Accuracy and performance

Accuracy is correct draw calls divided by settled predictions. Void matches and pending matches are excluded. Every percentage shows the number of predictions behind it.

28%draw calls correct, from 199 settled
28%average estimate we published
25%of all matches in covered leagues ended level, last 12 months (3,788 matches)

Record by period

PeriodPublishedCorrectIncorrectVoidPendingSettledAccuracy
Today400040No data
Last 7 days1868041443% of 14
Last 30 days13641901413131% of 131
All time2265614312619928% of 199

Periods are based on kickoff date in Africa/Lagos. Void matches (postponed, cancelled, awarded, or with no result after 3 days) are shown but never counted as wins or losses.

Hit rate as results come in

15%20%25%30%35%40%45%50%55%60%65%70%75%80%85%90%95%100% 1199 settled predictions
Correct so far 28%Average published estimate 28%Competition average 28%

The solid line is the share of draw calls that were correct so far. The dashed line is the average draw chance we published for those matches. If the model is well calibrated the two should drift together as the sample grows. Early on they will not.

By competition

CompetitionPublishedSettledCorrectAccuracy
Premier League3026623% of 26
La Liga1817741% of 17
Serie A66233% of 6
Ligue 1121119% of 11
Championship4540718% of 40
Eredivisie55120% of 5
Primeira Liga1010110% of 10
Série A (Brazil)3021629% of 21
NPFL1816638% of 16
Premier League1919737% of 19
Botola Pro1010440% of 10
Premier League22150% of 2
Ligue 11611436% of 11
Ligue 144250% of 4
CAF Champions League111100% of 1

By estimated draw chance

Published estimateSettledAverage estimateActual draw rate
Below 26%2126%29% of 21
26% to 28%9627%20% of 96
28% to 30%4429%34% of 44
30% and above3832%42% of 38

A calibrated model has an actual draw rate close to its average estimate in each row. Rows with few settled predictions are noisy.

Probability quality (live predictions)

MeasureOur estimatesBaseline: competition average for every match
Brier score (lower is better), n=1990.19970.2008
Log loss (lower is better), n=1990.58830.5910

These are measured only on matches we picked because the model liked them, so the baseline here is not a like-for-like contest. The backtest below compares on all eligible matches.

Backtest on past matches

Run with model draw-v1.0. Each week was predicted using only earlier matches.

MeasureModelBaseline
Brier score, n=5,2180.19250.1938
Log loss, n=5,2180.57260.5759

Improvement over baseline (Brier skill): 0.65%. A positive number means the model scored better than the baseline. Gains in draw prediction are usually small.

Matches at or above our publishing threshold (26%): 961 of 3409 ended level (28%), against 26% for all eligible matches.

Calibration
Estimate rangeMatchesAverage estimateActual draw rate
15% to 18%5017%12%
18% to 21%20820%16%
21% to 24%56423%20%
24% to 27%162326%26%
27% to 30%166128%28%
30% to 33%85831%29%
33% to 36%22934%35%

Ranges with fewer than 20 matches are hidden. Covers 2023-11-02 to 2026-11-14.

Estimates based on historical data. Historical patterns do not guarantee any result, and no prediction here is a sure bet.