Baseball Prop Edge: weather, Pitch DNA and AI models

Track record

Every pick we publish gets graded against the official box score. Picks are saved before first pitch and never changed after.

Last 30 slates 4,037 graded picks
67.9% Top 3 1+ Hit 53–25 record
39.2% Top 3 2+ Total Bases 31–48 record
14.3% Top 3 Home Run 11–66 record
30 Slates tracked 1,879 picks excluded (didn't start)

Hit rate by pick depth

How often the published top 1, 3, 5 and 10 in each market came through.

1+ Hit

Top 118–8
69.2%
Top 353–25
67.9%
Top 587–41
68.0%
Top 10171–86
66.5%

2+ Total Bases

Top 18–17
32.0%
Top 331–48
39.2%
Top 549–82
37.4%
Top 1095–161
37.1%

Home Run

Top 15–19
20.8%
Top 311–66
14.3%
Top 519–104
15.4%
Top 1032–217
12.9%

Estimate vs reality

For every estimate range, the average estimate (top bar) next to how often it actually happened (bottom bar). Perfectly calibrated estimates would match.

Model estimateActual

1+ Hit

4,037 graded estimates, Brier score 0.3068 (lower is better)

0–9%
55.5%
10–19%
56.9%
20–29%
50.9%
30–39%
52.6%
40–49%
56.8%
50–59%
61.0%
60–69%
64.0%
70–79%
59.7%
80–89%
60.0%
90–100%
62.5%

2+ Total Bases

4,037 graded estimates, Brier score 0.295 (lower is better)

0–9%
30.4%
10–19%
31.5%
20–29%
37.8%
30–39%
35.1%
40–49%
30.9%
50–59%
39.4%
60–69%
31.6%
70–79%
33.7%
80–89%
35.2%
90–100%
42.4%

Home Run

4,037 graded estimates, Brier score 0.1248 (lower is better)

0–9%
10.5%
10–19%
13.6%
20–29%
13.5%
30–39%
17.0%
40–49%
2.9%
50–59%
7.3%
60–69%
19.5%
70–79%
15.2%

How we grade

  • Picks are published before games start and archived the same day.
  • Archived rankings are never re-ranked or edited.
  • Only official starters with a recorded plate appearance are graded.
  • 4,037 graded, 1,879 excluded, 20 pending.

Historical results do not guarantee future outcomes. Players not in the official starting lineup, starters with zero plate appearances, and unfinished results are excluded from percentages.