Ask LNH: Doesn’t kenpom treat all games equally?

Ask LNH: Doesn’t kenpom treat all games equally?

This question came in response to our article regarding Wisconsin’s adjusted offensive efficiency. The Badgers now have a positive points-per-100 possessions adjustment of 7.0, or 5.8%, given to their actual per game average offensive efficiency. Both of these adjustments are larger than those given to any other team in the nation.

The answer to the question is “no.” When thinking about kenpom’s ranking system, always keep in mind that it’s a predictive system. Giving more weigh to recent games when predicting the immediate future doesn’t seem odd at all.

Let’s do some quick math to see if treating the Kentucky equally to all others would cause a different answer. Coming into the Final Four, Wisconsin’s AdjOE was 127.5. They had played 38 games. 127.5 * 38 = 4,845. After the Kentucky game, Wisconsin’s AdjOE is 128.5. An AdjOE of 128.5 * 39 games = 5,011.5.

If all games were treated equally, you’d expect Wisconsin’s AdjOE against Kentucky to be 166.5 (5,011.5 – 4,845). The Badgers scored 123.3 pp100p on a neutral court against Kentucky and their 86.5 AdjDE.

If we adjust for the defense they played against (using the Wildcats’ AdjDE and a national average of 102.0), we’d add 145.4 to Wisconson’s AdjOE for the season.

(4,845 + 145.4) = 39 = 128.0, an increase from 127.5, but only about half of the 1.0 boost given based on Saturday’s results. Indeed, all games are not treated equally by the kenpom predictive system.

One other note – Kentucky’s defense, which has been lauded as one of the greatest ever by many people this season, no longer has the nation’s best AdjDE after allowing Wisconsin to score at a rate essentially equal to what the Badgers have been averaging for the full season.

Does this mean Kentucky has the nation’s second best defensive efficiency this season? Of course not! It means that the kenpom predictive system is using a worse AdjDE for Kentucky than it is for Virginia.

The per game average actual defensive efficiencies show a sizable advantage for Kentucky: 84.7 to 86.8.

Our pick tonight? Same as it has been since before the season began, throughout the year, and throughout the tournament: Duke.

Sharing Options: