Opponent-mix bias
A narrow interval can still be misleading if the record overrepresents weak or strong opponents. Precision does not repair selection bias.
Probability
Match Outcome Confidence Calculator
Build a Wilson confidence interval around an observed win, draw, or loss rate and see how sample size changes the evidence behind a match-planning assumption.
MATCH-RECORD UNCERTAINTY
Put a defensible range around one observed match outcome
Choose win, draw, or loss. The calculator treats that outcome against all other results, applies a two-sided Wilson interval, and keeps the complete three-outcome record beside the selected rate.
Selected observed rate-
Wilson lower limit-
Wilson upper limit-
Interval width-
Comparable matches-
Evidence grade-
LIVE DECISION RECORD
Outcome evidence and interval register
Counts, observed shares, selected-outcome interval limits, and the sample-size interpretation for the current record.
| Record item | Count | Observed share | Confidence role | Decision reading |
|---|
CURRENT CALCULATION PROCESS
Formula, current substitution, intermediate values, and reconciliation
pHat = x / n; Wilson = (pHat + z^2/(2n) +/- z*sqrt(pHat(1-pHat)/n + z^2/(4n^2))) / (1 + z^2/n)
| Symbol | Meaning and unit | Current value |
|---|---|---|
| W | Wins, matches | 18 |
| D | Draws, matches | 7 |
| L | Losses, matches | 5 |
| outcome | Selected categorical result | win |
| C | Confidence level, percent | 95 |
Waiting for valid inputs.
WHO SHOULD USE THIS
A record-quality check for analysts, coaches, and operations planners
Use the interval when a decision depends on how precisely a comparable historical record supports one outcome rate. Do not use it as a standalone forecast when opponent quality, venue, lineup, competition format, or time trend changes the data-generating process.
FIVE-STEP WORKFLOW
From comparable record to an auditable range
- Define the match population before counting: competition, season window, venue, and team state.
- Enter mutually exclusive win, draw, and loss counts; confirm their sum matches the source record.
- Select the outcome that drives the current decision instead of choosing whichever result looks favorable.
- Choose the confidence level and read both limits, not only the observed percentage.
- Save the result with the match list and rerun a narrower segment if the pooled record hides a material split.
FUNDAMENTALS
Five ideas behind match-outcome confidence
- Comparable sample
- The interval is only as relevant as the matches admitted to the count.
- Selected proportion
- For a win interval, wins are selected events and draws plus losses are the complement.
- Sampling uncertainty
- Two records with the same rate can support very different precision when their totals differ.
- Confidence level
- Higher confidence asks the procedure to cover more often and therefore produces a wider range.
- Wilson limits
- The adjusted center and denominator keep the interval inside the logical 0% to 100% range.
FORMULA AND DEFAULT SUBSTITUTION
Work the 18-7-5 record without rounding the comparison
n = 18 + 7 + 5 = 30; x = 18; pHat = 18/30 = 0.6000; z = 1.959964 at 95%
The Wilson denominator is `1 + z^2/n = 1.1280486`. Substituting the unrounded values gives a center near 0.5887 and half-width near 0.1654, so the reported win interval is approximately 42.32% to 75.41%.
THREE DEEPER READINGS
What changes the decision value of the interval
Time instability
A coaching change or major lineup shift may make older matches less comparable. A shorter window may be wider but more relevant.
Decision asymmetry
A must-win plan may focus on the lower win limit, while contingency planning may focus on the upper loss limit. The same record can support different cautious readings.
TWO DECISION CASES
Normal and boundary records tell different stories
Squad-planning baseline
An 18-7-5 record has a 60% observed win rate, but its 95% Wilson range is much broader. A staffing plan should not budget as if 60% were exact; it can stress both ends of the interval.
No losses in five matches
A 5-0-0 start shows a 0% observed loss rate, yet the upper Wilson limit remains substantial. The correct reading is “no losses observed in a limited sample,” not “loss is impossible.”
MATCH CONFIDENCE GLOSSARY
Six terms to retain with the record
- Observed rate
- Selected outcome count divided by all comparable matches.
- Wilson interval
- A score-based binomial-proportion interval with bounded endpoints.
- Lower limit
- The cautious low end of the procedure's supported range.
- Upper limit
- The cautious high end of the procedure's supported range.
- Interval width
- Upper minus lower limit, used here as a precision diagnostic.
- Coverage
- The long-run frequency with which intervals from the procedure contain the true rate.
LIMITS, DISCLAIMER, AND EVIDENCE
What this interval does not know
- Matches are assumed relevant and approximately independent after the user defines the comparison set.
- The model does not adjust for opponent strength, venue, scoring margin, injuries, tactics, or bookmaker information.
- It treats one selected outcome as binary against the other two and does not model correlation across fixtures.
- Rounding is display-only; interval calculations and checks retain unrounded values.
- This is descriptive statistical support for planning, not betting, legal, or contractual advice.
Evidence to retain: dated fixture list, inclusion rules, result source, selected outcome, confidence level, and exported calculation. Recompute after corrections or material changes to the comparable population.
RELIABLE SOURCES
Primary statistical references
MATCH OUTCOME CONFIDENCE FAQ
Questions about scope, precision, and interpretation
Why is one outcome treated against the other two?
A confidence interval for a selected outcome is a binomial proportion: selected matches count as successes and every other result counts as not selected. The full win, draw, and loss shares remain visible for context.
Why use a Wilson interval instead of p plus or minus a standard error?
Wilson limits behave better for small records and rates near zero or one. A simple normal interval can extend below 0% or above 100%.
Does 95% confidence mean a 95% chance that this exact interval is true?
No. In repeated comparable samples, the Wilson procedure is designed to cover the underlying rate at approximately the stated frequency. The parameter is not randomly moving inside one completed interval.
Can I mix home and away matches?
Only when that combined record answers the decision. If venue materially changes outcomes, calculate separate intervals or use a model with venue as a predictor.
Why can zero observed losses still have a positive upper limit?
A finite sample with no observed losses does not prove the underlying loss probability is zero. The upper limit expresses that remaining uncertainty.
Is this suitable for betting decisions?
No. It ignores opponent strength, market prices, lineup news, dependence, and selection bias. It is a descriptive planning aid, not betting advice.
IMPORTANT DECISION NOTE
Separate uncertainty from relevance
The Wilson interval quantifies sampling uncertainty under a fixed comparison definition. It cannot prove that the chosen matches are representative of the next opponent or future competitive conditions.