Here’s an interesting challenge. Can you predict how well the Syracuse men’s basketball team will perform this year based on the ranks of the incoming recruits and statistics from past seasons?
I have been giving it a try, and it is not easy. But after experimenting with a model for a while, one thing becomes clear: Syracuse needs an unexpected star to emerge or it will experience a significant drop-off from its recent success. Take a look at the chart below to see where I’m coming from:
The red dots show my model’s predictions for total wins in each season. The blue dots display Syracuse’s actual wins. Two things should jump out. For one, the predicted wins in the 2009-2010 season are far below—just over half—the actual wins that season. Secondly, the prediction for 2012-2013 is a drop to 22 total wins, just one more than the 2007 NIT season.
Let’s unpack these results. If it hasn’t occurred to you yet, the discrepancy between the predicted result and actual result in 2009-2010 can be explained by Wesley Johnson. My model, which I will explain in detail below, does not incorporate transfers. Normally, this would not skew the results that much (you probably don’t see the effect of Kristof Ongenaet), but Johnson was so good that the model suggests he accounted for upwards of 14 additional wins.
As for the prediction for this coming year, let’s start by exploring how the model works. At its heart, it attempts to balance three variables: the value that Syracuse loses from players departing the previous year’s team, the value that incoming freshmen provide and the improvement in value of the retained roster from the previous season. There are countless ways these factors could be quantified, but I focused on minutes played.
The model calculates value lost as the percentage of minutes played by departing scholarship players weighted by the winning percentage of the previous season. It determines the value of incoming freshmen through a combination of their RSCI (a weighted average of scouting-service rankings) and the average number of minutes played by previous Syracuse freshmen of similar rankings. Finally, the model assigns a value to the retained roster based on minutes played the previous season and a weighted average of the previous two seasons’ win totals. It estimates that the production of the remaining scholarship players will increase by 50 percent, although the model pushes this higher or lower depending on the personnel lost and gained from the previous season.
The model is by no means perfect. Even excluding the 2009-2010 aberration, the correlation between predicted wins and actual wins is a moderate 0.69. Nonetheless, the model does manage to predict successfully whether the Orange’s win total will rise or fall in each season, and it keeps within a margin of error of plus or minus four wins. If this holds for 2012-2013, the Orange should expect to win about 22 games and at most 26. This is a significant drop from the 29.75 averaged over the last four seasons.
The reason for this prediction is that the Orange is losing a large part of its rotation. It loses fully 51 percent of total minutes played, the most since Eric Devendorf, Paul Harris and company departed in 2009. And the number of minutes played does not come close to doing justice to the value Syracuse’s three NBA draft picks provided to last year’s team.
The question now is whether Syracuse’s recruiting class can help pick up the slack. According to RSCI, incoming freshmen Jerami Grant and Dajuan Coleman rank as 12th best class this year. If you add in Trevor Cooney using his score coming out of high school, Syracuse would leapfrog Texas to become No. 11. This is a solid class. Including Cooney, it is Syracuse’s highest ranked by RSCI since 2007 and the second highest in the last 15 years (it’s just a hair better than the 2010 and 2002 classes).
But according to the model, which weights the freshmen’s value by the amount of minutes played by previous freshmen of similar ranks, it won’t be sufficient to replace the value lost from last year. Full disclosure: The model predicts Coleman will average 21.5 minutes per game, while Grant will average 15.5.
This does not mean SU is doomed to a mediocre season, however. Far from it.
Syracuse can maintain its run of recent success if it can find value from its roster that the model does not expect. In 2009-2010, this was Johnson. In 2012-2013, it could take the form of a vastly improved player (perhaps a Michael Carter-Williams) or an underrated recruit (perhaps a Trevor Cooney with a year of practice under his belt). The key is that this player must not merely play well—he must play better than expected.
Regardless, next year Syracuse will need to find its Wesley Johnson or it will mean a return to the Syracuse teams of the mid-aughts. And after the last few years, who wants that?