on NSE names or index futures.
What to hold constant while you accumulate N
- One primary timeframe and a short liquid universe—not a rotating scavenger hunt.
- Frozen definitions for setup tags; “other” should be rare, not a dumping ground.
- Outcomes in R-multiples so ego P&L does not rewrite the log.
- Process fields: planned invalidation, followed plan Y/N, regime note (trend / range / event-ish day).
- Anti-hindsight practice: candle-by-candle or hidden-date replay so you are not grading a finished story. Pair this with a disciplined manual backtest walk when you are testing rules rather than nerves.
Indian practice note
If you only sample the calm middle of a cash session, your N is biased. Open volatility, midday chop, and the last hour are different decision environments. Spread practice days across those shapes—still without turning any name into a tip. Diversity of conditions beats a larger N of identical lunch-hour ranges.
A decision rule you can actually use
- Under ~30 consistent trades: you may judge executability. You may not declare edge.
- Through a few hundred tagged practice decisions: judge setup-fit and rule adherence. Still withhold “proven strategy” language.
- Only with frozen rules, multi-regime coverage, and honest costs: start talking about whether an edge might exist—and even then, live behaviour can diverge.
When you are ready to choose what to practise next, use fit criteria—not a viral win-rate screenshot. Start with which trading strategy suits you.
Bottom line
Stop asking for one number that ends the argument. Ask which ladder you are on. Small samples sell confidence. Larger, cleaner samples buy clarity—and clarity still is not a guarantee of profits. Securities markets involve risk; practice is education, not a bypass.
Educational content only. Not investment advice. Not a recommendation to buy or sell any security. See SEBI investor education on key risks.