- There's a lot of a variance in a team's win-loss record because of things such as luck and clutch performance.
- For that reason, point differential is a better indicator of future performance than win-loss record. A team's Pythagorean win total measures how many games it should have won based on its point differential.
- Teams from 2005 to last season that have underperformed by at least three wins have hit the over on their next-season win total 61.3% of the time.
NBA regular-season win totals are out, and there will be no shortage of articles and podcasts diving into whether teams will go over or under them. One piece of data cited as evidence — either way — will be a team’s Pythagorean win total from the previous season.
That may sound complicated, but I promise it’s not: We have a good idea of how many games a team should win based on its season-long point differential. However, it’s not always a perfect indicator, as there’s a lot of variance in clutch situations.
Remember two seasons ago when Russell Westbrook put up ridiculous numbers in the clutch on the way to averaging a triple-double on the season? As expected, the Thunder overperformed, winning 47 games; their point differential suggested they were a 43-win squad.
And it’s not all performance in the clutch. Sometimes there’s a lot of luck that goes into the final few minutes of games.
Unless you’re a dominant team such as the Warriors and win every game by 25, you’re going to have your fair share of close affairs. A half-court shot at the buzzer to lose is worth a full “L,” but those three points aren’t worth a whole lot in the grand scheme of a team’s point differential.
Some teams over- or underperform their point differential, which is a better predictor of future performance than a plain win/loss record. It makes sense that a team that drastically underperformed would likely bounce back the next season — and it’s possible Vegas oddsmakers could lower that squad’s next-season win total based on the unlucky record.
Do oddsmakers account for over- or underperforming teams?
I wanted to test that theory, so I went back to the 2005-06 season and looked at every team’s win record versus their Pythagorean win record, and then whether they hit the over or under on their next-season win total. Below is the win rate for each Pythagorean difference (negative means a team underperformed its Pyth mark).
- Negative numbers (team underperformed): Hitting the over on the next-season win total would be a “win,” as it would suggest the team either bounced back or oddsmakers lowered the next-season mark too much.
- Positive numbers (team overperformed): Hitting the under on the next-season win total would be a “win,” as it would suggest the team either regressed or oddsmakers raised the next-season mark too much.
Overall, underperforming teams since 2005 have hit the over 52.0% of the time, while overperforming teams have hit the under 52.6% of the time. Therefore, it seems there’s not a huge edge in using Pythagorean win differential to bet NBA win totals.
In a more micro sense, however, there could be an edge in some of the significantly underperforming teams: Squads that had a Pyth win differential between -3 and -7 (so they underperformed by three to seven games), hit the over on their win total 61.3% of the time the next season.
That sample includes 75 teams since 2005, so it’s not an insignificant number. Teams that overperformed by three to seven wins hit the under 51.4% of the time, so this could be an edge solely with underperforming squads.
>> Sign up for The Action Network’s daily newsletter to get the smartest conversation delivered into your inbox each morning.
Applications to the 2018-19 season
Here’s a table indicating how teams finished in relation to their Pythagorean win total last season, and their 2018-19 win total (via 5Dimes):
The first thing that stands out is just how much the Dallas Mavericks underperformed last season. No team since 2005 has had a Pythagorean win differential above or below eight wins, and the Mavericks were at -9 last season. They were a 33-win team — in terms of point differential — but finished with just 24 wins.
If their 34 mark seems high, that could be why, and it’s probably not a bad bet to take the over, especially with the additions of DeAndre Jordan and future NBA MVP Luka Doncic.
Other teams in that -3 to -7 range include the Charlotte Hornets, Utah Jazz, Brooklyn Nets, New York Knicks, Memphis Grizzlies, Atlanta Hawks and Orlando Magic.
The Hornets definitely pop off the page this season: Their point differential suggested they were a 42-win team, but they finished at just 36. Their opening mark on 5Dimes is 35!
And, finally, the Jazz are also intriguing: They were a 53-win squad by point differential but won just 48 games, and it’s possible, if not likely, that stud sophomore Donovan Mitchell will take another small step forward in his quick development.
With a full season under their belts with Mitchell running the team, the Jazz could get out to a quick start. They could definitely be a candidate to win 50 games and hit the over.