This is excellent Evan, great work as always! Do you have anything in the works to assign value metrics to euroleague or international NCAA new comers? They aren't necessarily freshman, but they also aren't an NCAA transfer. As this becomes more common, it seems like their "fit" will have enough unique factors to make up their own category.
I know you teased this at the end, but one statistical category I'd be curious about would be to see how coaching changes affect this data. For example, VCU has a new coach that will have put together a roster for 2025-26 comprised of a handful of returning VCU players, two players coming with him from Bryant, and a mix of high school and other transfer for the rest. Specifically, how would your data of returning players factor into returning to the coaching and its system versus that school and overall program. I guess that will be a topic for a different day though...
This is really interesting and well done! I won’t pretend to be an expert, but in medical research we look for statistically significant differences. Did you perform that type of analysis with this data?
Also, this may be too hard to tease out but what about coaching changes? The roster stability is likely low (see UK example) but should those schools be viewed differently? If a player stays at their school during a transition it seems like that should not count toward roster stability (unless they follow their coach from a different school perhaps).
Yeah, doing a hypothesis test here would certainly make sense. The problem is that every year the landscape changes significantly, so it would involve making a lot of assumptions about the data that aren't actually realistic in order to claim any statistical significance.
This is excellent Evan, great work as always! Do you have anything in the works to assign value metrics to euroleague or international NCAA new comers? They aren't necessarily freshman, but they also aren't an NCAA transfer. As this becomes more common, it seems like their "fit" will have enough unique factors to make up their own category.
Yep good suggestion. It's something I'll add to this analysis in the future.
Very interesting piece of writing, thank you.
I know you teased this at the end, but one statistical category I'd be curious about would be to see how coaching changes affect this data. For example, VCU has a new coach that will have put together a roster for 2025-26 comprised of a handful of returning VCU players, two players coming with him from Bryant, and a mix of high school and other transfer for the rest. Specifically, how would your data of returning players factor into returning to the coaching and its system versus that school and overall program. I guess that will be a topic for a different day though...
This is really interesting and well done! I won’t pretend to be an expert, but in medical research we look for statistically significant differences. Did you perform that type of analysis with this data?
Also, this may be too hard to tease out but what about coaching changes? The roster stability is likely low (see UK example) but should those schools be viewed differently? If a player stays at their school during a transition it seems like that should not count toward roster stability (unless they follow their coach from a different school perhaps).
Yeah, doing a hypothesis test here would certainly make sense. The problem is that every year the landscape changes significantly, so it would involve making a lot of assumptions about the data that aren't actually realistic in order to claim any statistical significance.