Should Kevin Ollie have UConn foul when they're up late in games?

Last night, for the second time this season, Kevin Ollie had to make the most-debated decision in college basketball: Up three with seconds remaining, should a team intentionally foul to send their opponent to the line, or risk giving up a game-tying three? And, like he did the first time he was faced with that choice, he declined to foul. That choice backfired on him a month ago when Marquette's Junior Cadougan drilled a buzzer-beater to force overtime in a game Marquette won. Last night things went a lot better, as Bryce Cotton's attempt rimmed out, letting UConn escape with an 82-79 win.

The quick and lazy way to evaluate those two choices would be to look at their respective outcomes, something Ollie alluded to after the game when asked about his choice:

"I don't know if I'm a bad coach because I didn't foul," Ollie said. "At Marquette I was a bad coach because I didn't foul. I just go with my gut and he missed the shot and we got the rebound."

But, as you may know, we here at TheUConnBlog are much bigger fans of examining the process that motivates a decision, rather than its result, so let's look a little deeper.

Coincidentally, The Big Lead took a look at this exact issue yesterday and created a useful primer on the issue generally. The main proponent of the "you should always foul" theory is probably DePauw basketball coach Bill Fenlon, who put together a paper on the topic a few years back. Fenlon's numbers say that you have about a 19 percent chance of being sent to OT if you don't foul, against a 4.9 percent chance if you foul correctly combined with about another 1 percent chance of losing / going to overtime if you screw up somewhere along the line. Luke Winn did a good job of going through Fenlon's decision-making process a few years back, through the lens of his former player (and current Butler coach) Brad Stevens.

But there is a glaring problem with Fenlon's numbers: they're not based on anything -- they're just probabilities he guessed at. At the other end of the spectrum is John Ezekowitz of the Harvard Sports Analysis Collective, who ran the numbers for the year 2010 and found that there was no significant difference between fouling and not fouling (in fact, fouling slightly lowered a team's winning percentage). Unfortunately though, Esekowitz's numbers aren't perfect either, because there is no uniform data set for college games, something commenters on his post have taken issue with.

The Big Lead's conclusion was that fouling is generally a better strategy, but that teams shouldn't employ it all the time because their opponents would anticipate it and adjust. That seems sensible to me, however, that's for an average team and I think UConn's current situation might call for a different plan.

That, as you might expect, is because UConn is a terrible rebounding team, so the risk of the opponent getting their own rebound off a missed free throw is increased. That was especially true last night, when UConn was outrebounded by 31, but it was also true against Marquette, who outrebounded UConn by 10 and had 11 offensive rebounds (good for more than 30 percent of their available chances).

Not fouling last night was probably a doubly good idea because of the state of UConn's roster at that point in the game. Enosch Wolf, Tyler Olander and DeAndre Daniels had all fouled out, meaning that a UConn team that was already getting crushed on the boards would probably have had to rely on Phil Nolan to get a rebound. Additionally, the foul almost certainly would have been committed by Ryan Boatright (he guarded Cotton up the floor and would have had the best chance to) and though that wouldn't have removed him from the game, it would have been his fourth, putting even more stress on UConn in a potential double overtime.

It's also worth noting that neither Marquette nor Providence were able to get much going from the outside in the two games. Marquette was 0-14 from behind the arc before Cadougan's shot, and Providence was a paltry 3-13 before Cotton's attempt.

For all those reasons not fouling last night seems like it was certainly the right decision. The Marquette game presented a tougher choice, but considering the rebounding concern and Marquette's 0-14 night, I think Ollie probably made the right call -- though I think you could defend either option.

It's also important to note that UConn did one thing much better last night than they did against Marquette: defend.

Watch Cadougan's shot:

Boatright, who was guarding him, gave him a lot of space, protecting the line to force Cadougan out, and as a result giving him an open (albeit deep) look. Compare that to last night:

Not only did Boatright bother Cotton the entire way up the court, but Niels Giffey also stepped up to keep a hand in his face when he slipped the coverage. At least for now it seems like Ollie's preference is going to be to let the other team shoot. That might not always be the right choice, but it sure is a lot better choice to make if you know your team is going to pressure any potential shooters.