Table tennis data sources & methodology

Published observations and coverage

The current match publication identifies BetsAPI as its upstream source. Table Tennis Splits reads published data; it does not independently observe play. Source attribution is not an endorsement or claimed partnership.

The board and exact match pages use the available publication. They do not establish a complete league, player, or historical archive. If the publication cannot be verified, the site shows an unavailable state rather than substituting unverified data. An absent match does not establish that the fixture or its history never existed.

Which scores contribute?

Scored H2H observations must have complete, consistent point scores and match the players and available identity evidence. The reader excludes incomplete or inconsistent scores, future-dated observations relative to the publication, and identifiable duplicate observations. Conflicting fixture identities are withheld.

The scored observations shown in an analysis can therefore be fewer than the historical record count. A win/loss result without usable point scores does not become a zero-point observation. When no scored sample is available, a point average or rate is unavailable.

Totals, denominators, and windows

A match total is the sum of both players' points across the played sets. The mean is the sum of the usable totals divided by the scored sample size. Over and under comparisons are strict: a total must be greater or smaller than the selected line. Equality is separate.

For an illustrative sample with totals of 70 and 80, the mean is 75 and one of two totals is above 74.5, an observed over rate of 50%. Adding a match with missing scores does not change that scored denominator to three.

Available windows are up to 10, 20, 30, 40, 50, or 100 scored observations. A requested window is not a guarantee of that many results or verified latest-N coverage. Minimum-sample controls can withhold a percentage when its eligible sample is too small. Use the displayed count and contributing results.

Set conditions and player summaries

A first-two-set split means different players won sets one and two. It is different from a match merely having sets won by both players. A set-specific rate uses only observations eligible for that set or condition; its denominator can differ from the whole-match sample.

When available, separate published player summaries provide qualified win/loss observations against other opponents. They do not establish complete latest-match coverage, point-scoring averages, or the exact H2H sample. Do not combine those summaries with H2H contributors as if they were the same evidence.

Freshness, partial Live coverage, and corrections

Publication time describes when a snapshot was generated. A scheduled start describes a fixture. Historical observation dates describe the scored sample. Refreshing a publication does not make its older observations recent.

The Live page distinguishes observation age, partial coverage, and retained scores when that context is available. Retained scores can remain visible for context while being excluded from active set-play filters. A partial empty feed is not a claim that no matches are live.

Feeds can lag, schedules can change, and later corrections can alter previously displayed results. Historical rates describe observations, not calibrated forecasts or guaranteed returns. Read the sample size guide and total points guide, then choose a match from the current board.