The Closeline NFL betting model
The Closeline NFL betting model prices every game on the slate and surfaces only the bets where its number beats the vig-free market — spreads, totals and moneyline. Football is the sharpest market in North American sports: enormous handle, a full week for prices to mature, and margins that hinge on key numbers like 3 and 7. That makes NFL a demanding league for any model, and it is precisely why a disciplined, price-first process — rather than narrative or public sentiment — is the only approach that stands a chance.
Key numbers are where NFL edges live and die. Because so many games land on 3 and 7, a spread of +2.5 versus +3 is a genuinely different bet, and a model that ignores that distinction is leaving value on the table or walking into traps. Closeline prices around those key numbers deliberately, compares to the vig-free market, and only flags a bet when the edge clears the fees — then sizes it to the edge with fractional Kelly. In a market this efficient, patience and precision matter more than volume.
NFL runs during its season, and for the whole time it does the record is public and graded against the closing line — units, ROI, win rate and CLV, losses shown next to wins. Beating the close in football is hard, which is the point: sustained positive CLV against the sharpest market there is the strongest possible evidence a model is genuinely pricing games well. When the season is live, today's NFL plays appear on the free picks page; this page is the evergreen explainer for how the model approaches the league.
What the model prices in football
Team strength and matchup, quarterback situation, pace and pass/run tendencies, situational and rest factors, and — critically — the key numbers 3 and 7 that decide so many spreads. Those inputs produce a true spread, total and win probability, compared to the vig-free market. Each market is priced independently.
Why the NFL is the hardest test
A full week of price discovery and massive handle make NFL lines extremely efficient, so edges are thin and fleeting. That is exactly why we grade against the closing line and lead with CLV: beating the final number in the sharpest market is the clearest sign a model is adding real value rather than riding variance.