How much is air worth?
01.24.12
Wyoming hosts San Diego State tonight (if the Aztecs make it to Laramie) and it got me to wondering about the importance of altitude. The Aztecs won at New Mexico last week, and while observers were impressed with the win in the Lobos’ building, nobody really mentioned that they had to overcome the thin air of Albuquerque (elevation 5300 feet) as well. There are a few studies out there that have looked at the difference in a high-elevation team’s winning percentage between home and road games. For some reason, I’d rather know how many points altitude is worth.
Wyoming has played San Diego State home and away every season since 2000. In 12 games in Wyoming, the average margin as been +1.0 for Wyoming while the games at San Diego have produced an average margin of +9.4 for the Aztecs. The difference of 10.4 points can’t be explained by raw home court advantage alone, which has long been calculated to be around four points. There are 2.4 points left unaccounted for. This is only a 13 game sample, though, and the error bars are large enough so that we can’t get too carried away about the results.
A bigger sample is available if we look at the University of Colorado during its time in the Big 12. For 15 consecutive seasons the Buffaloes played a home and away series with Kansas, Kansas State, Missouri, Iowa State, and Nebraska. In 75 games at Boulder over that span, the average margin was Colorado +1.5 and in the 75 games at lower elevation, Colorado’s average margin was -13.0, for a whopping 14.5 point difference, leaving 6.5 points unexplained.
Six points seems a bit much to account for altitude but there’s another possibility to consider as well. Teams residing at altitude might have an increased disadvantage when they go to lower locations. I’m not a runner and I don’t completely understand the concept of “living high and training low” but it seems to be accepted as an optimum way for endurance runners to train.
Obviously, teams living at high elevation can’t take advantage of this because they also train at high elevation. This might explain the additional disadvantage for these teams when they play at lower elevations. Perhaps higher elevation teams are in poorer shape than their lower elevation counterparts, an effect that is overwhelmed when the lower elevation team is suddenly deprived of its usual oxygen, but appears at sea-level when both teams are on equal footing air-wise.
Whatever the reason, there almost surely is an altitude effect that low-elevation teams have to suffer through when playing in thin air. If San Diego State wins tonight, I expect they won’t get as much credit as they deserve. Wyoming is a better team than RPI believers, or those that rate teams solely based on quality wins, would think. But also, there’s the altitude issue. Throw in the travel problems as well and it’s not unreasonable to think that the Aztecs are dealing with something approaching a double home-court advantage tonight. Keep that in mind if they pull out an ugly victory.
The untrained eye: New Mexico vs. UNLV
01.22.12
UNLV beat New Mexico 80-63 in a 64-possession game Saturday night. I was there. This is what I saw.
Mike Moser is great offensive rebounder considering his shot selection.
I don’t think there’s a useful position descriptor for UNLV’s Mike Moser. Whatever you want to call him, he was a busy man on Saturday night, taking 16 shots, grabbing 13 boards and recording two blocks. It’s not Moser’s activity that makes him unique, though. It’s that his shot selection skews towards the perimeter, yet he’s a major factor on the offensive glass. This is best explained in graph form.
What you see here is the distribution of D-I players that have played at least 40% of their team’s minutes this season, according to their OR% and 3PA%. Two players have separated themselves from the cloud, er, crowd. Moser and Belmont’s Brandon Baker. Even though Baker takes a higher percentage of his shots from three-point range, he takes many fewer shots than Moser in general (Baker’s %Shots is 16.2 to Moser’s 25.6), and thus has a little more time to hang out near the rim than Moser does.
Carlos Lopez is no longer a minutes-thief
I was not a big fan of UNLV center Carlos Lopez last season mainly because I felt like he was taking minutes that rightfully belonged to man-child Quintrell Thomas. But this year you could make the case that Thomas is stealing Lopez’s minutes. If it weren’t for his part-time status, Lopez would be getting more credit for an amazing turnaround in his game. He’s gone from an ORtg/%Poss combo of 96/23 as a freshman to 126/21 as a sophomore and has made 73% of his twos so far. He didn’t disappoint in this game, either, getting to the free throw line 14 times in 14 minutes of action. Free Carlos!
Where has Steve Alford’s defense gone?
Back during Monte Carlo sim-time, I mused about Alford finally channeling the magic of the top-ranked ‘06 Iowa defense again. But the last two games have done a lot of damage to the Lobos’ defensive rating, and for good reason. The 1.25 points per possession surrendered to UNLV was the sixth-best offensive performance for the Rebels this season, with the top five occurring against teams outside the top 150. San Diego State has its third-best offensive game of the season (1.26 PPP) against New Mexico last Wednesday.
And the Aztecs’ two best games occurred against a non-D-I and 271st UC Riverside. You can’t single out just one weakness for New Mexico’s decline. The Lobos can’t stop dribble penetration, and the breakdowns leave the perimeter wide open too often. But also, whomever Alford plugs into the 3 and 4 spots are atrocious on the defensive glass. It’s essentially up to Drew Gordon to prevent second-chances, and as good as Gordon is at rebounding (tenth in the nation in DR%), he needs help.
The 2012 kPOY watch list
01.20.12
It’s time to release the first leaderboard of the nation’s premier statistical individual award. For those new to these parts, the kenpom.com Player of the Year is based solely on statistical data. It’s safe to say that “player of the year” means different things to different people. But I’ve tried to define in the way it’s assumed to be defined for other more famous awards (Wooden, Naismith, etc.).
I hope we can agree that “player of the year” is not exactly “the best player in the country”. If you put Thomas Robinson on Houston Baptist, he’d be the same player but would have no shot at winning any player of the year awards. So much like other POY awards, the kPOY combines personal achievement and team strength. “Player of the year” for these purposes is essentially “the player that means the most to a very good team”. (If it was defined as the player that means the most to any team, Quincy Roberts would be the runaway winner, and we wouldn’t that.)
The kPOY formula is top secret, but it isn’t arbitrary. At least not totally. The calculations are based on the research done to produce pre-season team ratings, where player minutes are weighted according to the player’s value to his team. The more arbitrary part of the formula is the team influence. An off-season meeting of the kPOY selection committee determined that the quality of a player’s team should have slightly less influence in the formula and the calculations have been tweaked accordingly for this season. More details on the kPOY are in last season’s introductory post.
Note that this award is not designed to predict or otherwise mimic the Wooden or Naismith awards, other than to maintain a similar definition for identifying the winner. While voters in other awards generally look to per-game stats and ignore turnovers, this award looks at aggregate production based on advanced metrics. This award also includes defensive contributions, although we all know that box score data limits us somewhat in identifying those. However, by including team strength, the quality of a team’s defense is also a part of the kPOY mix, thus the influence of a player’s defense in this award is more robust than the human-derived awards. I think the inclusion of these additional factors make the kPOY better at identifying college basketball’s player of the year.
Eventually, I’ll get the kPOY standings hard-wired into the site so that we can see the list change on a daily basis. In the meantime, here are the top ten players in the kPOY standings through Wednesday’s action:
1. Thomas Robinson, Kansas (kPOY rating: .558): Robinson has been the most dominant player in the country on a team that has Final Four potential. Hard to argue he doesn’t deserve to be the frontrunner.
2. Jared Sullinger, Ohio St. (.476): Any criticism of Sullinger having a disappointing sophomore season is unfounded. He’s been as productive as freshman Sully, and in an offense that’s not as efficient as last season.
3. Draymond Green, Michigan St. (.453): Green’s numbers haven’t changed much from last season - he still does a variety of things very well, even if he does nothing great - but the Spartans appear to be a much better team than last year’s edition, which makes Green’s candidacy more relevant.
4. Mike Scott, Virginia (.446): You need no more evidence of the ongoing scourge of pace discrimination than the ESPN.com POY straw poll in which Scott received the same amount of votes as Julian Mavunga. #EndPacism
5. Anthony Davis, Kentucky (.422): The only freshman on the list, Davis is here mostly due to his defense, but he’s not one dimensional. He’s making 69% of his two-point attempts while rarely turning the ball over.
6. Kevin Jones, West Virginia (.420): Jones has always produced stunningly low turnover numbers, but this season he’s combining that skill with a 63% conversion rate on his twos. Oddly, he’s never been able to get to the free throw line very often during his career.
7. Jordan Taylor, Wisconsin (.411): It’s in the kPOY bylaws that there must be a representative from Wisconsin in the top ten. Taylor’s doing an under-appreciated job propping up what would otherwise be an ineffective offense.
8. Will Barton, Memphis (.376): Barton the Elder is flying under the POY radar, but he’s made massive improvements in his game as a sophomore and is the hands-down go-to-guy for the Tigers.
9. Doug McDermott, Creighton (.364): McDermott is basically the Valley’s Derrick Williams. His shooting percentages are off the charts despite a heavy workload and he’s a solid rebounder. Williams got essentially no POY consideration last season, but McDermott is high on most people’s lists right now.
10. Marcus Denmon, Missouri (.351): Denmon’s game is not quite that of a spot-up shooter, but his stats work out that way. You have to be damn good at shooting to make this list when that’s your main skill.