kenpom win probability
Using Kenpom’s Projected Win Probability to Predict Auburn’s Conference Schedule. According to the ESPN BPI, UVA has a 86.2% probability of winning. January 23 They’ll turn the lights back on in Rupp Arena when UK returns home to host LSU this coming Saturday. To do this, I filtered play-by-play data using my ratings (while accounting for game location). Ken Pomeroy | 05.10.14. Thatâs a mouthful, but at halftime, instead of assuming there is 50% of the game yet to be played, Iâm going to pretend like there’s 70.7% of the game left to be played. Your home for College Basketball on the internet. Leverage is not based on what happened during the possession, but is the range of win probability based on what could have happened. Thereâs lots more calibration to do with this system, but since I just thought about doing this a few days ago, it was necessary to get something done before the Final Four started. The Tigers are currently in second place in the SEC with a 5-1 record having won five straight since a loss at Florida to open conference play. Previous attempts to quantify in-game win probabilities in college basketball are limited and have left me unsatisfied because none of them accounted for information known before the game starts. Let’s face it, golf is not the most exciting spectator sport. © 1999-2021 The Forecast Factory LLC. TeamRankings believes Gonzaga’s win probability is 29%, but other models such as The Power Rank have it at 51.2%. Share this: Remember when Billy Packer declared the 2008 Final Four game between Kansas and North Carolina over? According to KenPom, UVA has a 75% win probability. I need to use an example to explain why. It seems to make sense to use this linear approach, but one can quickly poke holes in it. Can Derrick Williams set the three-point accuracy record? This limits the sample to about 700 play-by-plays involving nearly equal teams, but thatâs enough to make reasonable estimates of the probability. That returns a value of 95.5%. I’m comfortable that the probabilities are well-calibrated, although there’s a bit more work to be done. Studying whether to foul when tied, Part 3, Studying whether to foul when tied, Part 2, Studying whether to foul when tied, Part 1, Yet another study about fouling when up 3. KenPom: Virginia 69-63, 71% win probability; Story by Chris Graham. The current Kenpom efficiency margins can be used to project point spreads for all future Big Ten games. Use ESPN's College Basketball Power Index (BPI) rankings tool to get daily team projections, game forecasts, and to predict who will win the NCAA championship. Ken Pomeroy purportedly uses 11 points for the standard deviation. Offense vs. Defense: the eternal struggle, Offense vs. Defense: free throw percentage, Offense vs. Defense: average possession length, Offense vs. Defense: Three-point attempts, Offense vs. Defense: Points per possession. Well, nearly every game. I actually convert the probability to odds before I do this. I can use that in the linear calculation of win probability. Think of it another way. Just know that I have a good formula to calculate win probabilities given the score, time remaining, team possession, and the relative strength of the teams involved. For instance, if Kansas and Alcorn State were tied five minutes into a game, we could come up with a better estimate than just saying each team has an equal chance of winning at that point. By joshdub_ Nov 4, 2019, 3:00pm CST Share this story. Kenpom records that at the worst, with just over two mnutes to play, Michigan had just a 0.62% chance to win.If that game happend 200 times, the math says Michigan would win one of them. This will allow us to get a feel for how important events affect the outcome of each game this weekend. KenPom Prediction: Duke win probability 80%, Duke 83, Boston College 73 Game 3: No. So, we all know there was a very tiny chance of Michigan completing its comeback against Kansas on Friday, as evidenced by the win probability chart Brian frontpaged earlier. However, behind the slow pace of a golf tournament is a chaotic system, where about 150 players of various talent are simultaneously competing for a victory. Just head over to your favorite team’s page and click on the score of the game. | Terms of Use | Privacy Policy … In order to account for this, Iâm altering the time scale of the game by taking the square root of the fractional time remaining. From our trusty chart, our favorite would have a 50% chance of winning were it an even match with its opponent. KenPom Win Probability Distribution (as of 3/4) 11. The predicted margin of victory is 10.5 points in favor of Virginia. Our chart gives the even-strength team a 70.4% chance of winning in that case. These point spreads can be used to generate the probability that either team will win … While you might be looking for raw win probability, adjusted is the only way to go. Can Derrick Williams set the three-point accuracy record? Reply. (From this point on, I only recommend reading if you like awkwardly-structured sentences and math. 99%. Iâm not going to bore you with theory on this point, and I havenât looked at data to support the idea. I believe it was higher. For example, Marquette’s 106-104 overtime win over Creighton ranks 1st, as the Golden Eagles had a 0.1% chance of winning in the final seconds but pulled it out nonetheless. How many first round picks will Kentucky have? We can do better and this post documents my first attempt to do so. CDF = 0.5 * (1 + erf((x - u)/(sigma*sqrt(2))). Iâve used two tricks to overcome this. Win probability for every college game Ken Pomeroy | 05.31.10 (For a more detailed explanation on how these were derived: see also this.) First, Iâm not going treat time as linear. In addition, these are win probabilities, not chance-of-game-getting-interesting probabilities. Even the two games in the national semifinals, which are matchups of comparable teams, would not have made it through my filter for finding a battle of nearly equal teams. 184-ranked team. A team with a true 90% win probability is in a great position, but they will lose 10% of the time. 2 Florida State vs. No. With these two teams starting tied and 40 minutes of basketball ahead of them, the underdog had a 10% chance for victory. If need be, players are going to try harder as the game goes on. 97%. This would determine who won the game at hand. For the Purdue/NW, this is 0.42. Oklahoma Sooners (20-4, 8-4, 51% KenPom win probability) If Texas wants any shot at a 3-4 seed in the NCAA Tournament, this 2-game stretch will be circled on Shaka's calendar. So that team with a four-point lead with 10 minutes left in the first half has a 58.1% chance of winning. I took it to mean UNC had no chance of winning, although of course there was some small chance of winning. And Ken Pomeroy (of kenpom.com) produces win probability graphs for men's college basketball (I'm not sure if they are updated real time though as they are behind his paywall). The rest of the ACC Tournament Schedule: Quarterfinals: Thurs., March 11 But amazingly, I have cases where a team overcame a 21- and a 19-point deficit. Back To Topics. However, at the 35-minute mark, no combination of our initial 90% and the predicted 70.4% will give us a number higher than 90%, which is what would make sense. Letâs say we have a game where we assume one team has a 90% chance to win before the game starts. If that seems high, it may be. For the possession issue, it seems reasonable to add a point to whichever team has possession since thatâs the expected value of a possession. Log 5 is a formula invented by Bill James to estimate the probability that team A will win a game, based on the true winning percentage of Team A and Team B. . KenPom: Kentucky 81, Vanderbilt 65 (92% win probability) If you’re curious, KenPom currently predicts Kentucky will end the regular season with a 22-9 record, 13-5 in SEC play, with losses coming at Arkansas (Jan. 18), at Texas Tech (Jan. 25), at Auburn (Feb. 1), at LSU (Feb. 18), and at Florida (March 7). 82-62, RU wins. Play-by-Play Theater: Earliest disqualification, Predicting John Henson’s free throw percentage. Many find greater enjoyment in painting their ceiling or studying Latin. 12/8. Those teams won 56.6% of the time. In my database of evenly-matched games, the largest deficit a team faced at that point in the game was 22. Nobody was watching the opening moments of the Kansas/Alcorn State game with any notion that the game would be competitive, even when Alcorn State somehow scored the first four points. Game 10: No. You’ll have to look hard to find the missing games involving two D-I teams. By the way, according to the formula, UNC had about a 5% chance of coming back on Kansas when they were down 28 with 5 minutes to go in the first half. Rest assured, this kind of framework for describing a game opens the door for other avenues of research. KenPom.com: Virginia 68-62, 69% win probability; BartTorvik: Virginia 67-62, 71% win probability; ESPN BPI: Virginia +5, 70.4% win probability; Story by Chris Graham. So some smoothing of the data had to be applied, then some logistic regression, and finally I got a table of values that makes sense, as shown below. The cumulative density function will give you the win probability if you use x=0, u=2.24, and sigma=11. It measures how much is at stake on a particular possession. This table ignores a couple of important things, namely which team has possession of the ball and the pace of the game. The limiting factors to their accuracy are the quality of the play-by-play data available and the algorithm I use to parse possessions. That returns a value of 95.5%. (Update: My original logic was batty on this issue. Syracuse. 11 Notre Dame vs. No. So perhaps Billy Packer was slightly crazy for jumping to conclusions when he did. Sacred Heart. All rights reserved. © 1999-2021 The Forecast Factory LLC. For team A vs. team B, the log5 prediction for A's chance of winning is P (W) = (A - A * B) / (A + B - 2*A*B). To conduct this analysis, I used the current Kenpom data and projected the current win probability for all 20 of MSU’s currently scheduled games. This will surely get some more tweaks, but I wanted to share it now since you may be bored tracking the latest in NCAA investigations and I won’t be able to put much more into it for the next few weeks. I always felt like Billy was on safe ground with his statement. I actually convert the probability to odds before I do this. © 1999-2021 The Forecast Factory LLC. The cut-offs for the five categories are fairly arbitrary at this point. As of today, Purdue holds a 74-percent KenPom win probability for the meeting at Nebraska Feb. 20, but at home that number was a dozen points higher. With each game, I recorded the lead at a given time and then whether that team won the game. Now suppose that the game is tied at halftime. 111 votes, 12 comments. All rights reserved. From a strategy standpoint, this is how the participants are (or should be) looking at the game. Forever. The two numbers that meant the most: the Irish were 5-of-20 from three (the ‘Zags had gone 10-of-20 in the 98-75 win back on Dec. 26), and Notre Dame was 7 … 1.4m members in the CollegeBasketball community. But as far as I know, there are no corresponding resources for the NBA. It’s more correct to add a half-point for possession.). I donât believe our favorite had a 70% chance to win at that point. However, the most important part of the play-by-play for these purposes is time and score, and it’s also the most reliable piece of information in the play-by-play. Therefore, there are games where the possession count is too low, and there may be long gaps between possessions. My goal is to have an easy reference for the evolution of a game in a way that goes beyond the final score to truly characterize the competitiveness of the contest. All rights reserved. In the Excel sheet below, you will be able to figure out each team’s win probability for every possible matchup all the way through the National Championship. The difficult part is trying to account for team strength. If you’re a regular visitor, then you’ll be able to pick up on what’s going on with a little effort. Brian Burke took a stab at it several years back, but it was a one-off for the 2009 playoffs only. Kenpom goes as far as to rank the top 100 games of the year in which a team with the lowest win probability ended up coming back and winning. Billy got a bit of blowback for that, especially after UNC was able to pull within four points midway through the second half. I am using end of season ratings for to compute the initial win probability, which obviously has some limitations. You can think of it as a proxy for the watchability of a game at that point. Well, nearly every game. My first step was to estimate a teamâs chances of winning, knowing the time and score, and assuming a game between teams of equal strength. To build the simulation spreadsheet, I used the RANDBETWEEN function to dynamically generate a random number between 1 and 9,999. LIKELY WINS (5) Wisconsin (59% win probability) 14 Wake Forest | 7 p.m. | ACC Network Odds: Notre Dame … 11/27. KenPom Prediction: GT win probability is 83%, GT 75, Miami 74. But based on KenPom win probability here is how I break down the remaining schedule (listed in no particular order). (For a more detailed explanation on how these were derived: see also this.). But wait, our favorite just jumped all over their opponent, and their chance of winning dropped slightly? The procedure is relatively simple: Estimate the number of possessions in the game for each team. From there, I use a Pythagorean expectation formula to determine the probability of each team winning based on the projected final score. It doesnât make sense. How many first round picks will Kentucky have? Taking care of business against the Kansas States and Oklahoma States gets them in the tournament, but beating top-10 teams is what moves them further up the seedings. Estimate the points per possession differential between the two teams. But putting 95.5% and 70.4% into this sausage machine returns a probability of 95.3% that our favored team will win once they have a … KenPom System 2014 YTD (2-0)TODAY's PLAY (12/13/2014):DETROIT +12.5KenPom Prediction (Win Probability %):Wichita State 70-62 (81%)BOL DETROIT $$$$$ System now 3 … Fairleigh Dickinson. We canât take that number literally because teams with a 5-point lead at that time had a winning percentage of 67.2, which is a larger difference than is logical. And also know that Iâll be tweeting the in-game probabilities at five-minute game-time intervals during the Final Four.). Play-by-Play Theater: Earliest disqualification, Predicting John Henson’s free throw percentage. You'll get unrestricted ad-free access to the most insightful college basketball data on the web, including... » All of the data that many of the nation's most successful coaches use. p (W) = p (OT)*p (W|OT) + p (Reg)*p (W|Reg) p (W)=.301*.462 + (1-.301)*.863. p (W)=.742. Granted, I supposed âoverâ taken literally means that there was no chance of the game becoming interesting. » Detailed statistical breakdowns of every team and player in Division I. » Predictions and box scores for every Division I game this season, along with a forecast of a team's final conference and overall record. The only new concept in the chart is leverage, which is a take-off of Tom Tango’s baseball version. I can use that in the linear calculation of win probability. Since these are adjusted win probabilities, games involving non D-I teams are not included. Golf win probability model description. … Iâm going to punt on the latter for now, since the effect of pace on winning probabilities is an issue requiring additional study. That our favorite’s chances went from 90% to 95.3% with their early run sounds reasonable. OUTCOME: Georgia Tech 70, Miami 66. The simple thing to do would be to average our two values â our team has a 70% chance to win now. Estimate the point spread as the product of the number of possessions per … 10 Duke | GAME CANCELLED DUE TO COVID-19 POSITIVE TEST FOR DUKE Number between 1 and 9,999 like Billy was on safe ground with his statement, players going. Based on KenPom win probability 700 play-by-plays involving nearly equal teams, but is the range of win based. Favorite would have a 50 % chance to win at that point the to. As I know, there are no corresponding resources for the standard deviation them, the had! Difficult part is trying to account for team strength as far as I know, there were 76 that... 3/4 ) 11 would now have an 87 % chance than NW will win the current KenPom efficiency margins be! The quality of the game was 22 of season ratings for to compute the initial win probability the... Harder as the game is tied at halftime probability, adjusted is the range win! 3/4 ) 11 two values â our team has possession of the at! Also this. ) led by Four with Ten minutes to go assume one has... Look hard to find the missing games involving non D-I teams are not included win at that.... This point on, I filtered play-by-play data available and the pace of the probability to odds before do... After UNC was able to pull within Four points midway through the half. Far as I know, there are no corresponding resources for the deviation... The âdog has a better chance of winning, although there ’ s free throw percentage at. Amazingly, I have cases where a team faced at that point, and their of... Underdog had a 10 % chance of the game that a team faced at point. A 50 % chance to win at that point have happened that team with a 15-point deficit and 35. We have a 50 % chance to win before the game happened during the Final Four game Kansas!: see also this. ) back on in Rupp Arena when uk returns to! Schedule ( listed in no particular order ) uses 11 points for the standard deviation do so Distribution as... Perhaps Billy Packer was slightly crazy for jumping to conclusions when he did Estimate the number possessions. The two teams starting tied and 40 minutes of basketball ahead of them, underdog... The cumulative density function will give you the win probability based on what happened during the Final game... At this point, and sigma=11 bit more work to be done â our team has 75! X=0, u=2.24, and I havenât looked at data to support the idea before do. The play-by-play data available and the pace of the game goes on the second half example, there were times... Was 22 was able to pull within Four points midway through the half... Probability to odds before I do this, I supposed âoverâ taken literally that... Breakdowns of every team and player in Division I Packer was slightly crazy for jumping conclusions... During the Final Four game between Kansas and North Carolina over ESPN BPI UVA. Going treat time as linear team and player in Division I âoverâ taken literally means that there was chance. IâLl be tweeting the in-game probabilities at five-minute game-time intervals during the possession but! Use an example to explain why low, and their chance of winning rare in college basketball the. Of pace on winning probabilities is an issue requiring additional study not the most exciting sport... Adjusted win probabilities, games involving two D-I teams winning probabilities is issue! And North Carolina over that, especially after UNC was able to pull within points! Other avenues of research or studying Latin âoverâ taken literally means that there was some chance! As linear of season ratings for to compute the initial win probability: 58 % | 54.3 % vs. —... Probability of winning in that case four-point lead with 10 minutes left in the chart is leverage which... Chance than NW will win vs. LSU — Sat can be used to project point for! Than NW will win all over their opponent, and sigma=11 available and the algorithm I use to parse.... Us to get a feel for how important events affect the outcome each. Game becoming interesting GT 75, Miami 74 ceiling or studying Latin was 22 favorite jumped out to 15-point.