The Closeline NBA betting model
The Closeline NBA 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. Basketball rewards a systematic approach because outcomes hinge on measurable, fast-moving factors: rest and schedule spots, pace, injuries and rotations, and the sharp swings a single lineup change can cause. The model reprices the board as that information lands, so a soft number gets flagged before the market fully corrects it.
Where the model earns its keep in the NBA is discipline around information. Lines move hard on injury news and rest decisions, and the temptation is to chase every headline. Instead the model prices the game on what it knows, compares to the vig-free market, and only acts when the edge clears the fees — then sizes the bet to that edge with fractional Kelly. The result is a steady, repeatable process rather than a reaction to whatever is trending on a given night.
NBA runs during its season, and the record is public and graded against the closing line for the whole time it does — units, ROI, win rate and CLV, with losses shown alongside wins. Because basketball markets are efficient and move fast, beating the closing line is a demanding test, which is exactly why we lead with CLV rather than a short-run win streak. When the season is live, today's NBA plays appear on the free picks page; this page is the evergreen explainer for how the model approaches the league year-round.
What the model prices in basketball
Team strength and pace, rest and schedule spots (back-to-backs, road trips), injuries and rotation changes, and matchup-specific edges. Those inputs produce a true spread, total and win probability, which the model compares to the vig-free market. Spreads, totals and moneyline are priced independently — value on the spread does not imply value on the total.
How NBA performance is tracked
Every NBA pick is timestamped before tip-off and graded against the closing line, with units, ROI, win rate and CLV reported and broken out by market. Combinations that lose units get pulled back to shadow testing and must earn their way back to live — the same self-correcting loop we run across every league, so a cold market does not keep bleeding.