Interactive tool · Free
Game log research lab
Analyze your own results, compare non-overlapping windows, and see why an observed frequency is not a forecast. Includes a worked example and every calculation explained.
Open the research lab →TAP Props research library
Read the context behind the number.
A statistic is evidence, not a conclusion. These guides explain how to compare short and long samples, separate role changes from noise, verify lineup assumptions, and build a matchup view that records uncertainty instead of hiding it.
Recent form
How to read a recent-game trend
Use a repeatable process to decide whether a five- or ten-game run reflects a changed opportunity, a changed skill signal, or ordinary variance.
Read the guide →Sample size
Small samples, variance, and false confidence
Learn why short samples move so quickly, how selection bias distorts a trend, and which comparison baselines make a number more useful.
Read the guide →Availability
Lineups, roles, and late information
Build an availability-first workflow for projected lineups, probable starters, minutes, usage, injuries, rest, and confirmation timing.
Read the guide →Matchups
A practical matchup-research checklist
Combine baseline performance, current role, opponent environment, schedule context, and uncertainty without counting the same signal twice.
Read the guide →