PredictionMachine.com is a website that I manage that plays the game 50,000 times before it’s actually played. For the NCAA Tournament, we have used what we call the “Predictalator,” a computer simulator I developed that plays through every play of every game, to not only provide picks and predictions for each of the first round games, but to actually play through the entire field and give the likelihood of any team making it to any level of the bracket.
This is my seventh tournament applying similar technology to the Madness. In the previous six seasons, as of Selection Sunday, this methodology has accurately forecast the eventual champion five times. For the 2010 NCAA Tournament, PredictionMachine.com likes Kansas to win the title more often than anyone else, yet just 15.7% of the time – leaving the field pretty open for other teams like Syracuse (10.6% likely to win the championship), Baylor (10.4%), Duke (10.2%), Ohio State (9.9%) or Kentucky (8.0%).
Paul Bessire joined Howard Eskin on WIP in Philadelphia to talk about the Predictalator, first round upsets, each region, and the most likely champion.
On how this works:
“We are actually playing the game a possession at a time, every single game 50,000 times. What we are really doing is playing the tournament 50,000 times. We advance all the teams and then we do it again and ultimately 50,000 times so we can good sense of the likelihood of any team making it to any level. The stats themselves that are being fed in – what is really important is per possession stats or pace-adjustments. Pay attention to what teams do every possession and how effective they are. If they get rebounds or cause turnovers to extend or shorten possessions that is important as well. On top of that, strength of schedule adjusting which is really difficult with 350 or so teams in college basketball. But that’s why we’re here. Long story short, we are really playing every single play all the way through of a game and then doing that 49,999 more times.”
On first round upsets:
“The biggest one would be Old Dominion, which is an 11 seed beating Notre Dame and they beat them 65% of the time, which is a pretty significant upset. Notre Dame appears to be grossly over-seeded. Another one would be Georgia Tech which we have winning as a 10 over a 7. Missouri wins over Clemson and Florida State over Gonzaga. (Wake Forest is also favored to beat Texas 52.4%.) Aside from the Old Dominion game, none of the other ones are that dramatic of upsets. However, people can go check this out. I would really encourage that if people want to force in a 5-12 where the 12 beats the 5, you can go see which 5 is most likely to be upset. In this case that’s Butler with UTEP. UTEP is a 41.1% favorite to win, so it’s not really likely, but it’s the most likely 12 over a 5 if you are looking for that.”
On the toughest region:
“The toughest region – by far – is the Midwest. Kansas is the number one overall team, not only to win the championship, but also in our Power Rankings where we look at all teams individually regardless of region. However, Ohio State is the second-best team in the country according to those same metrics. Obviously, being in the same region with Kansas is absolutely brutal to put them in there – for those team’s sake at least. And even Georgetown and Maryland are teams that we identified before the tournament as teams in that 3-4-5 range that could make the Final Four. So the Midwest is the toughest region. 32.4% of the champions come from that region.”
On the weakest region:
“The West is the weakest region by far. 19.7% of the time the champion comes out of that region. Syracuse is our most likely team. They are not nearly as strong as some of the other teams (in other regions). They make the Final Four 32.5% of the time. Kansas State makes it 23.4%. After that, there are seven or eight teams in the region that actually have a 3% chance or better to make the Final Four. So if it’s not Syracuse or Kansas State, the region is pretty open. The one thing that I would say about the West is that it may be the least exciting region that I have ever seen. Every team in there is an efficient, slow, defensive-minded, methodical team.”
On Temple and the East region:
“Two-thirds of the time Temple wins and a third of the time Cornell wins, so Temple is in good shape there. We have the projected score at 69-59. But you are right that Cornell is a good team. The East region really sets up for some chaos because Cornell, Washington and Missouri are all teams that are in that region that are double-digit seeds that could upset teams and even make a Sweet 16 or Elite Eight run. Wisconsin is a team that we really like because of their efficiency. Wisconsin’s chances to beat Temple are pretty good. Wisconsin’s chances to beat Kentucky are pretty good too because they are the opposite of Kentucky. Kentucky is at 28.2% likely to make the Final Four. But after that is Wisconsin at 21.7%. And after that is a surprising team in Marquette at 21%.”
On Villanova and the South region:
“Villanova makes the Sweet 16 47.5% of the time. That may not be more likely than 50%, but they have a higher percentage than anyone else so they would be in our most likely Sweet 16. After that Villanova tapers off pretty quickly because then they run into Baylor. And Baylor is one of our Final Four teams and is actually in our national championship game. Baylor is coming on strong and are very well constructed that is efficient offensively and defensively. Duke and Baylor are significantly more likely to make the Final Four. And even Texas A&M based on where they are is more likely to make the Final Four than Villanova.”