Public ledger

Track Record

Season-to-date backtest, flat 1 unit at -110 across all markets. We publish ROI, hit rate, and Brier score per market — the markets that aren't profitable yet stay visible here even when they get gated off the daily card.

Published Picks

Every play that's actually shipped on the daily card, graded after game completion. Real-money simulation at the prices we took. CLV captured at first pitch.

Mean CLV (full history): -0.502% over 14 snapped picks

30-day rolling

Picks

40(33 graded)

Hit rate

51.5%

Units P&L

+0.67u

ROI

+1.68%

Mean CLV (14 snapped)

-0.502%

Full history

Picks

40(33 graded)

Hit rate

51.5%

Units P&L

+0.67u

ROI

+1.68%

Mean CLV (14 snapped)

-0.502%

Recently resolved

DateMatchupPickCLVResultUnits
2026-05-04CIN@CHCUNDER 11.5LOSS-1.00u
2026-05-04CWS@LAAOVER 7.5+0.00%LOSS-1.00u
2026-05-04CIN@CHCCINWIN+0.96u
2026-05-04BAL@NYYOVER 8.5+0.26%WIN+0.96u
2026-05-04PHI@MIAUNDER 8.5+0.00%LOSS-1.00u
2026-05-04CIN@CHCUNDER 11.0LOSS-1.00u
2026-05-03MIL@WSHOVER 8.5LOSS-1.00u
2026-05-03TOR@MINOVER 7.0WIN+0.91u
2026-05-03ATL@COLOVER 9.5WIN+0.85u
2026-05-03LAD@STLSTLWIN+0.77u
2026-05-03CWS@SDCWSWIN+0.78u
2026-05-03TEX@DETUNDER 8.5+0.59%WIN+0.85u
2026-05-03CIN@PITOVER 7.5WIN+1.00u
2026-05-03ATL@COLOVER 10.0PUSH+0.00u
2026-05-03CLE@ATHUNDER 9.0-1.78%LOSS-1.00u
2026-05-03PHI@MIAUNDER 8.5-3.73%WIN+0.87u
2026-05-03LAD@STLUNDER 9.5WIN+0.94u
2026-05-03BAL@NYYBALLOSS-1.00u
2026-05-03AZ@CHCUNDER 11.5-2.74%WIN+0.95u
2026-05-02SF@TBSF-1.34%LOSS-1.00u

Calibration check

A 70% pick should win 70% of the time. The chart below bins our published picks by predicted probability and shows the actual hit rate. Dots near the diagonal mean the model's confidence numbers are honest.

Calibration: predicted vs. actual

33 graded picks

Each dot is a probability bucket. The diagonal is perfect calibration — a 70% pick wins 70% of the time. Above the line: we're under-confident; below: over-confident. Larger dots = more samples.

0%25%50%75%100%0%25%50%75%100%Predicted probabilityActual hit raten=4n=15n=8n=5n=1
BucketPicks (graded)Predicted avgActual rateDrift
<55%4 (4)47.1%75.0%+27.9 pts
55-60%15 (15)57.4%40.0%-17.4 pts
60-65%8 (8)61.2%75.0%+13.8 pts
65-70%5 (5)67.3%20.0%-47.3 pts
70-75%1 (1)70.8%100.0%+29.2 pts
75-80%— (0)
>80%— (0)

Drift = (actual − predicted). Positive = model under-predicted; negative = over-predicted. Buckets with zero graded picks show their total count in parens.

Backtest (season-to-date, every game-line at flat -110)

Independent of what we publish. Grades every line we could have played as a calibration check on the model itself. The market gate uses these numbers to decide what's eligible for the daily card.

Backtest bets

120,896

Hit rate

54.9%

Units P&L

+5634.74u

ROI

+4.66%

By market

MarketBetsHit %UnitsROIBrierGate

A market clears the gate when it shows ≥+1% ROI AND Brier < 0.246 over its rolling 200+ bet backtest. Failing the gate doesn't hide the market — it just keeps it off the daily card while we work on it.

Daily P&L (most recent first)

DateDaily unitsCumulative
2022-03-20-2.73u-2.73u
2022-03-21+0.81u-1.92u
2022-03-22+4.54u+2.62u
2022-03-23+20.81u+23.43u
2022-03-24+13.72u+37.15u
2022-03-25+29.54u+66.69u
2022-03-26+35.17u+101.86u
2022-03-27+19.45u+121.31u
2022-03-28+8.45u+129.75u
2022-03-29+1.63u+131.38u
2022-03-30+19.81u+151.19u
2022-03-31-12.64u+138.55u
2022-04-01-6.46u+132.09u
2022-04-02-3.37u+128.72u
2022-04-03+36.08u+164.80u
2022-04-04+3.99u+168.80u
2022-04-05+1.17u+169.97u
2022-04-06+0.45u+170.42u
2022-04-07-13.82u+156.60u
2022-04-08+18.99u+175.60u
2022-04-09-42.55u+133.04u
2022-04-10+12.72u+145.76u
2022-04-11+0.36u+146.12u
2022-04-12+3.26u+149.39u
2022-04-13-5.01u+144.38u
2022-04-14-15.91u+128.47u
2022-04-15-0.74u+127.73u
2022-04-16-3.46u+124.27u
2022-04-17+2.81u+127.08u
2022-04-18-6.73u+120.35u