Projecting top transfers

Transfers have become a bigger part of the game in recent seasons, and assessing their impact has become more important. This provides a new frontier for college basketball analysis. I know I've been somewhat confused at how seemingly ordinary players from lower leagues can transfer to a better team and have success. It's not even that they have success, it's that they're widely expected to have success before they play their first game.

For instance, T.J. McConnell was third-team All-Atlantic 10 in his freshman season at Duquesne, and yet he's being touted as a possible first-team All-Pac-12 player for Arizona by people in the know. As a freshman, Rodney Hood was the fifth offensive option on a Mississippi State squad that didn't threaten to get a bid to the NCAA tournament. Now many are expecting him to be one of the best players in the ACC in his first season at Duke.

I have faith in the experts in most cases, but one area where I feel less confident about them is their ability to see everything and everybody. This season, there are more than 200 transfers who will get significant playing time. A more objective way to predict the performance of transfer players can provide us with a way to estimate the production of some of these high-profile transfers, and also allow us to find some overlooked players who may excel as well.

With the help of Pabail S. Sidhu, basketball operations analyst at the University of Washington, I developed a model to predict how a transfer would perform with his new team. We're predicting offensive rating and usage here, so defense isn't a consideration. Although, defensive stats were considered as possible predictors of offensive performance. For instance, a high steal rate indicates that a player will be able to maintain a larger role in the offense when he transfers up.

What we found was that, in general, a player's role in the offense will tend to decrease as he transfers up, and vice versa. On the flip side, a player's efficiency improves as he transfers up and surrounds himself with more skilled teammates. So an efficient, high-usage player from a lower conference can have success transferring up, even though his role may shrink a bit. (For a recent example, think Luke Hancock.) None of this is too surprising, but quantifying the degree to which an increase in competition affects a player's production is useful.

Keep in mind that these forecasts do not involve playing time; we're determining what a player should be expected to do once he gets on the floor. For the most part, however, guys who are expected to be productive shouldn't have a problem getting playing time. Another caveat is that the forecasts do not take into account existing personnel on the transfer's team. Some of this is handled by the estimate of team strength that is an input to the model, but there are additional considerations that we'll need to address subjectively in each case.

Finally, we're looking only at players with previous experience at the Division I level. That means players who have a season where they played at least 10 percent of their team's minutes. So guys like Syracuse's Michael Gbinije and UNLV's Jelan Kendrick are not included in this analysis. This still leaves us with dozens of players changing teams, so I won't be able to cover them all, but I'll hit many of the notable transfers as well as some potential hidden gems.

With the preliminaries out of the way, let's check out the five transfers forecast to be the most productive in 2013-14.