The NBA stats that predict outcomes are useful starting points, not guarantees: offensive and defensive rating measure efficiency, the four factors help explain it, and pace helps translate it into expected points. Their value depends on the opponents faced, the players available and how much data you have. For how these feed into actual markets and prices, the NBA betting guide covers the groundwork.
Why do per-possession numbers beat per-game numbers?
Two teams can both score 115 a night and be nothing alike. In a hypothetical example, one scores 115 on 105 possessions; the other gets there on 96. Dividing points by possessions and multiplying by 100 gives roughly 109.5 versus 119.8 points per 100 possessions.
The second team produces more per possession in this example. That does not establish which team would win a meeting: defence and opponent quality are still missing.
NBA Stats defines offensive rating as points scored per 100 possessions and defensive rating as points allowed per 100. Their difference, net rating, measures scoring margin on the same scale. A team with an offensive rating of 118 and defensive rating of 113 has a net rating of +5.
That adds information the standings cannot show. A one-point win and a twenty-point win count equally in the standings, but differently in scoring margin. Net rating still needs context; it is not automatically a better forecast just because it differs from a team's record.
A useful comparison checks:
- Whether the teams faced opponents of similar strength.
- Whether the players responsible for those ratings are available.
- Whether the sample covers a season, a recent stretch or a small lineup stint.
Which NBA stats explain scoring efficiency?
The four factors separate efficiency into shooting, turnovers, offensive rebounding and free throws. Read them on both offence and defence: creating good shots and limiting opponents' good shots both matter.
| Factor | What it measures | What to check |
|---|---|---|
| Effective field goal percentage | Field goal accuracy adjusted for the extra value of threes | Shot locations, shooters and recent shooting variation |
| Turnover rate | Turnovers relative to possessions or an estimated opportunity count | The source's denominator and whether turnovers are live-ball |
| Offensive rebound percentage | The share of available offensive rebounds collected | Who is playing and the opponent's defensive rebounding |
| Free throw rate | Free throw attempts or makes relative to field goal attempts | Which definition the source uses |
NBA Stats provides definitions in its statistical glossary. Rate definitions can differ between data providers, so compare figures calculated the same way.
Effective field goal percentage fixes a limitation of raw field goal percentage: a made three is worth more than a made two. Its formula is:
eFG% = (field goals made + 0.5 × three-pointers made) ÷ field goal attempts × 100
For example, 40 made field goals, including 12 threes, from 90 attempts produces an eFG% of 51.1%, compared with a raw field goal percentage of 44.4%.
The factors describe different routes to efficient scoring. A turnover ends a possession without a shot; a live-ball turnover can also give the opponent a transition opportunity. An offensive rebound keeps an attack alive, while free throws add points without a made field goal.
There is no need to assign each factor a universal predictive weight. Their usefulness depends on the matchup, sample and forecasting method. Nor should their effects simply be added to net rating: the resulting points are already reflected in that rating.
How much damage does three-point variance do?
Three-point shooting can move a result substantially even when a team's underlying shooting ability has not changed.
Consider an illustrative model with 40 attempts, each having an independent, constant 36% chance of going in. Expected makes are 40 × 0.36 = 14.4. The standard deviation, a measure of variation around that average, is:
√(40 × 0.36 × 0.64) ≈ 3.04 made threes
Multiplying by three gives roughly 9.1 points of standard deviation from three-point scoring alone.
This is our calculation under simplified assumptions, not a measured estimate for every NBA team. Real attempts differ by shooter, distance and defensive pressure; shot volume also changes from game to game.
Opponent three-point percentage needs the same care. A high percentage allowed over a short stretch does not, by itself, establish poor defence. Check who took the shots and whether the defence conceded better opportunities.
Equally, recent hot shooting does not guarantee an immediate reversal. Regression means an estimate moves toward a more sustainable level as evidence accumulates; it does not mean a team is due to miss tonight.
How do you turn these numbers into a bet?
A workable process starts with a team-strength estimate and then accounts for the game being priced:
- Start from ratings and consider opponent strength and rest. The NBA back-to-back guide examines schedule context.
- Estimate possessions alongside each team's expected scoring efficiency.
- Adjust for minutes and availability, using player-level analysis where relevant.
- Compare your estimate with the spread or total and the offered payout. Allow for uncertainty in your own number.
For a total, multiply expected possessions by the sum of the two projected scoring rates per possession. At 100 possessions and projected efficiencies of 1.15 and 1.10, the result is 225 points. The difficult part is estimating those matchup-specific efficiencies; season averages alone do not resolve it.
Some bettors build statistical models; others concentrate on rotation changes or a small group of teams. Either approach needs a record of predictions made before games start. When testing a model, use games outside the data used to build it: explaining past results and predicting unseen results are different tests.
Compare current basketball odds across betting platforms.