by Ken Pomeroy on Friday, March 29, 2013
With action moving to Cowboys Stadium tonight, there’s liable to be some discussion about how such a large venue affects the participants’ shooting. Or at least there should be, considering both Florida and Michigan are likely to be more dependent than normal on three-point shooting. Brian Cook at MGoBlog recently summarized what little research there is on the subject. A few weeks ago, John Ezekowitz put some stuff together attempting to quantify the effect of Princeton’s voluminous Jadwin Gym on shooting percentages. This is one of the better datasets pertaining to the issue since we can compare what Princeton’s opponents do between home and road games.
As it turns out, the three-point accuracy of Tigers’ opponents has been nearly five percent worse at Jadwin when accounting for each opponent’s season-long shooting ability. Across college basketball, there’s about a one percent difference between home and away three-point shooting, so you might attribute three to four percent to the lack of a typical shooting background.
It’s perhaps not quite that easy to assume that the three percent is truly due to the background. When looking across all of Division I, Princeton’s opponents’ splits are among the biggest in the game, but not the biggest. Since the 2003 season, there are three arenas that have demonstrated a bigger split.
[Note here that for expediency, I have not eliminated neutral site games, and there are a few that have ended up in the home/away bins, which will contaminate the results a bit. Thus, this is not the most scientific thing either, but hopefully a little better than what we have, which is not very much.]
One is Hawaii’s Stan Sheriff Center which appears to be a conventional basketball arena. Hawaii is known for its epic home-court advantage which is probably rooted in the long plane ride opponents have to take to play there, and we might reasonably attribute such shooting differences to jet-lag. The others are Illinois State’s Redbird Arena and Wright State’s Nutter Center.
There’s nothing to suggest that Redbird Arena is anything other than a typical arena. The Nutter Center accommodates hockey, and the basketball court gets shoved to one end of the arena, so it’s possible there are some background issues at the one end of the floor where there is no seating behind the hoop. As with Jadwin Gym, there is a strong difference in free throw percentages in these three arenas as well.
The Hawaii and Illinois State cases serve to illustrate that home/road splits can be generated from something other than shooting backgrounds. Another odd case is at Campbell, who has split time in two different, and small, gyms over this period and ranks at the top of home/road shooting splits by its opponents. Even with a large sample, statistical noise can have some influence.
However, it’s worth noting that John (and others) had a theory about Jadwin, then looked at the data which as it turned out, supported the theory. This is a stronger approach than looking at the data, taking the outliers and making a big deal about them whether it makes sense to or not. So I’m inclined to think that the lack of a shooting background has some influence on a team’s accuracy.
Another piece of information you may be interested in is that the Carrier Dome shows almost no home/road split over this time. Of course, the Carrier Dome is not configured for basketball in the way Cowboys Stadium will be, with the court right in the middle of the football field. So maybe the numbers from Jadwin are more reflective of what we might expect at domes. My conclusion: There may be a shooting-background effect worth considering, which is on the order of what a good three-point defense could do. Even if we make that assumption, cold shooting at Cowboys Stadium will be due more to whatever normally causes cold shooting and less about the venue.