Methodology
Sample-size floors
Every CricketStudio statistical claim must pass a sample-size floor before it is published. Below-floor data is suppressed — never silently surfaced as a clean statistic. This page lists the exact thresholds and the reasoning behind each one.
| Metric type | Floor | Applies to |
|---|---|---|
| Batting strike rate | ≥ 30 balls faced | Orange Cap · Strike rate leaderboards · Powerplay / death batting SR · Phase split claims |
| Batting average | ≥ 30 balls faced | Batting average leaderboards · Player profile P3/P4 claims |
| Bowling economy | ≥ 15 balls bowled | Purple Cap · Economy leaderboards · Powerplay / death bowling economy · Phase split claims |
| Bowling average / bowling strike rate | ≥ 90 balls bowled | Bowling average leaderboards · Bowling SR leaderboards · Phase bowling SR claims |
| Trend insights (venue / conditional) | ≥ 3 matches | /trends/venue-sig-* · /trends/cond-* · Venue hub stats |
| Team / season aggregates | ≥ 5 matches | Team profile aggregates · Season hub claims · Captain matchup pages |
Rationale
≥ 30 balls facedBelow 30 balls the luck component in runs-per-ball is too high to distinguish skill from variance. At 30+ balls the sample captures at least 5 overs of pressure context.
≥ 30 balls facedSame threshold as strike rate. Average additionally requires at least 1 dismissal — an unconditional not-out across a tiny sample can produce a misleadingly infinite average.
≥ 15 balls bowled15 balls ≈ 2.5 overs, enough to see the bowler face at least one powerplay or death-overs stint. Economy can swing wildly on a single over; 15 balls provides a minimum stabilising window.
≥ 90 balls bowledAverage and SR are wicket-rate metrics — they require enough balls to contain a statistically meaningful wicket count (typically ≥3 wickets). 90 balls ≈ 15 overs is the minimum for a credible average.
≥ 3 matchesTrend pages aggregate across fixtures. Three matches is the minimum to distinguish a pattern from a single outlier event.
≥ 5 matchesTeam-level patterns (home/away splits, toss effects, phase strengths) need at least 5 matches to suppress single-match anomalies.
FAQ
- What is the minimum sample size for a batting claim?
- CricketStudio requires ≥30 balls faced before publishing any batting strike rate or average claim. Sub-floor entries are suppressed from leaderboards and trend pages.
- What is the minimum sample size for a bowling claim?
- Economy rate requires ≥15 balls bowled. Bowling average and bowling strike rate require ≥90 balls bowled (roughly 15 overs) to ensure a meaningful wicket count.
- Why use sample-size floors at all?
- Small samples inflate luck. A bowler who concedes 0 runs in 3 balls has an economy of 0.00 — meaningless as a performance signal. Floors ensure every published metric reflects a minimum window of competitive pressure.
- What happens to sub-floor data?
- Sub-floor entries are suppressed from leaderboards and trend pages, or rendered with an explicit "sub-floor" disclosure tag. They are never silently surfaced as clean statistics.
- Are floors the same across all leagues?
- Yes — the same thresholds apply to IPL 2026, IPL historical, and Major League Cricket. Consistency across leagues is enforced by the build-time guard validate-sample-floors.mjs.