Are the signals earning trust?

Performance

Assists carries the strongest edge (+23.1 pts over base). Batter Hits is the volume leader (5,634 graded, -3.7 pts).

What's working

Every market's measured edge against how much of it we actually serve, ranked by lift. The tables further down read served predictions only, so a market that is good but never clears the floor is invisible there — and looks the same as one that does not work.

MarketAll predictionsServed (70+)Share served

Assists+23.1 pp n=268+32.9 pp n=15859%

Rebounds+18.7 pp n=320+32.9 pp n=17454%

Batter Ks+16.6 pp n=138+28.2 pp n=1410%

SP Strikeouts+14.8 pp n=166+15.8 pp n=15090%

Points+8.3 pp n=444+21.9 pp n=24555%

SP Hits Allowed+3.8 pp n=86+6.4 pp n=7385%

Batter Hits-3.7 pp n=5,634+2.6 pp n=1,30023%

“Starved” means the market beats its base rate by 5+ points across everything it predicted, yet under a tenth of those ever clear the floor — the scale cannot reach, so the edge is never offered.

Last 30 days · Featured108–5765.5%n=16514 void · 0 pending · avg score 99 · 23 slates

All qualifying1314–80062.2%n=2114

Featured108–5765.5%n=165

Other qualifying1206–74361.9%n=1949

Deeper diagnostics

Market pulse

Recent slate hit rates compared with each market’s own average. The right edge shows the selected period and the latest three slates versus the prior three.

MarketPeriodAug28Aug27Aug26Aug25Aug24Aug23Aug22Aug21

Batter Hits62%n=104building sample————————

SP Strikeouts69%n=13building sample——100%n=3———67%n=350%n=6

SP Hits Allowed80%n=5building sample——100%n=1———100%n=250%n=2

Points67%n=9building sample100%n=2100%n=1——0%n=3100%n=3——

Rebounds53%n=17building sample67%n=333%n=350%n=233%n=3100%n=30%n=2——

Assists88%n=17building sample100%n=375%n=4100%n=1100%n=450%n=2100%n=3——

Above own averageNear own averageBelow own averageFaded cells have fewer than 5 decisions

Does a higher score perform better?

Not enough graded props yet to judge whether higher scores perform better.

Current scoring engine only — 117 predictions from superseded versions excluded. Pooling engines once made the top band look anti-predictive when it was not; per-version records are in the model table below.

Score bandRecordHit rateSampleVoidsvs base

99–10041–1869.5%n=593+45.7 pp base 24%

Over vs Under

Over108–5765.5%n=165

Under0–0Not gradedn=0

MarketOverUnder

Batter Hits62.5% n=104Not graded n=0

SP Strikeouts69.2% n=13Not graded n=0

SP Hits Allowed80.0% n=5Not graded n=0

Points66.7% n=9Not graded n=0

Rebounds52.9% n=17Not graded n=0

Assists88.2% n=17Not graded n=0

Edge finder

SegmentHit ratevs baseRecord30-day trend

Batter Hits62.5% n=104+1.9 pp base 61%65–3962% 30d

Below minimum sample

ReboundsSmall52.9% n=17+33.2 pp base 20%9–853% 30d

AssistsSmall88.2% n=17+69.8 pp base 18%15–288% 30d

SP StrikeoutsSmall69.2% n=13+35.8 pp base 33%9–469% 30d

PointsSmall66.7% n=9+45.9 pp base 21%6–367% 30d

SP Hits AllowedSmall80.0% n=5+46.6 pp base 33%4–180% 30d

Consistency

Last 7

72.5%

37–14 n=51

Last 30

65.5%

108–57 n=165

Prev 30

69.2%

18–8 n=26

Season

66.0%

126–65 n=191

All time

66.0%

126–65 n=191

By month

MonthRecordHit ratevs base

2026-0718–869.2% n=26+16.1 pp base 53%

2026-08108–5765.5% n=165+18.1 pp base 47%

Model versions

VersionActiveRecordHit ratevs base

MLB Batter Hits

4 earlier versions2026-07-05→2026-08-2077–4563.1% n=122+2.5 pp base 61%

WNBA Rebounds

wnba-pra-v42026-08-23→2026-08-288–850.0% n=16+29.7 pp base 20%

2 earlier versions2026-07-15→2026-08-201–233.3% n=3+7.8 pp base 26%

WNBA Assists

wnba-pra-v42026-08-23→2026-08-2815–288.2% n=17+69.8 pp base 18%

1 earlier version2026-07-15→2026-07-161–0100.0% n=1+56.6 pp base 43%

WNBA Points

wnba-pra-v42026-08-23→2026-08-286–366.7% n=9+45.9 pp base 21%

1 earlier version2026-07-15→2026-07-155–0100.0% n=5+55.8 pp base 44%

MLB SP Strikeouts

sp-v42026-08-21→2026-08-268–466.7% n=12+32.4 pp base 34%

1 earlier version2026-08-11→2026-08-111–0100.0% n=1+77.2 pp base 23%

MLB SP Hits Allowed

sp-v42026-08-21→2026-08-264–180.0% n=5+46.6 pp base 33%

All qualifying includes every published prediction scoring 70+. Featured is the main-page eight. Observed hit rates are not probabilities; prediction counts are correlated within the displayed slate count.