by Ken Pomeroy on Saturday, May 10, 2014
Let’s face it, golf is not the most exciting spectator sport. Many find greater enjoyment in painting their ceiling or studying Latin. 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. But four rounds of golf is often not enough to separate the best golfers in the field from everyone else.
In fact, no matter the skill of the golfer, the winner will have to play better than his long-term average. Thus, the typical professional golf tournament is a tribute to randomness. One with expert knowledge could go an entire year without successfully picking a tournament winner and not feel too bad about it. Who is going to play over his head this week? That is the question that must be answered to predict a winner and no one can know that answer with much certainty.
In order to understand this messed-up world, I’ve been trying to develop a credible win probability model for golf tournaments for a while. This is not something that will solve one of sport’s great mysteries, but perhaps it will make more sense of the wacky world of professional golf, where each tournament contains 2.5 times the entries of the NCAA tournament with significantly more parity than college hoops.
by Ken Pomeroy on Monday, April 28, 2014
Ever want to relive the 2002 season? Now you can, as tempo-free data for the 2002 season has been posted. Thanks a bunch to Josh Steele for filling in the holes of my data set to make this possible. Here are five interesting things I noticed about that season, but I’m sure you can find more!
5) It was not that long ago, but it was a different game. The average game in ‘02 contained 69.5 possessions for each team. That figure would put a team around 50th in adjusted tempo in 2014. Kansas, for one, played just one game that had fewer than 70 possessions. The average ACC game had 74.2 possessions. This season, it had 61.8, which represents an 18 percent decline. There was also just more going on in college basketball back then. More offensive rebounds and turnovers, especially. Perhaps this is bordering on “the game was better in my day” talk. I’m not saying the game was better, but it was more fun. The data backs this up. (Except for blocked shots. There are more blocked shots now.)
4) The SEC was the best conference in America. This wasn’t the last time it would be rated as the top conference by my system - it’s happened as recently as 2006 - but I suspect the ACC got more notoriety by having two one-seeds. However, the bottom of the ACC was clearly worse than the bottom of the SEC. It was the season UNC lost 20 games and Florida State and Clemson were equally bad. The SEC on the other hand put all 12 of its teams in the top 100. That balanced hindered the top teams from putting up a great record and only one SEC team lost fewer than six conference games.
by Ken Pomeroy on Monday, April 21, 2014
Whose listed weight is the most misleading? You didn’t know you wanted to know that and I didn’t either, until I started trying to predict a player’s weight based on other information. And the only reason I did that was because sometimes a player’s weight is not revealed. Mostly in cases of players on Utah Valley’s roster as the Wolverines are the lone team in college hoops that treats its players’ weights like Bill Belichick treats his players’ injury information.
You might think it is really dumb to be bothered by such minutiae, and that’s fine, but you are not the intended audience of this site. Go to some other place and read “normal” things about “people” and stuff. In the real world, when one entity out of 351 refuses to do something the other 350 are doing, it can create a headache. I mean, that “effective weight” stat people have been demanding is not possible because of the Wolverines.
But there is no need to let Utah Valley hold us back from having a complete data set of players’ weights. One can look at available players’ weights and all of their accumulated statistics and figure out what matters when predicting what a player’s listed weight will be. And this produces a pretty decent estimate. By pretty decent, I mean a standard error of 13.5 pounds for the 2014 season by using data from the 2007 through 2013 seasons to build the model.
by Ken Pomeroy on Tuesday, April 8, 2014
Congratulations to the winner of the 2014 kenpom.com player of the year award, Russ Smith of Louisville. Smith earned his second consecutive kPOY by posting a 114 offensive rating while using 31 percent of the Cardinals’ possessions when he was on the floor. That was the fourth-highest offensive rating among players using at least 30 percent of their team’s possessions. In addition, he recorded 3.96 steals per 100 possessions, which ranked 32nd in the nation, while playing 29 minutes per game for the nation’s fourth-ranked defense. In 2014, no player combined offensive production and defensive impact like Russ Smith. Russ joins past winners Jared Sullinger (2011), Draymond Green (2012), and himself (2013).
by Ken Pomeroy on Monday, April 7, 2014
Among the things officials had to deal with this season were new interpretations (“hey defenders, stop touching ball-handlers”) and real changes to rules. There were 29 changes to the rules this season. A lot of them had to do with expanding the jurisdiction of replay or cracking down on certain uniform issues. My personal favorite was this one…
Dead ball. (Rule 6-6.3). A try in flight shall not become dead when a shooter’s teammate excessively swings his elbows without making contact.
If you’ve been at a game where that has ever come up, well, I’d call you a liar. I cannot imagine a scenario, except maybe a game of NBA JAM, where that has come up.
by Ken Pomeroy on Wednesday, March 26, 2014
San Diego State enters Thursday’s game with Arizona with what seems to be a remarkable streak. They’ve won 119 consecutive games when they have led with five minutes to go. The last time they blew a lead with five minutes left was on January 9, 2010. The Aztecs were up 14 with the ball with 4:47 to go at Wyoming and managed to lose 85-83. It was one of the more remarkable comebacks that season. That was one of three Mountain West wins for the Cowboys, who were pummeled by lowly Air Force in the play-in game of the conference tourney. San Diego State would go on to earn an 11-seed in the NCAA tournament.
It is an amazing streak, but winning the vast majority of the games where one leads with five minutes left is not terribly unusual. There are 30 other teams that have won at least 95 percent of such games over the past five seasons. The closest competitor to the Aztecs’ supremacy is Louisiana Tech, who has won 94 times in 95 chances during that span.
Of course, there’s probably a reason whoever researched that fact chose five minutes as the special time. In Arizona’s memorable 68-67 win over the Aztecs in the Diamond Head Classic last season, San Diego State led with 4:58 remaining, but the game was tied with five minutes left. That game effectively blocks using 4 minutes or 3 minutes as a cut-off. And the Aztecs blew a lead with six minutes to go against Air Force during the 2012 season. So specifying five minutes is critical here.
by Ken Pomeroy on Tuesday, March 25, 2014
I’ve tweeted about this a couple times, but it’s worth documenting in a more permanent location: This season, free throws were shot at a better rate than in any previous season in the history of college basketball. To date, my calculations indicate that D-I teams have made 69.82 percent of their attempts from the free throw line. According to the NCAA record book, the previous best was the 69.7 percent made in the 1979 season.
It’s an odd phenomenon, but analysts, journalists, and coaches appear to be programmed to bash fundamentals. That’s another subject deserving of its own article, but criticizing modern free-throw shooting has always been a dubious exercise within that realm. After all, that’s one fundamental we can measure, and free-throw percentage has essentially been constant for the last 50 years.
But my perception is that most people in the game feel like free-throw shooting was better way back when. And if people can’t get that right, one should be skeptical when other fundamentals are criticized. Are players really worse at setting screens, or scoring with their off hand than they were 30 years ago? I think it’s reasonable to wonder whether we’ve been lied to all along about those things as well.
by Ken Pomeroy on Sunday, March 23, 2014
Here’s the log5 for the Sweet 16…
Elite8 Final4 Final Champ 1 in Pvs 1S Florida 70.8% 58.9% 38.5% 22.6% 4 12.9% 1W Arizona 72.6 50.7 30.2 18.0 6 15.9 4MW Louisville 69.8 46.6 27.9 16.6 6 12.3 1E Virginia 63.0 43.3 24.0 12.6 8 11.5 11MW Tennessee 54.4 22.4 10.1 4.6 22 0.9 2W Wisconsin 58.3 23.0 9.8 4.2 24 2.3 4E Michigan St. 37.0 20.9 9.1 3.6 28 2.4 4S UCLA 29.2 19.5 8.7 3.3 31 1.3 2MW Michigan 45.6 16.9 6.9 2.8 35 2.1 3E Iowa St. 50.1 17.9 6.8 2.3 43 1.0 7E UConn 49.9 17.9 6.8 2.3 44 0.7 8MW Kentucky 30.2 14.1 5.7 2.3 44 0.7 4W San Diego St 27.4 12.9 4.7 1.7 57 1.0 6W Baylor 41.7 13.4 4.7 1.7 60 0.4 10S Stanford 56.1 13.1 4.0 1.0 99 0.1 11S Dayton 43.9 8.5 2.2 0.5 217 0.07
by Ken Pomeroy on Wednesday, March 19, 2014
Given Wichita State’s challenging draw, I thought I’d calculate the chances of the Shockers winning the tournament if they were given other seeds in other regions. Keep in mind their current situation gives them an 8.0% chance to win the title*.
Wichita State got the one-seed is in the Midwest region. What if the Shockers were seeded worse in the Midwest? Here are their log5-style chances of winning six games given a different location in the bracket:
by Ken Pomeroy on Monday, March 17, 2014
What follows is the master log5 for the 68-team field. At the risk of being lumped in with Rick Reilly, I will recycle the disclaimers from last year:
- The calculations below represent the chance of each team advancing to a specific round based on the log5 formula and each team’s pythagorean rating. Unfamiliar with log5? Here’s more than you want to know.
- This does not necessarily represent Ken Pomeroy’s opinion. There are subjective factors which this analysis does not include. Please use it responsibly.
- Don’t interpret these numbers as saying Ken Pomeroy’s computer is predicting Arizona will win the tournament. It’s saying there’s an 84% chance they won’t.
- If you doubt that seeding doesn’t matter (much) check out how similar the order of these teams is to the team’s ranking in my system. Louisville doesn’t suffer much as a four-seed. However, it’s a different story if you have to play extra game like Tennessee.
- No proximity bonus is included in any games here.