November net rating is mostly schedule
Every November the same ritual plays out. A team that was supposed to be mediocre opens the season with a gaudy net rating, and the takes arrive on schedule: the new coach has unlocked something, the young core has arrived, the preseason projections were lazy. A contender sits near zero and the word “concern” starts appearing in headlines. Net rating, points scored minus points allowed per hundred possessions, is the best single team number basketball has, which is exactly why people trust it too early. In the first month of the season it is measuring the schedule at least as much as it is measuring the team.
A dozen games is a tiny sample of possessions
By mid-November most teams have played somewhere around a dozen games. That sounds like a meaningful slice of the season until you translate it into the units net rating is built on. A dozen games is roughly twelve hundred possessions on each end, and the variance in points per possession is large. Shooting luck alone, a few extra corner threes falling in one week, a cold stretch from the line the next, can swing a team’s offensive rating by several points per hundred over that span without any change in how well it is actually generating shots.
The defensive side is worse, because opponent shooting is the least stable part of defense. A team whose opponents happen to miss open threes for two weeks will post an elite defensive rating that tells you almost nothing about its defense. Over a full season those swings mostly cancel out. Over twelve games they have not had time to, and the number you are looking at is a blend of real quality and noise in proportions nobody can separate yet.
The schedule is not balanced yet
The NBA schedule balances out over eighty-two games, roughly. It does not balance out over twelve. One team’s opening month might include a long homestand against rebuilding teams; another might open with a West Coast road trip, three sets of back-to-backs, and the two best teams in its conference. By the end of the year those differences fade. In November they are a large share of the story.
Home court is the easiest piece to see. Home teams win more often and outscore opponents by a few points per game on average, so a team that has played eight of its first twelve at home carries a built-in boost to its net rating that will reverse later. Rest matters too. Teams on the second night of a back-to-back, especially on the road, tend to play worse, and the distribution of those games across the first month is uneven. Then there is opponent quality, which is the hardest to judge early because everybody’s rating is noisy. A team that has beaten three opponents who look bad in November may have beaten three teams that are merely unlucky so far.
One blowout can carry the month
Net rating is a margin stat, and margins are where single games do outsized damage. A thirty-five-point win counts for as much as several solid ten-point wins. In a full season one blowout is diluted. In a dozen games it can account for a large share of a team’s entire point differential.
Blowouts are also the games least like the rest of the season. Once the margin gets large, starters sit, rotations deepen, and the final quarter is played by lineups that will barely see the floor in close games. The trailing team often stops competing hard. A big win says something real, but the last twenty points of it say much less, and early-season net rating counts all of them at full value. Stripping out garbage time, the minutes when the game is already decided, usually narrows the gap between the early hot starts and everyone else.
What regression looks like here
None of this means early net rating is useless. It carries signal, just less than it appears to, and the right way to read it is to blend it with what you believed before the season started. Preseason expectations, built on last year’s results, roster changes, and player age curves, are imperfect, but they are built on thousands of possessions. Twelve games of new information should move that belief, not replace it.
A useful habit is to ask how much of a surprising start would survive if the team’s opponents had shot their normal percentages. If a team’s defensive rating is elite mostly because opponents are shooting far below their usual mark from three, expect that to fade. If its offense is generating more shots at the rim, fewer turnovers, and more free throws than last season, that is a change in process, and process stats stabilize faster than shooting percentages do. Shot profile, turnover rate, and offensive rebounding settle down sooner than the shooting numbers that ride on top of them.
Adjusting before you conclude anything
The simplest improvement is to adjust for opponent and venue. Several public sites publish schedule-adjusted ratings that account for who each team has played and where. Early in the season those adjustments are large, and they often reorder the middle of the table considerably. A team that sits fifth in raw net rating can drop toward the middle once its soft early schedule is accounted for.
The second improvement is to look at the distribution, not just the average. A team with a modest positive net rating built from many close wins and a couple of narrow losses is a different team than one with the same number built from two blowout wins and a handful of losses. The first is probably closer to what it is. The second is a coin that has landed on its edge a couple of times.
The third is patience on narratives. Coaching changes, new systems, and young players taking leaps all happen, and early numbers are where they first show up. But the early numbers also show up for teams where nothing has changed at all. The way to tell the two apart is to look for a mechanism: more shots at the rim, a different pace, a new defensive scheme that is visibly forcing worse shots. If the explanation is only that the net rating is high, the explanation is the schedule until proven otherwise.
The number earns trust in January
Net rating is a good statistic. Over a full season it predicts future results better than win-loss record and better than most individual metrics. The problem is not the measure, it is the calendar. Every point of net rating in November carries a heavy load of opponents, travel, rest, and shooting luck, and that load lightens game by game. By the time teams have played thirty or forty games, the schedule has evened out, the blowouts have been diluted, and the shooting has drifted toward normal. That is when the number starts telling you about the team. Until then, it is telling you a lot about the schedule.