NCAA BASEBALL · MODEL PERFORMANCE

The whole record, one page

Two products, both priced at -110 and both graded against the closing line. The samples are not alike: the spread has 1,087 decided picks behind it and the total has 173, so the second percentage is the less settled of the two whatever it reads today.

✓ SEASON COMPLETEEvery pick published before first pitch and graded against the closing line.

5,048

Graded games

57.2%

ATS win rate

Jun 21, 2026

Graded through

SEASON

ATSPREMIUM · EDGE 2–6 RUNS

57.2%

622W – 465L · 1,087 graded · 95% CI 54.3–60.1%

published · break-even 52.4%

+9.3%

ROI · flat 1u at −110

+101.0u

Units won

O/UPREMIUM · EDGE ≥ 4 RUNS

62.4%

108W – 65L · 173 graded · 95% CI 55.0–69.3%

published · break-even 52.4%

+19.2%

ROI · flat 1u at −110

+33.3u

Units won

LINE QUALITYATS

45.3%

Share of games BBMI landed closer than the market

computed from 1,087 paired games, on the run margin

4.85

BBMI run margin MAE

4.60

Market run margin MAE

Calibration by edge

ATS · 1,087 decided

Every pick in this product, split by how far the model and the market were apart. This feed stores no tier on the pick, so the split is APPLIED from the published band rather than read off the row, and it is a split rather than a product: both halves were sold the same way. The total is the product card's own cohort.

EDGE (RUNS)	N	WIN %	ROI
AS POSTED · THE TIER EACH PICK ACTUALLY FIRED UNDER
All picks as posted	1,087	57.2%	+9.3%
BY EDGE INTERVAL · DESCRIPTIVE, SAME PICKS
2.0 to 3.0 runs	421	54.6%	+4.3%
3.0 to 4.0 runs	397	57.2%	+9.2%
4.0 to 5.0 runs	200	60.5%	+15.6%
5.0 to 6.0 runs	69	63.8%	+21.8%

The interval rows below are the same picks cut by raw edge. They do NOT add to the headline the way the tier rows do: a band counts every decided pick at that edge, which is what the thresholds are selecting from rather than what they delivered.

Season to date

ATS

Cumulative win rate after each week, against the 52.4% break-even. The curve is the running rate, not each week's own, so a late swing moves it less than an early one.

19 weeksbreak-even 52.4%now 57.2%

BY WEEK ▼	PICKS	WIN %	ROI	CUMULATIVE
6/19 ’26	1	0.0%	-100.0%	57.2%
6/12 ’26	1	0.0%	-100.0%	57.3%
6/5 ’26	2	0.0%	-100.0%	57.3%
5/29 ’26	32	50.0%	-4.5%	57.4%
5/22 ’26	47	53.2%	+1.6%	57.7%
5/15 ’26	97	56.7%	+8.3%	57.9%
5/8 ’26	89	55.1%	+5.2%	58.0%
5/1 ’26	60	68.3%	+30.5%	58.3%
4/24 ’26	82	59.8%	+14.1%	57.5%
4/17 ’26	90	64.4%	+23.1%	57.2%
4/10 ’26	55	54.5%	+4.2%	56.1%
4/3 ’26	69	55.1%	+5.2%	56.3%
3/27 ’26	90	53.3%	+1.9%	56.5%
3/20 ’26	42	57.1%	+9.1%	57.3%
3/13 ’26	74	56.8%	+8.4%	57.3%
3/6 ’26	51	54.9%	+4.9%	57.4%
2/27 ’26	73	57.5%	+9.9%	58.0%
2/20 ’26	83	56.6%	+8.2%	58.3%
2/13 ’26	49	61.2%	+16.9%	61.2%

BBMI vs Vegas

1,087 games

Computed from this product’s own 1,087 paired games, on the run margin. The Record card publishes no closer split for this sport, so it is computed rather than read.

WHICH LANDED CLOSER492 / 572 / 23 tie

BBMI45.3%

VEGAS52.6%

RUN MARGIN MAEVegas by 0.25 runs

BBMI4.85

VEGAS4.60

Published, across every priced game. The rows above are computed from this product’s own 1,087 paired games. These are the site’s own aggregates over 5,048 games, full 2026 season (model launch Feb 14), read rather than recomputed, each with the verdict the Record card reached.

PROJECTED WINNER ACCURACYBBMI by 1.2 pts

Share of games each projection’s implied winner was right

BBMI63.9%

VEGAS62.8%

BRIER SCOREVegas by 0.002

Probability accuracy: rewards honest odds, not bold ones

BBMI0.223

VEGAS0.221

GAME-TOTAL MAE, RUNSVegas by 0.15 runs

Projected total vs actual combined score, in runs

BBMI4.61

VEGAS4.46

Where the lines diverge

1,087 games

2.0 to 3.0 runs38.7%

3.0 to 4.0 runs36.5%

4.0 to 5.0 runs18.4%

5.0 to 6.0 runs6.3%

How far apart the two numbers were. Every pick here already clears its product’s floor, so the narrow bands are thin by construction rather than because the model and the market rarely disagree.

Calibration

Brier 0.223 · Published · full 5,048-game sample

A calibrated 60% should win about 60% of the time. Predicted win probability against the actual win rate, across every priced game rather than only the ones a pick fired on. This panel does not follow the product pill: the scorecard publishes one cohort for the sport.

HOME WIN P	N	PRED	ACTUAL	GAP
< 35%	8	33.2%	25.0%	-8.2
35–45%	226	41.9%	30.1%	-11.8
45–55%	1,508	51.4%	43.9%	-7.5
55–65%	2,581	59.6%	64.9%	+5.4
≥ 65%	725	68.8%	84.6%	+15.7

GAP is actual minus predicted, so a positive number means the model was too cautious in that bucket. Brier 0.223 over 5,048 games; lower is better, and 0.25 is a coin flip.

Products & rules

thresholds from config

These are the definitions in force today. They change through the season as the model is refined, and a change applies from its date forward only: a pick is never refitted to a rule that did not exist when it fired, and the log is never edited to match a newer one. Every pick is counted in the record exactly as it posted.

PRODUCT	THE BET	PREMIUM BAND	GRADING
Spreads	Against the spread, the sign of the edge picks the side.	edge 2 to 6 runs	pushes excluded
Totals	Full-game over / under, the sign of the edge picks the call.	edge ≥ 4 runs	pushes excluded
Moneyline · all	Backs the side the price underrates, favorite or underdog.	edge ≥ 1pp of win probability	posted odds · ties are no action
Moneyline · underdogs	The same picks, plus-money half only.	plus-money price	posted odds · ties are no action

Spreads and totals settle at -110, so their win rates are directly comparable. A push is not a loss and does not appear in either at all: this sport drops them from the cohort, so every row in those logs is decided and the log length is the denominator. Thresholds are read from the published config, never restated here.

The two moneyline products do NOT compare to those two on win rate and do not compare to each other. They settle at the posted price, so the win rate that breaks even moves with the price: the underdog half wins well under half its picks and pays, and the favorite half wins well over half and does not. Read ROI on those two. The underdog cut is a SLICE of the all-picks cut and not a separate product, so its picks are counted in both.

All sports →

The log

1,087 rows

Every pick in this cut, with the market’s number beside ours and which one landed closer. BBMI ERR and VEGAS ERR are each side’s projected margin against the actual, so CLOSER is decided by those two columns and nothing else.

					PICK			FINAL
2026-06-21	North Carolina @ Oklahoma	+1.5	-1.5	3.0	Oklahoma	53%	LOSS	6–2	5.5	2.5	VEGAS
2026-06-17	North Carolina @ West Virginia	+1.5	-0.5	2.0	West Virginia	51%	LOSS	12–7	5.5	3.5	VEGAS
2026-06-06	Mississippi State @ Georgia	-1.5	-3.5	2.0	Georgia	62%	LOSS	12–13	2.5	0.5	VEGAS
2026-06-06	West Virginia @ Cal Poly	+8.5	+3.5	5.0	Cal Poly	39%	LOSS	17–1	12.5	7.5	VEGAS
2026-06-01	Troy @ Florida	-2.5	-4.5	2.0	Florida	66%	LOSS	10–2	12.5	10.5	VEGAS
2026-06-01	USC @ Texas A&M	-1.5	-3.5	2.0	Texas A&M	62%	LOSS	7–1	9.5	7.5	VEGAS
2026-05-31	Ole Miss @ Nebraska	+1.5	-2.5	4.0	Nebraska	59%	LOSS	6–3	5.5	1.5	VEGAS
2026-05-31	VCU @ East Carolina	-2.5	-0.5	2.0	VCU	53%	LOSS	0–10	9.5	7.5	VEGAS
2026-05-31	Arkansas @ Northeastern	+3.5	-0.5	4.0	Northeastern	53%	WIN	10–9	1.5	2.5	BBMI
2026-05-31	Saint John's @ Northern Illinois	+1.5	-1.5	3.0	Northern Illinois	54%	LOSS	21–8	14.5	11.5	VEGAS
2026-05-31	Virginia @ Jacksonville State	-1.5	+0.5	2.0	Virginia	48%	WIN	6–7	1.5	0.5	VEGAS
2026-05-31	East Carolina @ North Carolina	-3.5	-1.5	2.0	East Carolina	57%	LOSS	3–9	4.5	2.5	VEGAS
2026-05-31	West Virginia @ Kentucky	+1.5	-0.5	2.0	Kentucky	51%	LOSS	11–9	2.5	0.5	VEGAS
2026-05-31	UCF @ Auburn	-2.5	-4.5	2.0	Auburn	64%	WIN	3–9	1.5	3.5	BBMI
2026-05-31	Texas @ UC Santa Barbara	+4.5	+2.5	2.0	UC Santa Barbara	41%	WIN	6–4	0.5	2.5	BBMI
2026-05-31	Kansas @ Arkansas	+1.5	-2.5	4.0	Arkansas	59%	LOSS	13–10	5.5	1.5	VEGAS
2026-05-31	Florida State @ Northern Illinois	+6.5	+1.5	5.0	Northern Illinois	47%	WIN	7–4	1.5	3.5	BBMI
2026-05-31	Alabama @ Oklahoma State	-1.5	+1.5	3.0	Alabama	46%	WIN	9–7	0.5	3.5	BBMI
2026-05-31	Texas A&M @ USC	+1.5	-1.5	3.0	USC	53%	WIN	3–14	9.5	12.5	BBMI
2026-05-30	North Carolina State @ UCF	+1.5	-1.5	3.0	UCF	56%	WIN	3–9	4.5	7.5	BBMI
2026-05-30	Tennessee @ VCU	+3.5	+0.5	3.0	VCU	48%	WIN	4–5	1.5	4.5	BBMI
2026-05-30	Oklahoma State @ Alabama State	+5.5	+3.5	2.0	Alabama State	39%	WIN	8–7	2.5	4.5	BBMI
2026-05-30	Northeastern @ Missouri State	-1.5	+3.5	5.0	Northeastern	39%	WIN	5–1	0.5	5.5	BBMI
2026-05-30	Auburn @ North Carolina State	+2.5	-0.5	3.0	North Carolina State	53%	LOSS	17–13	4.5	1.5	VEGAS
2026-05-30	USC @ Lamar	+3.5	+0.5	3.0	Lamar	48%	LOSS	19–6	12.5	9.5	VEGAS

All picks locked before first pitch and graded against the closing line. Past performance is not a guarantee.