History (and famous baseball sabermetrician, Bill James) has taught us that point differential is a better indicator of a team’s winning percentage year to year than winning percentage itself. Lately, more teams, players and fans have used stats that account for how a player performs per possession (or per 100 possessions) instead of per game. Net rating (points scored per 100 possessions – points scored against per 100 possessions) differs from point differential (points scored per game – points allowed per game).
James also derived a formula using point differential to determine how many games a team should expect to win using the differential and called it Pythagorean winning expectation, and Daryl Morey altered it slightly to be used in basketball. The formula uses season-long point differential which accounts for some of the “luck” that a team has in a few games.
NBA.com gives net rating back to the 1996-1997 season for individual teams and is the source of my information. I used winning percentage instead of total wins in most cases because of the two strike-shortened seasons.
The picture above shows the correlation between net differential (x-axis) and winning percentage (y-axis). Obviously the bigger the net differential, the higher percentage of games the team is going to win, but a few of the teams stand out.
The Bulls’ team shown at the top of the chart is the season after the 72-10 season and is probably the best team since ’97. The team outscored opponents by 11.6 points per 100 possessions (the ’96 Bulls outscored opponents by 13.4 points per 100 possessions), and there’s a noticeable gap between that team and the next two highest (’08 Celtics and ’15 Warriors at +11 and +10.1 points per 100 possessions, respectively).
The other side of the graph shows the worst team during the period, the ’12 Charlotte Bobcats. The Bobcats won less than 11 percent of their games and were outscored by almost 15 points per 100 possessions (the next worst was the ’98 Nuggets, who were outscored by nearly 13 points per 100 possessions).
Two years ago, Kevin Love and the Minnesota Timberwolves were one of the unluckiest teams in recent memory. Despite a net differential of +2.6, the team won only 40 games finishing with eight fewer wins than expected. As it turns out, the Timberwolves were unlucky throughout Love’s tenure. In six seasons with the team, the Timberwolves won almost 26 fewer games than expected. While Love only played in 18 games during one of those seasons, it seems too easy to label each of the six seasons as a coincidence.
Former Timberwolves’ coach Rick Adelman decided to play J.J. Barea over Ricky Rubio (who was obviously the better player at that point) during the 2014 season, so it’s not too difficult to see why that team finished eight games below where they were expected to for that season. But it’s not as easy to determine why the Wolves consistently underperformed during Love’s tenure, especially during the season Love played the least (2013).
Not every team was as unlucky as the Wolves, however.
For every win the Timberwolves should have gotten, the Grizzlies did get. Since Zach Randolph and Marc Gasol have been on the same team, the Grizzlies have won at least 40 games every season (six total seasons), including three seasons above 50 wins.
However, the Grizzlies have been extremely lucky by winning 16 more games than expected. There is probably some information that isn’t included to explain this. Their offensive style is an inside-out approach — they work their offense through the two big men — which has worked for the majority of their time together. But when the Grizzlies get down by a large margin, they are unable to come back due to their inability to shoot threes, leaving them with larger deficits than a team that could shoot from the outside would have.
In just four seasons together, the Clippers have underperformed by almost six wins, but as with all data, figuring out the reasoning is a little more difficult. In the Clippers’ case, it could be due to the absence of a bench. Chris Paul, Blake Griffin and DeAndre Jordan have been one of the best trios since their inception. Maybe the three build big leads only to see their bench give back a few of the points.
Since 1997, the two best teams in the league have been the San Antonio Spurs and the Los Angeles Lakers. The chart above shows that the Spurs have outscored opponents by an average of 6 points per 100 possessions each season during that time, and the Lakers have outscored opponents by 5 points per 100 possessions. However, the Lakers have outperformed their expected wins by almost 21 wins over that time period, and despite winning more than 50 games in every season except for 1997, the Spurs have somehow underperformed by about 21 wins. The Lakers have been one of the “luckiest” team of the past 20 years — according to their expected wins — while the Spurs have curiously been one of the most “unlucky” teams.
The team that has best represented their expected wins is the Portland Trail Blazers. After almost 20 years of data, the Blazers have only differed from their expected wins by one measly win.
If you’d like to see more data on the expected wins, or just play with some of the information yourself, you can go here.