Public Proof / Settled only

Historical evidence without the live signal

A read-only record of settled model forecasts and parlays. Pending events never count toward these aggregates. Explore the public predictions and individual match receipts on the Predictions board.

DATA SCOPE

Resolved outcomes only

SCORING

Accuracy + Brier

EXCLUDED

Live and actionable data

Individual match receipts

Recent public forecast outcomes

Soccer shows its graded evaluation cohort. Other sports show forecasts published in the last 35 days. These are different cohorts and their counts should not be compared as one total.

The historical frozen ledgers below use separate evaluation cohorts, so their totals may not match the recent receipt cards.

Tennis / Frozen ledger

Settled forecast record

Published

SETTLED

1000

WINS

624

MISSES

376

ACCURACY

62.4%

Probability calibration

1000 settled forecasts with a comparable market probability.

0.223

MODEL BRIER

0.223

MARKET BRIER

0.224

Model Brier is 0.001 lower on this sample.

Parlay record

Fully graded frozen slips only.

18.2%

SETTLED

192

HITS

35

MISSES

157

A parlay counts as a hit only when every frozen leg resolves correctly.

MLB / Frozen ledger

Settled forecast record

Published

SETTLED

1738

WINS

942

MISSES

796

ACCURACY

54.2%

Probability calibration

551 settled forecasts with a comparable market probability.

0.241

MODEL BRIER

0.241

MARKET BRIER

0.239

Market Brier is 0.002 lower on this sample.

Parlay record

Fully graded frozen slips only.

35.5%

SETTLED

124

HITS

44

MISSES

80

A parlay counts as a hit only when every frozen leg resolves correctly.

Freeze rule

Forecasts are frozen before the scheduled event starts and are not rewritten after the outcome.

Grade rule

Only officially resolved outcomes count as wins or misses. Pending events are excluded.

Score rule

Accuracy measures direction; Brier score measures probability quality. Lower Brier is better.

Snapshot generated Oct 19, 2026, 8:13 AM UTC

Historical model results are informational and do not guarantee future outcomes. Small samples are preliminary and are shown with their sample size instead of being hidden.

Review terminal access