Model PerformanceNCAA Baseball
NCAA Baseball · Product

NCAA Baseball — Total (O/U)

Premium · edge ≥ 4 runs · 62.2% O/U win rate · n=172 · live
All picks (edge ≥ 1.5 runs): 54.7% on n=1,052 (575–477)
live

Methodology

Picks qualify at edge ≥ 4 runs. Two cuts are distinguished by the Q1.9 amendment (11th drift pattern). The emission gate is the product STREAM — the pipeline emits every O/U pick at edge ≥ 1.5 (OVER_MIN_EDGE=1.5 Phase 2 validated; OU_OVER_MIN=OU_UNDER_MIN=1.5 on the picks page; no upper cap, OU_MAX_EDGE=7 in config is a historical-display constant only). The Premium TIER is the redactor's predicate — `_is_ncaab_baseball_ou_premium` at `premium_redact.py:423` locks picks at |bbmiTotal − vegasTotal| ≥ 4. Free users receive the [1.5, 4) band; the [4, ∞) band is paywalled. The headline reports the Premium tier; the secondary line ("All picks") reports the full emission stream for transparency. The per-band table below shows every band — the [1.5, 4) bands are explicitly the free territory, the [4, ∞) bands are the paywalled tier. All 2026 season, single-window live.

All figures computed from logged pick data at build · updated with each deploy.

Per-band breakdown

Each band shows sample size, record, win rate, and the 95% Wilson confidence interval. Bands inside the production gate are shown — wider edges have higher conviction but smaller samples.

Edge bandnW–LWin %95% Wilson CI
1.5–2.5 runs54428725752.8%[48.6, 56.9]
2.5–3.5 runs28315013353.0%[47.2, 58.7]
3.5–4.5 runs136855162.5%[54.1, 70.2]
4.5–5.5 runs68402858.8%[47.0, 69.7]
5.5–7 runs159660.0%[35.7, 80.2]
7+ runs64266.7%[30.0, 90.3]
Total1,05257547754.7%[51.6, 57.6]

BBMI vs Vegas

BBMI MAE
4.71 runs
Vegas MAE
4.41 runs
✓ Tighter predictor

Mean absolute error in predicting the actual combined total. Lower is a better predictor; Vegas remains the tighter total predictor.

n = 1,093 graded games