Data-driven NBA predictions from season scoring rates and live odds. Updated every 15 minutes throughout the season.
No NBA tips published yet. The card goes up each morning, UK time, when there are fixtures to price.
🔒 Pro members get the full NBA card with value edge % and stake recommendations on every game
Unlock Pro →The NBA model is deliberately simple, and we would rather describe it accurately than dress it up. For each team it takes points scored and points allowed per game across the current season, expresses both relative to the league average, and uses them to project a score for each side. Home court adds 3.5 points. The projected margin becomes a win probability, which is compared against the bookmaker's implied probability with the margin stripped out.
What it does not use: pace, possession-adjusted efficiency, rest days, travel, injuries or recent form. Those matter in basketball, and a model that used them would be a better model. Ours does not, and the record on this page is the record of the model as described.
NBA tips are published throughout the regular season (October to April) and the playoffs (April to June), across all 30 teams.
Home court advantage — worth roughly 3 to 4 points on average in the NBA. Our model applies a flat 3.5 to the home side, the same figure at every arena.
How many games have been played — season averages taken from a handful of games are noisy, so a tip on a team with fewer than fifteen games behind it carries a lower confidence figure. That penalty is the only adjustment the NBA model makes beyond the two scoring numbers.
The price, not just the pick — a selection is only staked when the model's probability is higher than the price implies once the bookmaker's margin is removed — the rest are published as insight, with no stake. A team we expect to win is not a tip if the odds already say so.
Playoff basketball is different from the regular season — teams shorten rotations, coaching matters more, and series momentum is real. Our model does not adjust for any of that. It applies the same season scoring averages to a postseason game as to a game in November. We publish playoff tips on that basis and say so here rather than claiming an adjustment we have not built.