Preseason ratings are up and the method to produce these is largely the same as used in previous seasons. However, this season I have added transfers and injured players with prior D-I experience to the mix.
The construction here is surely more clumsy than Dan Hanner’s lineup-based approach, but at least it will better handle the situations where most of a team’s production is expected to come from newly-eligible transfers or players that missed last season. Western Kentucky may still be ranked too low, but they are higher than they would have been in previous iterations of the model. Basically, all players with previous D-I experience are recognized in some way by the model. Or at least, they should be.
The one exception is that I’m not including second-semester transfers at this point. Over the seven seasons I’ve produced preseason ratings, my thinking on the purpose of them has evolved from trying to project end of season ratings to trying to predict how good a team is right now.
Most of the time, these two things are similar, but if you are trying to predict what will happen on November 11th, then it doesn’t do much good to have information about a player not eligible until December 15th.1
That said, my policy towards other players hasn’t changed. Unless a player has been ruled out for a large portion of the season, I include him in the model. I realize this isn’t totally consistent with my approach to transfers, but parsing words from coaches on a player’s health – especially when “no timetable for return” has become the default statement in so many cases – is not something I want to take on.
Finally, there is no change on incoming freshmen. The top 30 or so have an impact on a team’s rating and beyond that the computer is mostly blind to newcomers. That’s not to say it can’t make some guesses, though. In fact, it’s kind of a fun challenge to predict the impact of recruiting classes without any information on the recruiting class itself. Things like basketball budget, conference affiliation, recent performance, and whether the coach is returning handle some of this. But history says you can also glean some information from what kinds of players have left a team.
This is the case with Ohio State, who is ranked higher here than anywhere else. They had a young team last season, and the other indicators in the model are very positive. Furthermore, even though three rotation players transferred, those players were replacement-level quality for the Big Ten.
The fact that they are leaving is viewed as a positive in the model because if those players thought they would get more playing time, they would stay. And if they don’t expect to get more minutes, then those minutes figure to be taken by better players, which often means better players are coming into the program even if those players are ranked highly by recruiting services.
In Ohio State’s case, they have just one top 100 RSCI freshman joining the team, so the computer’s assumptions fail a bit with respect to the Buckeyes. Still, the news of transfers leaving Ohio State was not a bad thing and even without a stellar recruiting class, there’s a good chance the minutes that need replacing will end up being more productive this season than last.
|^1||This also applies to my guesses at the national average of tempo and efficiency. These values reflect what I expect them to be on opening day.|