Track record
MLB record
HitsHRRHRTBPitcher KsTeam totalMoneylineRun lineGame totalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
NFL record
MoneylineSpreadGame totalReceptionsRec yardsRush yardsPass yardsAnytime TDCarriesTargetsPass TDsCompletionsPass attemptsTackles + astSacksFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
NFL markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
MLS record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
MLS markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
NBA record
MoneylineFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
NBA markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
Premier League record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
Premier League markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
La Liga record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
La Liga markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
Bundesliga record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
Bundesliga markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
Serie A record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
Serie A markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.
Ligue 1 record
Home winDrawAway winShotsShots on targetAnytime goalFailed to load results for this window.
finished right· Flagged read· Flagged reads
graded · none flagged · last 14d seasonno flagged markets in window
projection gradedprojections graded
Ligue 1 markets are published as projections and graded here, hits and misses alike. None was flagged in this window, so there is no flagged record to show. For a flagged record, see MLB's Flagged record.
– hit–miss · all markets graded · last 14d season · model models · model serving now: models serving now:
Past results do not indicate future outcomes.
How honest are the percentages?
Reliability
predicted vs. observed · last 14d seasonWas the model's stated probability honest? Each row is a predicted-probability bucket vs the rate actually observed — within ±5pp is well-calibrated. A bucket is only judged once it clears the same 200-resolved-outcome threshold a calibration curve needs; under it the rate is shown unjudged, because the sample cannot settle the question either way.
No calibration data in this window
Calibration appears once enough flagged markets resolve in the selected range. The model needs ~200 resolved outcomes per market before a curve is trustworthy.
| Market | Bucket | Predicted | Observed | n |
|---|
Reliability compares the model's predicted rate to the observed rate per bucket, judged only past 200 resolved outcomes. Each market's fitted calibration curve is under Calibration on the Results page. Past results do not indicate future outcomes.