Visit Bucky's 5th Quarter Get fresh Badger content from me at Bucky's 5th Quarter.
Subscribe to the new RSS feed Subscribe to my RSS feed from the new site.

Monday, January 25, 2010

Plus/Minus For Comeback Wins

Tracking "plus/minus" statistics is something I've always wanted to do but never got around to before. Traditionally known as a hockey stat, it's only more recently been used for basketball since the influx of advanced statistics into the sport.

The numbers are easy to follow. A player's plus/minus score is calculated by taking the number of points his team scores while he is on the court and subtracting the number of points the opponents score while he is off the court. This stat seeks to account for those "intangibles" that cannot be measured by other stats to connect a player's true value to what you see on the scoreboard.

But plus/minus alone is not always very helpful. Take blowouts for example. A team can build a 10-point lead with a player in the lineup and even when the player sits, his team could extend that lead to 20 or 25 points. Is the player really that instrumental in building the lead or is the opponent just vastly inferior? Also, you need to be able to compare players who play a different number of minutes.

Blogs like The Hoya Prospectus, which has used plus/minus for years (thanks Dylan), account for both problems. This is accomplished by also measuring the point differential when the player is off the floor ("Off") and subtracting it from the traditional "On" score for a Net +/- number. These numbers can also be adjusted per 40 minutes to compare players. I unabashedly copied this model when calculating the plus/minus scores for Wisconsin's past two games, which both happened to be comebacks of epic proportions.

The disclaimer is that plus/minus numbers are only another layer to add to existing stats to try to get the full picture. They are far from perfect. I sorted the players by Net +/- score, though I strongly considered sorting by minutes played to easily weed out data that is unreliable. As you might notice below, statistical outliers are more likely to warp the "net +/- per 40 min" numbers for the most-used and least-used players in individual games.

1/20: @Wisconsin 54, Michigan 48:
Now you can see the problem with calculating Net +/- per minute for individual games. We know that Rob Wilson and Trevon Hughes sparked the comeback against Michigan, but Hughes gets penalized immensely in the Net/40 column because while he took a quick two-minute breather, Wisconsin hit a 3-pointer and the Wolverines did not score.

Concluding that Pop was the least valuable Badger per-minute that day would not be accurate. The calculations will be more valuable when we can average out a whole season of data, which I intend to do. Just keep this in mind as you view the charts.

1/24: @Wisconsin 79, Penn State 71:
Jordan Taylor and Keaton Nankivil were incredibly valuable on average against Penn State. Mike Bruesewitz was once again terribly ineffective. With only four minutes played, Bruiser didn't get a chance to make his +/- more respectable, but that was a wise choice by Bo Ryan. "You gotta knooow when to hold'em ... know when to fold 'em ..."

Stay tuned for +/- for previous games and a season-to-date +/- board.

No comments:

Post a Comment

Do you like the NCAA tournament's new 8-team "first-round" format?

What should Wisconsin do with the newly vacated scholarship?

Poll: Who will win the Big Ten?

Poll: Who was Wisconsin's first-half MVP (thru 16 games)?

Poll: How Many Regular Season games will wisconsin win in 2009-10?