Margin of victory and what it says about a team's true level
Average point margin is one of the best quick guides to how good a team really is. See why it beats the win column, how much noise it carries, and how to adjust it for blowouts and schedule.
Every basketball game produces two facts: who won, and by how much. The standings keep only the first. A 1-point win and a 31-point win each add one to the win column and nothing else, even though they describe wildly different games. Anyone who has watched a team squeak past a weak opponent and then watched another team take apart a strong one knows those results do not mean the same thing. Margin of victory, the average number of points by which a team outscores its opponents, restores the information the win column throws away.
Margin has been a favorite tool of basketball statisticians for a simple reason: it predicts. A team's average margin through the first part of a season is a better guide to its record in the rest of the season than its record so far. That finding is so consistent that the rule among analysts is to trust margin more than wins when the two disagree, though not infinitely more, and not without caveats.
This article explains why margin is so informative, how to translate it into expected wins, how much noise it contains, and where it needs adjusting: for blowouts, for strength of schedule, for pace and for the quirks of garbage time. It is the natural companion to our guides to net rating and Pythagorean wins, which approach the same truth from slightly different angles. Game results and margins for every contest are on the scores page.
Why margin carries more information than wins
Think of each game as a measurement of the gap between two teams. A win-loss result is a very coarse measurement: it only tells you the sign of the gap on that night. A margin is a finer one. It tells you the sign and the size. If you wanted to estimate someone's height and could only ask whether they were taller than a certain door, you would learn less than if you could measure them against a tape.
The practical effect shows up in the numbers. Suppose two teams each win 45 of 60 games. One has done it with an average margin of plus 9, the other with plus 4. The first team has not only won; it has dominated. If you had to bet on which would win the next game against an average opponent, you would choose the first, and you would be right more often than not. The win totals look identical, but the margin separates them.
Behind this is a statistical idea: the more levels of an outcome you record, the less luck contributes to the estimate. Basketball is especially friendly to this approach because scoring is plentiful. A team scores around a hundred points in a game; a handful of possessions shifting here or there changes the margin by a few points, but the long-run average is stable and informative.
From margin to wins
The usual translation is a rule of thumb: over a full 82-game NBA season, each point of average margin is worth roughly two and a half wins above or below .500. An average team wins about 41 games. A team with a plus 4 margin should win near 51, and a team with plus 10 should win in the high 60s. The rule is approximate, since the exact slope varies a little from year to year and from method to method, but it is reliable enough to build a mental picture.
A few historic teams illustrate the top of the scale. The Chicago Bulls of 1995-96, who won 72 games, the Los Angeles Lakers of 1971-72, who won 69, and the Golden State Warriors of 2015-16, who won 73, all had average margins in the neighborhood of plus 11 to plus 12 points per game. That range is what it takes to win close to 70 games in a season. Margins like that are rare because the typical good team sits well below them; teams with margins around plus 5 are comfortably among the league's top contenders in a typical year.
A more precise approach, Pythagorean wins, uses the ratio of points scored to points allowed instead of the difference. The two methods agree closely for teams in the normal range, which is why analysts often use them interchangeably for a rough view.
How noisy is a margin?
A margin is more stable than a record, but it is not exact. The amount of noise comes from two sources.
The first is the spread of results in single games. Even between teams of unequal quality, NBA game margins vary a great deal. The typical difference between the final margin and the margin you might have predicted from the ratings is something like a dozen points. That is a large number: it means that a team that is truly six points better than its opponent will lose a noticeable share of its games, and that any single game tells you relatively little.
The second source is the number of games. If single-game results vary by about 12 points around their expected value, then the average of 82 games varies by about 12 divided by the square root of 82, which is roughly 1.3 points. So a team's observed season margin is typically within a point or two of its true level, and sometimes a bit more. After 20 games, the same calculation gives about 2.7 points of uncertainty. After 10 games it is closer to 3.8. This is why early-season margins are so unsettled: a team that is plus 8 after ten games might be a genuine star team or just a good team that has been fortunate.
This logic is the heart of our discussion of small samples. The mathematics says that a few weeks of results can contain a large share of luck, and the sensible response is to pull early extremes back toward the league average, which is exactly what regression to the mean describes.
A way to combine evidence
A good habit is to treat a team's true level as a blend of two things: what the team has done this season and what you expected of it beforehand. If you had good reasons to expect a team to be around average, because of its roster and last year's results, then an early plus 8 margin should move your estimate up but not all the way to plus 8. As the games pile up, the observed margin earns more weight and the prior expectation less. Many rating systems do exactly this behind the scenes, beginning each season with last year's rating as a starting point and updating as games are played.
The adjustments that matter
Raw margin is a fine first look, but four adjustments sharpen it.
Blowouts and garbage time
When a game is decided, the outcome of the remaining minutes tells you less about the teams than the early part did. The leading team sits its starters, the trailing team may do the same, and the game finishes with reserves who rarely play together. The margin can shrink or swell in ways that do not reflect the teams' true quality.
For that reason many analysts cap or discount large margins. If a game ends 130-90, a model may count it as a win by 20 or 25 rather than 40. The cap does not reflect a belief that the extra points never matter; it reflects the observation that the first 20 points of margin are more informative than the next 20. Capping also reduces the reward for running up the score, a concern in college basketball where rating systems influence selection and seeding.
Strength of schedule
A margin of plus 8 against weak opponents is not as impressive as plus 8 against strong ones. The fix is to adjust each game's margin for the quality of the opponent, and for the location of the game, and then solve for every team's rating at once. The method is described in strength of schedule and how to adjust for it. In the NBA the schedule is nearly balanced, so the correction is small; in college basketball, with conferences of widely different quality, it can be large.
Home and road
Home teams outscore road teams on average, by a margin that has varied over the years but has been consistently positive. A team that played an unusually large share of its games at home will have an inflated raw margin, and a team that traveled more will have a deflated one. A home-court correction of a couple of points per game is the standard remedy. If you are curious how large the effect is and where it comes from, see home-court advantage in the numbers.
Pace
A game between two fast teams has more possessions, so scoring margins in it are stretched. A game between two slow teams is compressed. Points per game and point margin are therefore tied to tempo: a team that plays at a high pace tends to have a bigger absolute margin than an equally good team that plays slowly. Converting to per-possession terms, as in net rating, removes this distortion. For most NBA comparisons the pace differences between teams are modest, so margin and net rating tell similar stories, but in college basketball, where tempo varies enormously, the per-possession view is important. Our guide to pace covers the details.
Margin versus the record: reading the gap
The most revealing use of margin is to compare it with the record. The following table summarizes the typical patterns.
| Pattern | What it usually means |
|---|---|
| Record better than margin implies | Good fortune in close games; likely to cool off |
| Record worse than margin implies | Bad fortune in close games; likely to rebound |
| Record and margin agree | The record is probably a fair picture of the team |
| Big margin built on a few blowouts | Check whether the average hides inconsistent results |
The first two rows are the heart of the matter. A team that wins close games at an extraordinary rate is not necessarily a better team; it may simply have had the ball bounce its way. Studies of close-game results across many seasons show that the ability to win them repeatedly is weakly persistent at best, a point developed in our article on clutch statistics. A team with a great record and a mediocre margin should be viewed with the knowledge that part of its success is likely to fade.
There are legitimate reasons for a gap that have nothing to do with luck, and good analysis considers them. A team with an elite closer and a strong late-game system may win a few more close games than a comparable team without one. A team that rests its starters in games it has clinched will see its margin shrink while its record is unchanged. A team that makes a major trade midseason may have a margin that blends two different rosters. None of these explanations should be assumed; they should be checked against the games.
Margin in the playoffs
Margin also helps explain playoff series. A best-of-seven series is a longer test than a single game, and the better team wins it more often than it wins any single game. If one team is better by 5 points per game on a neutral floor, it wins a single game about 65 percent of the time and a seven-game series a good deal more often than that. Over many seasons, teams with larger regular-season margins have won series at a much higher rate than their records alone would suggest.
The caveat is that playoff conditions differ. Rotations shorten, opponents prepare specifically for one another, and the pace often slows. A team's regular-season margin is a useful starting point but not a final answer. A team that was great with a deep bench may be less impressive when only eight players play, and a team with top-heavy talent may be better than its regular-season margin suggested. Our look at how coaches adjust in a playoff series describes why a series can diverge from the regular-season script.
Splitting a margin into offense and defense
A margin is the difference between two numbers, and the two numbers can tell different stories. A team that is plus 6 because it scores 6 points more than the league average and allows the league average is a different animal from a team that is plus 6 because it allows 6 fewer. They may have the same expected record, but they will not necessarily behave the same way when circumstances change.
The two halves are not equally stable. Defense is harder to read, because so much of what happens after a shot is released, whether it goes in, is outside a defender's control. Opponents' three-point shooting, in particular, is volatile from month to month, and a defense can look better or worse for a stretch purely because of it. Offense, which depends heavily on the team's own shot selection and shooting, tends to be a little steadier. That does not mean defense is unimportant; it means a margin built mostly on a defensive surge deserves an extra look at how the opponents shot, a theme we take up in defensive metrics: why defense is hard to measure.
Margin beyond the NBA
The same logic applies in every league, with changes of scale. In the WNBA, the season is shorter, so a team's margin rests on fewer games and carries more uncertainty; the interval around a team's true level is wider, and early-season swings should be discounted even more. In college basketball, the schedule is the main complication. A margin of plus 15 in a weak conference may represent less quality than plus 8 in a strong one, and the huge differences in tempo mean that points-per-game margins can mislead. That is why the major college ratings adjust per possession and for opponent, as outlined in college efficiency metrics. In international competition, where teams may play only a handful of games against one another in a year, margin is the best summary available, but the sample is so small that rosters and context have to carry much of the weight.
In every setting, the principle is the same: more information per game, collected over enough games and corrected for obvious distortions, beats a simple tally of wins and losses.
A short guide to using margin well
- Use the average, not the extremes. One 35-point win does not make a team great; a season of plus 8 does.
- Count the games. A margin after ten games is a rumor; after forty it is a reasonable estimate; after eighty it is a settled picture.
- Check the schedule. Ask who the margins were earned against and where.
- Convert for pace if tempo differs. When comparing teams that play at different speeds, per-possession measures are fairer.
- Compare with the record. A large gap is a clue, not a conclusion.
- Look at the trend. A team whose margin has improved over the last month may have changed in ways the season average hides. You can follow the trend game by game on the standings page and in the scores.
What to take away
Margin of victory is the simplest way to get closer to a team's true level than the win column allows. It uses the full size of each result rather than the sign alone, it predicts the future better than the record does, and it translates cleanly into expected wins at a rate of about two and a half wins per point over an 82-game season. It is noisy over short stretches, and it needs adjustments for blowouts, schedule, venue and pace. Keep those in mind, treat early-season margins with the caution that a small sample deserves, and let the gap between margin and record be a prompt to ask why, not a substitute for the answer.