|
|
|
|
NCAA Tournament PredictionsThe men's tournament kicks off tonight with the play-in game (err... Opening Round Game) between Mount St. Mary's and Coppin St. As we've done before, we've estimated the probability of each team winning the championship and advancing to the Final Four. The numbers are based on simulating a million tournaments using team ratings from the end of the regular season. (In fact, we do more than that -- we draw random team strengths from the distribution of possible team strengths, which allows us to capture the uncertainty inherent in any rating parameter.) The results are not surprising with Kansas winning 14.8% of simulations and advancing to the Final Four over a third of the time. Although the four #1 seeds are all favored to come out of their regions, they account for less than 50% of tournament winners. Furthermore, even though each has a high probability of making the Final Four, there's less than a 1% chance that they all will. Keep this in mind when you hear talking heads tell you about upsets. Some are inevitable, but knowing which ones will happen before they happen is virtually impossible. Probability (%)
Championship Final Four
Kansas 14.81 35.78
Memphis 11.73 32.21
UCLA 10.26 29.17
North Carolina 8.16 27.39
Duke 7.25 22.44
Wisconsin 5.01 15.97
Texas 3.65 16.00
Georgetown 3.23 12.20
Louisville 2.92 14.28
Stanford 2.72 12.70
Washington St. 2.66 12.88
Tennessee 2.36 12.57
Clemson 1.84 8.67
Xavier 1.72 8.48
Marquette 1.58 8.51
Drake 1.49 7.18
Michigan St. 1.34 7.47
Davidson 1.21 5.49
Butler 1.19 7.34
Southern California 1.09 5.35
Indiana 1.03 6.24
Pittsburgh 1.01 6.21
Notre Dame 0.95 6.22
Texas A&M 0.91 4.81
West Virginia 0.87 4.72
Connecticut 0.85 5.44
Purdue 0.76 4.52
Arizona 0.76 4.24
Gonzaga 0.67 3.71
Brigham Young 0.56 3.27
Kansas St. 0.51 3.06
Oregon 0.50 3.59
Baylor 0.42 2.97
Oklahoma 0.42 3.75
Miami FL 0.37 3.12
Mississippi St. 0.35 2.84
St. Mary's 0.33 2.78
Saint Joseph's 0.32 2.95
Nevada Las Vegas 0.28 2.22
Kent St. 0.27 2.14
Western Kentucky 0.24 2.01
Arkansas 0.24 2.22
Vanderbilt 0.20 2.03
South Alabama 0.19 2.00
Villanova 0.17 1.65
Temple 0.14 1.51
Oral Roberts 0.13 1.35
Kentucky 0.07 0.96
George Mason 0.06 0.91
Siena 0.05 0.75
Cal St. Fullerton 0.05 0.56
Cornell 0.03 0.50
Winthrop 0.03 0.58
Georgia 0.02 0.39
Boise St. 0.02 0.39
MD Baltimore County 0.01 0.25
San Diego 0.01 0.26
Portland St. 0.01 0.19
American 0.01 0.21
Austin Peay 0.00 0.14
Belmont 0.00 0.11
Texas Arlington 0.00 0.12
Mount St. Mary's 0.00 0.08
Mississippi Valley St. 0.00 0.00
Coppin St. 0.00 0.00
Kansas Finishes the Regular Season #1Kansas regained the top spot in our basketball ratings, with Memphis and UCLA in a virtual dead-heat for second place. At this point in the season, dramatic moves in the rankings are much less likely than in the polls owing to the relatively large amount of data already at our disposal and voters' propensity to over-weight recent results. Unlike the football ratings (which estimate team abilities from our original algorithm), the basketball ratings are determined using a reduced-dimensional approximation (which makes it possible to rank 330+ teams without having the computer explode). For those interested, we've computed strength of schedule too. (What more could you ask for?) SportsQuant Bowl PredictionsWe've done it for the last couple years, and we did it again this year -- predicting all 32 bowl games. For those interested, we posted another solid bowl season, going 23-9 (72%). SportsQuant in the NewsOur October 13 rankings were featured in the October 17 edition of the Charlotte Observer (again, saved for posterity). We were profiled at the Purdue University Statistics Department in October. Our rankings (and a rather flattering profile) were printed in the October 18, 2006 edition of the Charlotte Observer. Learning from our previous mistake, we made sure to save a copy before the paper archived it. Our rankings were mentioned in the November 18, 2005 Kansas City Star. Until recently you could read about them at the newspaper's website, but News isn't news if it's old, and the article has been relegated to the archives.
Objective Sports AnalysisHave you ever complained at your favorite team's personnel decisions? Have you ever groused about your coach's play selection? Have you ever argued about who really is #1? So have we, and that's why we're here. At SportsQuant, we apply operations research1 methods to sports problems such as game strategy, player evaluation, injury profiling and salary-cap management. Conventional wisdom is easy to follow, and you'll seldom be criticized if it backfires. But is it really the best? Or is it a self-fulfilling prophecy. If everybody does it this way, where's the incentive to buck the trend and try something perceived as risky and unorthodox? Our strategy page asks these types of questions and offers answers based on mathematical modeling and analysis rather than opinion. At our research page, you'll find a range of techniques and problems we've been thinking about. There, you'll find links to interactive simulations, musings on player evaluation, an introduction to the nearly world-famous SportsQuant college football rankings and some research papers which we've authored. 1. OR, which has its origins in the Allied effort in WWII, is the quantitative study of complex systems. Of course the methods have advanced considerably over the past six decades, and continue to advance, including our contributions. |
Send mail to
webmaster@sportsquant.com with
questions or comments about this web site.
|
