Research Model
To Record a Hit Picks
This is a research model — hit picks are not included in the official picks feed. "To record a hit" props carry significant juice, so this board surfaces per-batter probability as a research and exploration tool. Only high-conviction plays surface as candidates. See the performance page for tier accuracy data.
Shadow variant in testing: 92.9% top-pick WP (26W–2L, 41 slates)
A projected-lineup shadow variant is tracking early-access hit picks. The top-pick tier has outperformed the confirmed-lineup parent model. Selective promotion being evaluated — picks gate still building.
Today's hit picks aren't ready yet
Current-day file exists, but no rows are ready yet.
Picks typically arrive by late morning ET once lineups are posted.
Research Performance
Research onlyBacktested across 103 slates (2026-04-10 through 2026-09-23). 5837 hits on 9527 graded batter-rows. Overall rate: 61.3%. This model is not part of the official public W/L record.
Hit rate by probability bucket
Higher buckets should produce higher realized rates. Small-sample buckets are noisy.
| Bucket | Picks | Hits | Rate |
|---|---|---|---|
| <55 | 3960 | 2291 | 57.9% |
| 55-59 | 2585 | 1581 | 61.2% |
| 60-64 | 2410 | 1590 | 66.0% |
| 65-69 | 569 | 372 | 65.4% |
| 70+ | 3 | 3 | 100.0% |
Hit rate by slate rank
Top-N per slate. Useful for picking parlay / SGP leg candidates.
| Rank | Slates | Picks | Hits | Rate |
|---|---|---|---|---|
| Top 3 | 87 | 193 | 113 | 58.5% |
| Top 5 | 91 | 318 | 190 | 59.7% |
| Top 10 | 98 | 617 | 390 | 63.2% |
| Top 15 | 103 | 901 | 585 | 64.9% |
Hit rate by lineup spot
Unknown rows did not have confirmed lineups at pick time.
| Order | Picks | Hits | Rate |
|---|---|---|---|
| 1-3 | 3487 | 2295 | 65.8% |
| 4-5 | 2195 | 1402 | 63.9% |
| 6-9 | 3845 | 2140 | 55.7% |
| unknown | 0 | 0 | — |
Parlay simulation
Top-N ranked batters per slate. Parlay hits when every leg hits. Pure co-hit rate, no odds assumed.
| Legs | Slates | Parlays hit | Parlay rate |
|---|---|---|---|
| 2 | 91 | 37 | 40.7% |
| 3 | 77 | 25 | 32.5% |
| 4 | 67 | 9 | 13.4% |
Baselines
Beats these and the model is adding value. Matches these and it isn't.
| Strategy | Slates | Hits | Hit rate |
|---|---|---|---|
| top_order_1_3 Random eligible batter in lineup slots 1-3 per slate (seeded). | 102 | 72 | 70.6% |
| highest_ba Highest-BA eligible batter per slate (season BA from pick snapshot). | 103 | 65 | 63.1% |
| random_eligible Random eligible batter per slate, seeded for determinism. | 103 | 63 | 61.2% |
- Research model. Not included in the official public pick record.
- Useful for candidate discovery for hit parlays / SGP legs.
- Parlay simulation selects the top-N ranked batters per slate. No odds are assumed; the metric is pure same-day leg co-hit rate.
- Probability buckets are computed across all graded batter rows, not just top-tier picks, so sample sizes are large.
Summary generated 2026-09-24T07:13:45Z. Deterministic, offline, no paid APIs.