Track record

Every bet the model published, settled and unedited. Nothing is removed after the fact.

32547 settled
Rated 3+
Record13207−19179−16140.8% of 32386 decided · 32547 settled
Return on investment−7.25%per unit staked, flat stakes
Worst stretch−33.08%Thu 3 Sep, 1070 bets
Scope

Bets rated 3 and above, one per side, at the Consensus price — the average across the books we cover. Sun 30 Aug – Sat 19 Sep.

Daily ROI18 losing days of 21
SegmentW−L−DHitROIN
Moneyline220−322−040.6%−17.24%542
Spread212−273−343.7%—488
Total197−193−1050.5%—400
Props12578−18391−14840.6%−7.08%31117

How these are priced. Each bet is graded at the last Consensus price before the game — close to the closing number, not the price on the board when the bet appeared. At the first price a reader could have taken, the same bets returned −6.54%.

The settled bets

The 25 most recently graded, at the best price among the books this site shops. Every graded bet in this scope is in the download. Download this scope as CSV

RatingDayMatchBetPriceResult
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBJayden Mandal U 1.5Player Passing Touchdowns−168won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBLuke Weaver U 1.5Player Passing Touchdowns−139won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBCarson Clark U 4.5Player Receptions−118lost
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBRyan Browne U 1.5Player Passing Touchdowns−215lost
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBXavier Townsend O 3.5Player Receptions−136won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBWayne Knight O 1.5Player Receptions−130won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBSemaj Morgan O 2.5Player Receptions−125won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBBrian Rowe U 4.5Player Receptions−155won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBRyan Browne U 20.5Player Passing Completions−110won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBNico Iamaleava O 15.5Player Passing Completions−125won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBNico Iamaleava U 10.5Player Rushing Attempts−120won
Sat 19 SepMontana Grizzlies at Oregon State BeaversNCAA FBMontana Grizzlies +17.5Point Spread−110lost
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBFresno State BulldogsMoneyline−192won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBO 48.5Over/Under−105lost
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBJayden Mandal O 184.5Player Passing Yards−112won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBLuke Weaver U 237.5Player Passing Yards−112lost
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBJabari Bates O 61.5Player Rushing Yards−112won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBBryson Donelson U 65.5Player Rushing Yards−116won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBLuke Weaver O 17.5Player Rushing Yards−112lost
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBJosiah Freeman O 46.5Player Receiving Yards−115won
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBJosiah Freeman U 4.5Player Receptions−118lost
Sat 19 SepFresno State Bulldogs at San Jose State SpartansNCAA FBCooper Hoch O 3.5Player Receptions−148lost
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBPurdue Boilermakers +13.5Point Spread−100lost
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBO 52.5Over/Under−110won
Sat 19 SepPurdue Boilermakers at UCLA BruinsNCAA FBRyan Browne U 218.5Player Passing Yards−115lost

What the model claimed, and what happened

Sat 19 Sep · 221,687 settled predictions
33.6%claimed, averaged
33.5%actually happened
0.1482Brier score, lower is better
+0.334skill against a constant guess

Close to honest on average — 33.6% claimed against 33.5% observed. That average hides the finding below it.

No trend line yet. The 30-day window is still filling, so the sample grows every night and skill rises with it whether or not the model does.

The further we depart from the market, the more we overstate

Every prediction, grouped by the rating the model gave it. Low ratings track the price closely and are calibrated; the strongest depart furthest, and that is where the claim and the outcome come apart.

RatingClaimedActualGapGap, to scalen
Every prediction33.6%33.5%+0.1221,687
Rated 016.8%18.8%−2.0132,174
Rated 170.5%72.1%−1.646,638
Rated 330.2%22.0%+8.126,916
Rated 469.2%63.5%+5.711,608
Rated 572.6%56.0%+16.64,351

Why a rating is capped rather than trusted at the top of its range, and why this site publishes no return by rating: the ordering of these buckets is stable, their levels are not.

Calibration, market by market

Claimed probability along the bottom, the fitted correction up the side; the faint diagonal is perfect calibration. 13 markets carry a fitted curve — the rest are measured but not yet fitted, and are in the table below rather than dropped.

Batter hitsn = 31,868
claimed →
Batter total basesn = 31,683
claimed →
Batter doublesn = 10,705
claimed →
Batter RBIn = 10,537
claimed →
receiving_yardsn = 5,819
claimed →
offense_receptionsn = 3,160
claimed →
rushing_yardsn = 2,839
claimed →
passing_yardsn = 1,233
claimed →
Pitcher earned runsn = 930
claimed →
Pitcher outsn = 926
claimed →
Pitcher hits allowedn = 918
claimed →
passing_touchdownsn = 683
claimed →
rushing_attemptsn = 609
claimed →

Where we are worst

Every market we publish a board for, sorted by skill — the Brier score against a constant guess at the same base rate. Below zero the model lost to that guess: 14 of 52 markets. Nothing is removed and nothing is sorted flatteringly.

17 sit on fewer than 500 settled predictions. A skill score swings hard on a small sample, so those rows follow the rest rather than leading them — reported, not ranked.

MarketClaimedActualBrierSkilln
SpreadNCAA FB50.1%49.6%0.2770−0.108832
TotalNCAA FB49.9%49.8%0.2734−0.094894
SpreadMLB49.8%49.6%0.2723−0.089679
Pitcher outsMLB50.0%50.0%0.2712−0.0851,206
Pitcher earned runsMLB50.0%50.0%0.2597−0.0391,002
Pitcher hits allowedMLB50.0%50.0%0.2551−0.020990
TotalMLB50.0%50.0%0.2525−0.010662
Racing · show37.7%42.3%0.2436+0.00114,921
Racing · place26.4%29.8%0.2080+0.00616,844
horse_win_baseline15.8%15.8%0.1309+0.01817,773
Racing · win14.5%15.8%0.1302+0.02317,773
Batter singlesMLB49.9%49.9%0.2408+0.03713,865
horse_place_baseline29.8%29.8%0.1984+0.05216,844
MoneylineMLB49.8%50.0%0.2368+0.053688
Racing · all pools25.5%28.5%0.1908+0.06449,538
Batter total basesMLB27.1%27.6%0.1831+0.08338,852
horse_show_baseline42.3%42.3%0.2195+0.10014,921
horse_all_baseline28.5%28.5%0.1805+0.11549,538
rushing_yardsNCAA FB35.2%33.3%0.1896+0.1461,726
offense_receptionsNCAA FB38.6%39.2%0.1988+0.1651,269
passing_yardsNCAA FB42.2%40.9%0.2002+0.172867
receiving_yardsNCAA FB28.0%27.9%0.1655+0.1772,919
Batter RBIMLB50.0%50.0%0.2041+0.18413,685
offense_touchdownsNFL7.3%9.9%0.0688+0.2291,040
Pitcher strikeoutsMLB38.4%38.0%0.1726+0.2676,064
receiving_yardsNFL28.6%28.0%0.1462+0.2752,900
Batter hitsMLB30.1%30.1%0.1510+0.28239,071
offense_receptionsNFL35.0%34.8%0.1584+0.3021,891
Batter runsMLB25.4%25.2%0.1276+0.32325,223
rushing_yardsNFL36.5%37.4%0.1462+0.3751,113
MoneylineNCAA FB49.5%49.7%0.1457+0.417678
Batter doublesMLB49.9%49.8%0.1250+0.50011,543
Batter home runsMLB26.1%25.6%0.0742+0.61125,331
Batter stolen basesMLB33.0%32.1%0.0659+0.69715,284
Batter triplesMLB32.6%32.5%0.0149+0.9327,407
TotalNFL50.5%48.3%0.3412−0.367143
team_totalNCAA FB50.0%50.0%0.3321−0.3288
SpreadNFL54.5%49.3%0.2729−0.092136
team_totalNFL50.0%50.0%0.2633−0.0536
rushing_attemptsNCAA FB50.0%50.3%0.2595−0.038183
passing_attemptsNCAA FB46.4%38.8%0.2420−0.019250
rushing_touchdownsNFL50.0%50.0%0.2524−0.0106
passing_completionsNCAA FB50.0%50.0%0.2480+0.00898
passing_interceptionsNFL50.0%50.0%0.2441+0.02366
passing_completionsNFL50.0%50.0%0.2375+0.05064
offense_touchdownsNCAA FB15.7%10.1%0.0818+0.100178
MoneylineNFL51.0%48.5%0.2220+0.111136
rushing_attemptsNFL42.9%43.7%0.2021+0.178426
passing_attemptsNFL33.8%31.3%0.1714+0.202224
passing_touchdownsNFL40.1%41.5%0.1857+0.235205
passing_yardsNFL40.5%41.8%0.1858+0.236366
passing_touchdownsNCAA FB39.8%43.1%0.1845+0.247478

When the model changed

Every day the blend’s source weights were rewritten. A record is only meaningful against a model that was not quietly swapped underneath it.

  • Sun 20 Sep169 weights · 9 boards · 25 sources
  • Mon 14 Sep2 weights · 1 board · 1 source
  • Fri 28 Aug119 weights · 4 boards · 9 sources
  • Mon 29 Jun9 weights · 1 board · 5 sources
  • Tue 12 May91 weights · 2 boards · 9 sources
  • Mon 6 Apr4 weights · 1 board · 2 sources
  • Mon 9 Feb51 weights · 1 board · 8 sources
  • Sun 18 Jan19 weights · 1 board · 4 sources