Transparent by design
How PitchMath reads a match
PitchMath reads completed matches and describes what the record actually shows — the pattern, the sample it came from and the question the data cannot answer. Every number on the site comes from the same stored analysis; nothing is estimated to fill a gap.
How to read the numbers
Four things worth knowing before you read a page
Season, Last 10, Last 5, Home, Away
A window is a set of that team's completed matches before the cutoff shown on the page. Last 5 is the five most recent; Home and Away are that team's matches in that venue. Early in a season a window can hold fewer matches than its name — the sample beside it always says how many.
Competition percentile
Illustrative exampleExample: Arsenal 14.5 shots per match · 83rd percentile
In that comparison group, the club recorded more shots than roughly 83% of the teams in it. It says where a number sits among its peers, not whether it is good.
Evidence grades
A is a strong sample with wide metric coverage; B is good; C is limited; D is very limited. The exact thresholds are the analyzer's own and are listed in the legend below — they are read from it, never retyped here.
When a statistic is missing
PitchMath leaves a statistic empty when the source did not provide a trustworthy value. Missing data is never turned into a zero, and a comparison that only one club has a value for is not drawn at all.
Step by step
How a match analysis is built
Collect finished matches
Only completed fixtures before the cutoff enter the sample. The analysed game and future games are excluded from pre-match windows.
Build windows
Season, last-10, last-5 and home/away windows are computed as unweighted averages or medians. Small venue samples stay visible.
Compare vs competition baselines
A percentile places a team's value among other teams in the same competition and season — relative position, never a chance of success.
Grade the evidence
Every profile gets A–D from sample size and metric coverage. The grade measures completeness of the record, not certainty of an outcome.
Write factors with caveats
Only patterns the sample supports become factors, each with its evidence and limits. What the data can't answer is stated plainly.
Evidence
Evidence grade legend
Thresholds come straight from the analyzer — never hand-copied numbers that drift.
≥10 matches, ≥80% coverage — strong.
≥5 matches, ≥65% coverage — useful with a limitation.
≥2 matches, ≥35% coverage — limited.
Below the minimums — unavailable.
How to read
The favored mark
Tap or hover the mark to see the gate that has to pass.
Marked only when both samples are ≥5 and the gap is ≥5%.
The boundary
What we don't do
- Winner probabilities
- Predicted scores
- Betting advice
- Invented statistics
Historical samples cannot know lineups, injuries, tactical changes or match-state decisions. PitchMath does not calculate winner probabilities, predicted scores or advice about what someone should do.
Detailed definitions
Season, last 10 and last 5
Each window is an unweighted average, median or frequency from completed matches before the analysis cutoff. The current fixture and future fixtures are excluded from pre-match analysis.
Home and away samples
The home team’s home matches and the away team’s away matches are shown separately. Their sample sizes stay visible because venue splits can be small.
Competition percentiles
A percentile places a recorded team value among other teams in the same competition and season. It describes relative position on that metric; it is not a chance of success.
Style profiles
Descriptive dimensions use only available source metrics. Every dimension names its inputs and coverage.
Evidence grades
Grade A means strong samples and coverage. Grade B is useful with a limitation. Grade C is limited. Grade D is unavailable. A grade measures completeness, not certainty.
Missing and null data
A dash means unavailable or below the observation minimum. PitchMath preserves a recorded zero but never converts missing data to zero.
Data cutoff and recaps
Every page exposes its snapshot cutoff. After full time, a recap may use the same canonical page; the result is never presented as something the pre-match page foresaw.