Having trouble getting registered or subscribing? Email us at info@kysportsreport.com or Private Message CitizenBBN and we'll get you set up!


Platoon system shows advantage of tempo-free stats

Rate this Entry
This is something I've been thinking about since I saw the pre-season all-conference and all-American teams come out and saw what is in my opinion an under-representation of UK players on these lists. So I just decided to do what is probably a too-long, very geeky article and get it out once and for all.

A few years back I bought into the idea of tempo-free stats as espoused by Dean Oliver, Ken Pomeroy and others being a preferable method of gauging performance than raw numbers.

For example, I believe their work shows that looking at points per possession (PPP) is a better way to judge a team's offensive or defensive performance than points per game. A team can be playing at a slow pace and putting up low points/game while still making the most of each possession. It's a way to take tempo out of the equation and look at teams' efficiency (PPP) on an apples-to-apples basis. And John Gasaway has written several screeds on why rebounding % is a preferable statistic to rebounding margin.

This year's UK team shows the benefit of looking at teams (and players) from a tempo-free perspective. Take last night's Columbia game, for example. No citizen of Big Blue Nation will disagree with the opinion that UK did not play particularly well last night. Falling behind early 11-0 and shooting only .118 from 3, UK looked uninterested for most of the 1st half, only to turn it on when it got important. UK only scored 56 points; obviously a poor offensive performance, right?

Well, not necessarily. Keep in mind that Columbia slowed the pace to "watching grass grow" proportions; the game only had 51 possessions, UK's slowest game of the year. Scoring 56 points in only 51 possessions means that UK scored 1.098 PPP, well above the NCAA average of 0.996 PPP.

At the risk of being pedantic (and isn't using the word "pedantic" being pedantic in and of itself?), let's look at it this way: UK scored 56 points in defeating Columbia last night, and 71 points in the win over Buffalo. If one were judging offensive performance solely by points per game, one would assume the Buffalo game was the better offensive performance. And one would be wrong, because UK had 15 more possessions to score those points in the Buffalo game. Looking at if from an apples-to-apples (point per possession) standpoint, UK's offensive performance against Columbia (1.10 PPP) was better than its performance against Buffalo (1.07 PPP).

Similarly, UK held Columbia to 46 points and Texas broke the 50-point barrier, scoring 51 vs. UK. The defensive performance vs. Columbia was clearly superior, right? Not from a point-per-possession standpoint, because again, there were more possessions in the Texas game (67)than the Columbia game (51). When you take the points these 2 UK opponents scored and divide by the total possessions*, you see that Texas scored only 0.76 PPP while Columbia scored 0.90 PPP, the best by a UK opponent this season.

So all of that was discussing team performance. What about individual performance? It can also apply there. We're already familiar with it when it comes to shooting. We don't look at the raw numbers of baskets players make to judge their shooting ability; we look at their Field Goal percentage.

Taking that approach with other aspects of individual performance also gives insight into how they're performing. That's especially important when players are getting relatively few minutes, as is the case with UK due to the platoon system.

Take Karl-Anthony Towns and blocks, for instance. Unless something changes radically, Towns is not going to come close to breaking Anthony Davis' season record for total blocks. Towns isn't playing enough; due to the platoon system, he's only playing about 19 minutes/game. Through 10 games, he's on pace for about 112 blocks, which would put him more than 70 short of Davis' 186 blocks.

Looking at it from a per-game standpoint gives one a slightly better way of gauging his performance. So far he's blocking 2.8/game. That's obviously very good, but again it puts him well behind Davis' single-season record of 4.65/game. So maybe Towns just isn't having as good a shot-blocking season as Davis, right?

Not necessarily. You have to look at the number of possessions each played. As a result of the platoon system, Towns is playing many fewer possessions/game (47%) than Davis played (80%). That will skew the per-game stats.

The tempo-free way of looking at blocks is Block%: the % of an opponent's 2pt Field Goal attempts a player blocks while the player is in the game. Using that metric, Towns is having a superior season to Davis: so far, Towns is blocking 17.9% of opponents' 2pt FGAs when he's in the game, significantly better than Davis' 13.8% (a record for UK players back to 2001-02, the first season Pomeroy compiles tempo-free stats).

Another example: rebounds. If one were judging solely by rebounds/game, one would look at Dakari Johnson's 6.1 rebounds/game and say he's having an OK-but-not-great season. And that would not give him the credit he's due. Johnson's only playing about 18 minutes/game due to the platoon system. When one looks at the total % of rebounds he's getting when he's in the game, it paints a different picture, especially on the offensive glass. Johnson's currently grabbing 18.2% of UK's misses when he's in the game. That's the 12th-best Offensive Rebound% in the country according to Pomeroy and 2nd-best by a UK player** in the KenPom Era.

There's a very good chance that players like Towns and Johnson may be underrepresented on post-season all-Conference and all-American teams because the voters won't see the eye-popping per-game numbers they tend to (over)emphasize. You can bet the pros aren't missing the value of those performances and won't undervalue them at draft time, because the pros generally dig deeper in evaluating performance.

If you want to track this yourself during the season, KenPom's site is excellent; even though most of the good stuff is behind a paywall ($20 annually), it's money well-spent, IMHO. Another resource is Sports Reference.com's college basketball section. It may not be as user-friendly as Pomeroy's, but it still tracks tempo-free stats nationally, by team and by player.

*--If you ever want to track possessions during the game, the formula used by Pomeroy is
(.475 * FTAs) + FGAs - Offensive Rebounds + Turnovers
Do that for each team, add them up and divide by 2.
Yes, it's kind of a pain to do by hand, but knowledge is its own reward! Or you can follow me on Twitter (@KSRBEvans), where I live-tweet this stuff during most UK games.

**--DeMarcus Cousins put up a 19.6% Offensive Rebound% in 2009-10.

Submit "Platoon system shows advantage of tempo-free stats" to Google Submit "Platoon system shows advantage of tempo-free stats" to Facebook

Updated 12-11-2014 at 10:02 AM by KSRBEvans

Tags: None Add / Edit Tags