The kPOY: on the eve of conference play
by Ken Pomeroy on Wednesday, December 29, 2010
Let’s start this week’s kPOY update with an e-mail.
Ken,
Neat idea. Just curious if you were concerned about using raw totals not adjusted for the level of competition? I suppose the team strength brings that into play, but doesn’t Jimmer Fredette have some kind of advantage over most of the rest of the players by the end of the season if his conference is weaker?
Thanks,
Alan
There’s nothing more boring that writing formulas, and I can’t imagine it’s any fun for most of my audience to read them. But in its simplest form, the kPOY is computed from this…
player's internal value * team strength
Neither of these factors are directly related to the strength of the competition. The internal value of a player is relative to his teammates, so even if we adjust for competition it doesn’t change the equation. The goal is…
The kPOY: a history
by Ken Pomeroy on Wednesday, December 22, 2010
Response to the initial announcement of the kPOY has been overwhelming. I’ve received correspondence from as far away as Ohio and while there have been many different questions, the most common one involves who would have won the award in past seasons. We only have comprehensive and accurate tempo-free player stats going back to the 2005 season, so let’s start there and see if this thing has any credibility. (And also, feel free to cut and paste these standings into the Wikipedia page, whomever creates it.)
2005
1. Sean May, UNC. (.497)
2. Andrew Bogut, Utah (.482)
3. Shelden Williams, Duke. (.405)
4. Wayne Simien, Kansas. (.401)
5. JJ Redick, Duke. (.394)
6. Spencer Nelson, Utah State. (.394)
7. Taj Gray, Oklahoma. (.378)
8. Joey Graham, Oklahoma State. (.370)
9. Hakim Warrick, Syracuse. (.364)
…
The head-to-head fallacy
by Ken Pomeroy on Friday, December 17, 2010
Question - when Texas A&M beat Washington 63-62 in College Station last Saturday, were you convinced that the Aggies were the better team? If you are like most fans and media members, you were. However, if anything, that result proved the opposite.
Fortunately, there are plenty of rematches in the sport, so one can use real data to determine how much a head-to-head victory is a true verdict of superiority. Last season, there were 1,049 regular-season conference games which were a rematch of an earlier game. The winners of the initial game won 61.1% of the rematches. Not exactly a figure that inspires confidence that the outcome of a single game is useful to compare two teams. Keep that in mind as you make arguments about one team being better than another.
We should break this down further, though. Because we’re only looking at conference regular-season games in…
Introducing the kPOY
by Ken Pomeroy on Wednesday, December 15, 2010
The only frustration of following college hoops is that it’s impossible to follow everything. And I don’t like trusting The Man to tell me what I’ve missed, because The Man is missing stuff, too. That premise is the basis for this site, and it’s also the basis for America’s newest player of the year award.
In the effort to produce something more objective than currently exists, I’m introducing the first annual kenpom Player of the Year. Of course, it’s not completely objective - nothing is - but my intentions are pure. I’ve taken the work that I did to produce pre-season ratings and applied it here.
To refresh: In order to determine the impact of departing players, a player’s minutes were weighted by the amount of value they provided to the team. On offense, I used a combination a player’s offensive rating and his possessions used, which was valued…
Prediction tracker
by Ken Pomeroy on Thursday, December 2, 2010
Here at kenpom.com, we care deeply about how well our predictions do. It’s one thing to say you’ve picked 76% of winners, but that’s a meaningless statistic, especially in a sport with as many mismatches as college hoops.
That’s why this season we are tracking the performance of our forecasts. Eventually, we’ll use this information to calibrate the model a little better. I think people that follow this stuff closely understand that the percentages given to winning in my system are a little too aggressive. (For this reason, I water down the initial chance of winning in the win probability graphs.) But I won’t know how much until I aggregate the performance of predictions over the course of the season.
Here’s the breakdown on predictions through Tuesday’s games.
Prediction W-L Pct 50-55% 38-32 .543 56-60% 34-29 .540 61-65% 38-28 .576 66-70% 56-24 .700 71-75% 50-24 .676 76-80%…