clock menu more-arrow no yes

Filed under:

More Bad News for the Pacers From the Statistical Frontier

New, comments


The Pacers have struggled in the preseason thus far, losing their first three games by a combined score of 306-263. Fans have been able to maintain optimism by reminding themselves it's only the preseason. Unfortunately, Scott Sereday, of the ESPN Truehoop blog 48 Minutes of Hell, brought a little rain to that parade today.

Sereday posted an analysis of a team's record as a predictor of their regular season success. His numbers, which go back to 2001/2002, show that a team's preseason performance is actually a reasonably accurate predictor of how the team will fair in the regular season. For each team over that time period he compared their wins, to their estimated Win% for the previous season, the preseason, and the first 8 games of the regular season. His findings in his own words:

Not only do good teams play better in the preseason, but, after accounting for the number of games played, the preseason isn’t much worse at gauging regular season success than the regular season itself.

 


One other factor that he looked at was the skewed minute distribution which occurs during the preseason. As teams experiment with various lineup combinations, the best players on each team inevitably see fewer minutes than they would during the regular season. Sereday found that starters typically play about 70% of their regular season minutes during the preseason. The closer a team was to that minute distribution during the preseason, the better a predictor the preseason was of their regular season success. In fact he found:
After adjusting for differences in player minutes, I estimate that the preseason is approximately 90% as effective of a predictor as the regular season.
The preseason is a significant factor for predicting regular season success. In fact, preseason performance is comparable to regular season performance for predicting future wins and losses. After accounting for the number of starter minutes played, the difference becomes even smaller.
 
To incorporate the minute distribution for the Pacers' I needed to predict how many minutes the starters will likely play during the regular season. I used what I think are fairly reasonable estimates, for the players who look like the opening night starters:
  • Roy Hibbert - 32 MPG
  • Josh McRoberts - 32 MPG
  • Danny Granger - 36 MPG
  • Mike Dunleavy - 30 MPG
  • Darren Collison - 32 MPG

With their MPG averages so far in the preseason, the Pacers' starting lineup is playing roughly 74% of the minutes we would expect them to play during the regular season. Unfortunately, this number would indicate that their preseason performance is a more accurate predictor of what to expect in the regular season.

The Pacers' still have four preseason games left in which they could turn things around, and all is not lost. In fact, Sereday cautions Spurs fans not to read too much into their team's preseason performance, despite his analysis. It's still the preseason, I know, but with numbers like this my worries begin to feel a lot more justified.