I have a longstanding grievance with the term “toss-up.” It implies an outcome that resembles the toss of a fair coin, one whose prospects are about 50-50. But in political contexts, the term is sometimes applied far too liberally. Any reasonably competitive race is a “toss-up.” On the eve of the 2012 presidential election, for example, Karl Rove described 12 states as “toss-ups,” including Wisconsin and Nevada, where Mitt Romney had not led a poll in months. (Rove nevertheless seemed to be surprised by the election’s outcome.)
Poker players use a related term, “coin flip,” to describe certain hands that get all-in before the flop. Big edges are hard to come by in poker, and so poker players are more precise in their use of probabilistic language. Even so, there is some slack; a matchup between a pair of queens and an ace and king of different suit is often described as a “coin flip,” even though the queens have a 57-43 edge.
But Monday night’s NCAA men’s basketball national championship game? It really is a toss-up, at least according to the FiveThirtyEight forecast model, which gives Connecticut a 50.1 percent chance of winning and Kentucky a 49.9 percent chance.
You may have noticed that our odds changed since Saturday’s national semifinals. (You can find a complete archive of updates to our projections here). Right after Kentucky beat Wisconsin, we had the Wildcats as a 55 percent favorite to beat Connecticut.
But Kentucky’s chances dropped slightly with the news that Willie Cauley-Stein, Kentucky’s starting center, will miss the championship game. Our model adjusts for injuries, accounting for the value of individual players through a formula derived from Sports-Reference win shares. Cauley-Stein was the Wildcats’ fourth most valuable player, according to that system. His loss can obviously be overcome — Kentucky beat Wisconsin and Michigan without Cauley-Stein. But Kentucky’s edge against Connecticut was so narrow to begin with that the news was enough to tip the odds to 50-50.
Las Vegas still had the Wildcats favored by 2.5 points as of late Sunday night. I’m a bit curious as to why. Las Vegas odds usually track computer ratings closely, and Connecticut and Kentucky are about even in most computer systems based on their resumes this season. But those systems assume both teams are at full strength.
One reason may be the perception that Kentucky is simply the more talented team. I wouldn’t dispute that. (The FiveThirtyEight model’s way of accounting for a team’s talent is by looking at its preseason ranking, and Kentucky was No. 1 in the preseason. That’s one reason the model was more bullish on the Wildcats than the seeding committee.) On the other hand, if the perception is that Kentucky has been the hotter team, I’m not sure that’s true. Both teams have overcome exceptionally tough draws to reach the championship, but Connecticut has been more emphatic while doing so, winning by an average of 8.2 points so far in the tournament as compared to 3.6 for Kentucky.
In fact, Kentucky’s streak of close wins could be historic. Since the tournament expanded to 64 teams in 1985, the team to win the championship by the lowest aggregate scoring margin was Villanova that year. Villanova outscored its opponents by just 30 points combined over six wins (five points per game). Kentucky will break that record if it beats Connecticut by 11 points or fewer.