NBA Net Rating: Reading Team Strength Without the Hype

Net rating, explained in one minute

Every December I get the same conversation with mates at the pub. Someone points at the standings, sees a team at 18-8, and declares them a contender. I ask what their net rating is. Blank stares. Net rating is the single most useful summary number for assessing NBA team strength, and most casual punters never look at it. The standings tell you what has already happened. Net rating tells you why it happened and whether it should keep happening.

In one sentence: net rating is a team’s offensive rating minus its defensive rating, both measured per 100 possessions. A positive net rating means the team scores more than it concedes on a possession-by-possession basis. A negative net rating means the opposite. The scale is intuitive. Each net rating point is worth roughly one expected point of margin per 100 possessions, which is roughly one expected point per NBA game given current pace. A team at +5 net rating is, on average, beating its opponents by about five points per game when matched against a league-average opponent. That is most of what you need to know to start using it.

The formula and how it reads

Net rating sounds simple because it is simple. Take offensive rating, which is the number of points a team scores per 100 possessions, and subtract defensive rating, the number of points a team concedes per 100 possessions. The current league-wide offensive rating in 2025-26 is around 114.3 points per 100 possessions, which by definition is also the league-wide defensive rating because one team’s offence is another team’s defence. The league’s net rating, summed across all 30 teams, is zero. Always.

So what does a healthy net rating look like? Last season’s championship contender tier mostly sat at +5 to +9. Play-in teams operated around 0 to +2. Lottery teams ran negative, sometimes deeply negative — a team at -7 net rating is genuinely losing every game by an expected margin of seven points before luck and variance get involved. The 2024-25 home advantage stats are useful here for context too: the league-wide home win rate was 54.4 percent, a historic low compared with the 60 percent norm of the 2000s and 2010s. That tells you home court is contributing fewer net rating points than the prior era, which means a team’s underlying net rating is doing more of the work in determining outcomes than the schedule used to.

Net rating is also influenced by pace, but indirectly. The per-100-possessions framing strips out the pace effect on raw scoring, so two teams playing at very different paces can still be compared fairly. What pace does affect is the variance around the central net rating estimate. Fast teams with high pace see more events per game, which produces a noisier per-game margin around the same per-100 average. Slow teams produce more predictable per-game margins around their net rating. For betting, this is useful: a team with a +6 net rating and slow pace will cover spreads more consistently than a team with the same net rating playing at high pace.

Lineup net rating versus team net rating

This is where net rating goes from a useful summary to an actual edge stat. The team-level number is an average over every five-man lineup that has played. The lineup-level number breaks that average open and shows you which lineups are doing the work.

The cleanest example is the starting five versus everyone else split. Most teams’ starters carry a substantially higher net rating than the team as a whole, because starters typically play together for longer stretches and rotate against opponents’ best players. The team’s bench lineups often run flat or negative net ratings, with the second unit having to make up ground against more dangerous opponent rotations than its talent suggests it should face. The team’s overall number is the weighted average of these wildly different units.

For punters, this matters because lineups change night to night. A team’s net rating with its top three players on the floor might be +14. With one of them out, it might drop to +6. With two of them out, it might be -3. The team-level net rating you read in a season summary smooths over all of that. Pre-game, especially when stars are listed as questionable or doubtful, the lineup net rating is what you need. The fatigue research on injury timing is relevant here too — work modelling more than 70,000 NBA games showed that the probability of a player picking up an injury increases by 2.87 percent for every 96 minutes played and decreases by nearly 16 percent for every additional day of rest. Lineup integrity, in other words, is not just a question of game-by-game star availability but of how the schedule has worn on the rotation. Reading ORtg and DRtg directly is the parent skill that makes lineup-level work possible — net rating is just the headline number.

When net rating disagrees with the standings

This is the question that pays the rent. A team’s net rating and its win-loss record do not always tell the same story, and the gap between them is one of the most reliable signals available to a punter.

The standard pattern is the “lucky” team. They sit at 26-15, which everyone treats as a strong season, but their net rating is +1.8 — barely better than mid-table. They have won close games. They have benefited from opponent injuries on specific nights. Their bench has held leads that, by net rating, they should not have held. This team is going to regress. The market will price them as a contender for several more weeks because the standings are the most visible signal. The window where their spread prices are inflated is the window to fade them on close lines.

The mirror pattern is the unlucky team. They sit at 18-23 but their net rating is +3.5. They have lost too many close games. The variance is going to swing back. Their spread prices, anchored to their record, undervalue what they actually are. Over the following 20 games, they tend to outperform the public’s expectations. This is the side of the trade that less-experienced punters miss because it requires backing a team that the standings call mediocre.

The trick is to know when the standings are telling the truth and when net rating is. Net rating is more reliable from about game 20 onward, once enough variance has averaged out. Before that, a hot or cold shooting stretch can distort the underlying number. After game 50, if the standings and net rating still disagree, the standings are more often the misleading signal.

Three betting uses for net rating

I use net rating in three places in my workflow, in this order of frequency. First, futures markets. Net rating is the single best summary stat for evaluating who actually deserves to win the conference. Twenty-game rolling net rating beats season-long record for spotting market overreactions. Second, spread pricing on mid-season games where the standings are misleading. Comparing two teams’ net ratings, adjusted for current rotation health, often produces a more honest spread than the line on offer. Third, totals work. Adding the two teams’ offensive ratings gives you an expected combined points-per-100 figure, which combined with projected pace produces a clean total projection. The current league pace of 104.5 possessions and offensive rating of 114.3 are the league-wide anchors against which any specific matchup deviates.

Created by the "NBA Stats For Betting" editorial team.