Now that the field of 68 has been set, the next question is who will be lucky enough to win the NCAA Tournament? This has been the year of parity as we have seen a bevy of teams fall from the #1 ranking throughout the season. Also, this season features the most number of losses of #1 seeds in the history of the tournament with 23 combined losses between Kansas, Oregon, North Carolina, and Virginia. The previous record was from 2000 with 20 losses.
Nate Silver’s 538 page has a pretty cool interactive graphic that displays Kansas with a 19% chance of winning the tournament, followed by North Carolina with 15%. However 538 derived their probabilities, through a battery of Elo ratings and other various analytics (some based on play by play), this is just one interpretation of how to develop a probability structure. For our, we consider building an iterative model based on cross-correlated head-to-head match-ups (like in Colley’s method); score-based beta regression over game locations, time between games, player injuries and suspensions, and head-to-head-match-ups. It’s the very reason we predicted Connecticut to win the A-10 conference tournament when most other Bracketologists labeled UCONN as a “Bubble” needing to win the tournament. We’re not saying we’re better than others, we’re just saying our interpretations sometimes gets things right when others get them wrong (and vice-versa). We still have a hard time predicting Holy Cross’ run through the Patriot League.
Using our rankings system, we build a Markov simulation and run 10,000,000 simulations of the NCAA tournament. We tally the winners and then divide by 10,000,000 to obtain the probabilities. Sounds simple? It actually is… Anyways, here are the results:
Probability of Winning the NCAA Tournament
|Team||Probability of Winning|
|St Joseph’s PA||0.3999|
|Ark Little Rock||0.0214|
|S Dakota St||0.0146|
|WI Green Bay||0.0007|
|FL Gulf Coast||0.0000|
As we can see there is no clear cut winner. Kansas has the best chances to win, however, due to their ability to lose games to Michigan State, Iowa State, West Virginia, and (most ugly) Oklahoma State; Kansas also has the ability to lose as early as the Sweet Sixteen, when matched up against Maryland or California.
This isn’t the days of last year, when Duke and Kentucky crossed the 10 percent mark, with Wisconsin hovering near 9 percent. This is the worst fielded tournament in its 35 year history; in terms or losses. This will be an interesting tournament, filled with potential upsets everywhere.
How do you think the tournament will turn out this year? Who are you rooting for?