Build NBA Combos With Data-Driven Analytics

NBA combos are built from one game’s score distribution, shaped by pace, offensive and defensive efficiency, rest and — above all — who is actually available. The builder takes the current spread, moneyline and total from the public feed, folds in the injury report and recent form, and prices every leg from the resulting distribution: spread, moneyline, game total and team totals.

Basketball has a correlation the other sports mostly lack: blowout risk. When a big favourite covers by twenty, its starters often sit the fourth quarter and the bench plays out the clock, so the scoring pace of the last twelve minutes looks nothing like the first thirty-six. The shared model prices that relationship into any ticket that pairs a large spread with a total or a team total. Saved tickets are tracked; every leg is graded from the official final score, and a leg that lands on its line is a push and drops out of the ticket. Late scratches can land minutes before tip-off, and the model only knows what the feed has reported — every price is an estimate.

Take a home favourite at −9.5 with the underdog’s team total at 108.5. Independently, each leg may sit near even money. But the scenarios where the favourite wins by ten or more overlap heavily with the scenarios where the underdog’s offense stalled — “favourite covers” and “underdog under 108.5” arrive together far more often than independence suggests, so the naive product underprices the pair.

Pair that same spread with the game total over instead and the picture is murkier: a favourite that covers by playing fast usually does so in a high-scoring game, but a favourite that covers by defending pushes the total down, and a blowout sends the starters to the bench with minutes still on the clock. The builder reads all three legs from one distribution fitted to this game’s pace and availability, so the joint price changes with the combination rather than staying fixed. Availability news can change everything late, which is why the game page shows the current report.

The general version of this argument — why one score model beats multiplied legs — is on the combos overview, and the rules used to grade every saved ticket are on the pick tracker.