Cornell's upset chances

By Jordan Brenner and Peter Keatingand
No. 1 Kentucky (21.7) vs. No. 12 Cornell (88.4)
UPSET CHANCE: 19.2 percent

Hard to pick against the SEC champs who have lost only two games all season and who boast a top-10 offense (117.5 points per 100 possessions, 10th in the NCAA) and defense (87.5, eighth). And the Wildcats are particularly dominant inside: Kentucky, the fifth-tallest team in the country, shoots 54.5 percent on 2-point FGs (sixth in the NCAA) while leading the country in blocks (17.8 percent of opponent shots) and limiting opponents to just 41.3 percent (sixth in the nation) on 2-point FGs.



But as a result, 38.4 percent of the Wildcats' scoring margin comes from free throws, which is an incredibly dangerous stat according to our model -- so much so that we've been calling it the Magic Bullet for a couple of years. Why is it such a big deal? Start with the Wildcats' offense. If they run into a disciplined defense or a ref calling a loose game inside, they might not get the same calls they're used to seeing. But that's not the biggest concern, because Kentucky is right around the national average in terms of percentage of points that come from the line; the Wildcats have other ways to score. The big problem is on defense. Normally, the Wildcats' ability to protect the paint without fouling is a major advantage -- only 18.8 percent of opponents' points come from the line, which is 272nd in the country. But that means nothing to Cornell, because the Big Red's game plan is based around bombs, not dunks. Just 15.9 percent of Cornell's scoring comes from the FT line, and that's 341st in the country. In other words, Cornell has proven it can win without venturing into the lane, which takes Kentucky's shot-blocking prowess out of the equation. Instead, it's all about the long ball -- the kind of high-risk/high-reward strategy successful GKs tend to employ. It also, interestingly, makes Cornell's offense more efficient than Kentucky's.



Even without getting into a discussion about whether John Calipari-led teams typically exhibit championship-level discipline, the Wildcats have a few other weaknesses. They're not good at generating turnovers (19.7 percent of opponent possessions, 219th in the country) and they let opponents hit the offensive glass (offensive rebounds on 31.6 percent of missed shots, 122nd). They're also just average at protecting the ball (turnovers on 20.2 percent of possessions, 159th). So the key to this game will be whether Cornell can exploit those flaws and maximize its own possessions.



By now, you know Ryan Wittman, Louis Dale, Jon Jaques & Co. will hit their 3s; Cornell has shot 17-for-39 (43.6 percent) from beyond the arc in the tournament so far, which is just about what it's done all season (43.9 percent, best in the NCAA). But the Big Red need offensive rebounds and turnovers to give themselves enough opportunities for those shots to translate into points. And just as we pointed out with St. Mary's, it's worth noting that the model gives Cornell a much better chance of pulling an upset in this game than it did against Wisconsin.

MULECHOPS wrote:
By now, you know Ryan Wittman, Louis Dale, Jon Jaques & Co. will hit their 3s; Cornell has shot 17-for-39 (43.6 percent) from beyond the arc in the tournament so far, which is just about what it's done all season (43.9 percent, best in the NCAA). But the Big Red need offensive rebounds and turnovers to give themselves enough opportunities for those shots to translate into points. And just as we pointed out with St. Mary's, it's worth noting that the model gives Cornell a much better chance of pulling an upset in this game than it did against Wisconsin.

Uh, no.

Here's where his vaunted logic breaks down.

He predicts that Cornell will continue playing the way they've played, but dismisses any thought that Kentucky will continue to play the way they've played. The FACT remains that if each team plays the way they've played so far, UK will win by 20. I'll accept Cornell's threes as long as we get them into a running game. That'll do it.