NFL · MODEL PERFORMANCE

The whole record, one page

Every game we priced, with the market’s number beside ours and the result that settled it. Spreads and totals are separate products with separate thresholds, and both are graded against the closing line. Nothing is removed from the record after the fact.

✓ LIVE RECORDPicks frozen at kickoff and graded from final scores as the season is played.

Games priced, 2026

52.8%

Totals win rate

W7 ’26

Graded through

SEASON

TOTALSPREMIUM 4–6 · PREMIUM ★ 6+

52.8%

19W – 17L · 36 graded · 95% CI 37.0–68.0%

2026 · break-even 52.4%

+0.8%

ROI, 1u at −110

+0.3u

Units won

SPREADSPREMIUM

45.2%

19W – 23L · 42 graded · 95% CI 31.2–60.1%

2026 · break-even 52.4%

-13.6%

ROI, 1u at −110

-5.7u

Units lost

LINE QUALITYTotals

44.9%

Share of games BBMI landed closer than the market

computed from 78 paired games in scope

10.90

BBMI game-total MAE

10.29

Market game-total MAE

Calibration by edge

Totals · 36 decided

Pick rules and edge bands

Every pick that fired, split by the tier it fired at. This sport does not store a tier on the pick, so the tier is APPLIED from the published thresholds rather than read off the row. The total is the product card’s own cohort, which is what makes the two agree.

EDGE (PTS)	TIER AS APPLIED	N	WIN %	ROI
AS POSTED · THE TIER EACH PICK ACTUALLY FIRED UNDER
Edge 4.0 to 6.0	Premium	24	54.2%	+3.5%
Edge 6.0+	Premium ★	12	50.0%	-4.5%
All picks by tier	TOTAL	36	52.8%	+0.8%
BY EDGE INTERVAL · DESCRIPTIVE, SAME PICKS
Under 4.0 pts	No pick	41	43.9%	-16.1%
4.0 to 6.0 pts	Premium	24	54.2%	+3.5%
6.0+ pts	Premium ★	12	50.0%	-4.5%

How posted picks are grouped

The interval rows below are the same games cut by raw edge, INCLUDING the intervals that publish no pick. A band marked No pick is not a record: nothing was posted at that edge, and its win rate is what the thresholds are keeping out rather than what they delivered.

Season to date

2026

Win rate versus break-even

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

5 weeksbreak-even 52.4%now 52.8%

BY WEEK ▼	PICKS	WIN %	ROI	CUMULATIVE
W6 ’26	5	60.0%	+14.6%	52.8%
W5 ’26	7	71.4%	+36.4%	51.6%
W4 ’26	8	25.0%	-52.3%	45.8%
W3 ’26	8	62.5%	+19.4%	56.3%
W2 ’26	8	50.0%	-4.5%	50.0%

BBMI vs Vegas

78 games

Model versus market line quality

Both numbers measured against the same results, on the same games. Computed from the log: the Record card publishes no closer split for this sport.

WHICH LANDED CLOSER35 / 43 / 0 tie

BBMI44.9%

VEGAS55.1%

GAME-TOTAL MAEVegas by 0.60 pts

BBMI10.90

VEGAS10.29

WINNER ACCURACY · 93 GAMESVegas by 0.7 pp

BBMI62.4%

VEGAS63.0%

How winner accuracy is measured

Computed from 93 logged games with a decided winner. Level games are out: a tie has no winner to be accurate about.

Winner accuracy for this season

Projected winner, 2026. BBMI named the winner on 62.4% of 93 graded games against the market’s 63.0%. Break-even on a straight call is a coin flip at 50%, not 52.4%: no line is laid. The published row below is all seasons and does not move with this chip.

Published figures across priced games

Published, across every priced game. The rows above are computed from this market’s own 78 logged games. These are the site’s own aggregates over 910 games, walk-forward 2023-25 plus live 2026, graded, read rather than recomputed, each with the verdict the Record card reached.

PROJECTED WINNER ACCURACYVegas by 4.0 pts

Share of games each projection’s implied winner was right

BBMI66.2%

VEGAS70.2%

BRIER SCOREVegas by 0.013

Probability accuracy: rewards honest odds, not bold ones

VEGAS is derived from the closing spread, σ = 13.4 pts; the market posts no probability. Lower is better, and the two cells are scored over the same games under the same rule, so they compare directly.

BBMI0.217

VEGAS0.203

MARGIN MAE, PTSVegas by 0.66 pts

Projected home margin vs actual margin, in pts

BBMI9.95

VEGAS9.29

GAME-TOTAL MAE, PTSVegas by 0.80 pts

Projected total vs actual combined score, in pts

BBMI10.33

VEGAS9.53

Where the lines diverge

78 games

Under 1 pt17.9%

1 to 2 pts9.0%

2 to 3 pts12.8%

3 to 5 pts35.9%

5+ pts24.4%

How far the projections differ

How far apart the two numbers were. This needs no result, so it counts every game carrying both.

Calibration

Brier 0.217 · Published · full 910-game sample

Predicted versus actual win rates

A calibrated 60% should win about 60% of the time. Predicted win probability against the actual win rate, over every graded game rather than only the ones a pick fired on. This panel is the only block on the page that does NOT follow the season chip: the scorecard publishes one cohort across 2023 to 2026 and no per-season cut, so showing it under a 2026 heading would be claiming a season it does not measure. The published sample pools walk-forward seasons with the live season; it is one cohort, labeled as both.

HOME WIN P	N	PRED	ACTUAL	GAP
< 35%	160	26.5%	26.9%	+0.4
35–45%	149	40.2%	39.6%	-0.6
45–55%	154	49.7%	48.7%	-1.0
55–65%	174	59.9%	60.3%	+0.4
≥ 65%	273	75.4%	76.2%	+0.8

How to read GAP and Brier

GAP is actual minus predicted, so a positive number means the model was too cautious in that bucket and a negative one means too confident. Brier 0.217 over 910 graded games; lower is better, and 0.25 is what a coin flip scores.

Products & rules

thresholds from config

Rules change; posted picks stay fixed

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	PREMIUM ★
Spreads	Against the spread, the sign of the edge picks the side.	Edge 2.0 or more, Vegas spread 6.5 or less
Totals	Full-game over / under, the sign of the edge picks the call.	edge 4.0 to 6.0	6.0+
Straight-up	measured, not sold as a pick

A pick fires at an edge of 2.0 points or more with a Vegas spread of 6.5 or less, and every pick is Premium.

All sports →

The log

78 rows

How to read this game log

Every priced game in this market, with the market’s number beside ours and which one landed closer. BBMI ERR and VEGAS ERR are each side’s projected game total against the actual, so CLOSER is decided by those two columns and nothing else. A WEEK marked live is from the live record, frozen at kickoff and graded from the final score; an unmarked week is walk-forward, priced out of sample on a completed season.

What the SIDE column means

SIDE is blank on this market. A totals pick takes over or under, not a team, so the column would be answering a different question from the one it asks.

								FINAL
W7 ’26 live	NE @ CHI	51.0	51.1	0.1	52%	No pick	WIN	24–30	2.9	3.0	BBMI
W6 ’26 live	SEA @ DEN	40.5	42.6	2.1	49%	No pick	LOSS	14–17	11.6	9.5	VEGAS
W6 ’26 live	HOU @ JAX	42.0	41.8	0.2	56%	No pick	LOSS	34–31	23.2	23.0	VEGAS
W6 ’26 live	CHI @ ATL	51.5	48.0	3.5	51%	No pick	LOSS	31–23	6.0	2.5	VEGAS
W6 ’26 live	BAL @ CLE	39.5	39.9	0.4	53%	No pick	LOSS	10–13	16.9	16.5	VEGAS
W6 ’26 live	TEN @ IND	43.5	47.9	4.4	77%	Premium	WIN	26–20	1.9	2.5	BBMI
W6 ’26 live	NYJ @ NE	42.5	50.7	8.2	79%	Premium ★	WIN	19–26	5.7	2.5	VEGAS
W6 ’26 live	NO @ NYG	46.0	44.6	1.4	63%	No pick	LOSS	34–27	16.4	15.0	VEGAS
W6 ’26 live	CAR @ PHI	39.0	43.2	4.2	44%	Premium	PUSH	25–14	4.2	0.0	VEGAS
W6 ’26 live	PIT @ TB	38.5	42.8	4.3	38%	Premium	WIN	26–19	2.2	6.5	BBMI
W6 ’26 live	ARI @ LA	51.0	50.9	0.1	89%	No pick	WIN	18–16	16.9	17.0	BBMI
W6 ’26 live	LAC @ KC	41.5	43.8	2.3	73%	No pick	LOSS	6–22	15.8	13.5	VEGAS
W6 ’26 live	BUF @ LV	54.0	48.7	5.3	27%	Premium	LOSS	33–30	14.3	9.0	VEGAS
W6 ’26 live	DAL @ GB	54.0	59.0	5.0	54%	Premium	LOSS	13–23	23.0	18.0	VEGAS
W6 ’26 live	WAS @ SF	51.5	53.9	2.4	79%	No pick	LOSS	17–23	13.9	11.5	VEGAS
W5 ’26 live	TB @ DAL	49.0	52.0	3.0	77%	No pick	LOSS	24–16	12.0	9.0	VEGAS
W5 ’26 live	PHI @ JAX	41.5	43.7	2.2	66%	No pick	LOSS	17–19	7.7	5.5	VEGAS
W5 ’26 live	CIN @ MIA	44.5	51.8	7.3	22%	Premium ★	WIN	31–24	3.2	10.5	BBMI
W5 ’26 live	LV @ NE	45.5	45.5	0.0	68%	No pick	LOSS	17–45	16.5	16.5	VEGAS
W5 ’26 live	MIN @ NO	41.5	37.4	4.1	44%	Premium	WIN	23–13	1.4	5.5	BBMI
W5 ’26 live	CLE @ NYJ	38.5	38.3	0.2	50%	No pick	WIN	12–17	9.3	9.5	BBMI
W5 ’26 live	IND @ PIT	41.5	46.2	4.7	50%	Premium	WIN	22–34	9.8	14.5	BBMI
W5 ’26 live	HOU @ TEN	37.5	42.4	4.9	33%	Premium	LOSS	26–3	13.4	8.5	VEGAS
W5 ’26 live	NYG @ WAS	41.5	47.3	5.8	45%	Premium	WIN	27–27	6.7	12.5	BBMI
W5 ’26 live	CHI @ GB	46.5	51.5	5.0	51%	Premium	LOSS	7–31	13.5	8.5	VEGAS

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