Monday, January 25, 2010
Developing Player Value: Shot Value per 20 Minutes (ShV/20)
Last week, I spent a considerable amount of time explaining Defensive Value per 20 Minutes, or DV/20, and applying the metric to last year's defensemen. I also found out that I tend to be watched closely by players that have historically sucked at the statistic, which is harrowing but I may be in the clear this go around. Who doesn't like shooting?
Uh, did anybody see that? It looked like Craig Ludwig. Okay, play it cool...
So when compiling these player values, I want to set it up so that we have a balanced set of four metrics, including two defensive and two offensive sets of numbers, so that when they are brought together we hold players to task for defensive deficiencies as well as offensive deficiencies. In this case, we are handling the defensemen player values, which are analyzed by player performance over 20 minutes (the approximate average time for a starting defenseman).
After handling a metric that favored defensive defensemen last week, I want to shift to a more offensive-oriented metric this go-around. And so we come to Shot Value per 20 Minutes, or ShV/20.
We can all agree that shots have value, to a point. They are a goal opportunity, and the more shots a player takes on net, the more opportunities there are to score (by straight shot, tip-ins, and rebounds). On the other hand, not all shots are created equal, and some hockey stat-heads have looked at shot value by taking account of the distance from which the shots are taken (scroll down a bit, it's there). The problem, as Gabriel Desjardins (of Behind the Net) notes, is that a lot of the leg-work would have to be done to get the data together (like, we're talking months of data collection, something I would not do without something to cover my exhorbitant Coca-Cola and Swiss Roll expenses).
For these reasons, we will work with shots on net and missed shots as our main stat totals, using the missed shots as a way to balance for those who either are a.) horribly inaccurate, and/or b.) taking high difficulty shots. First of all, for both categories I take the totals for that year and determine how many shots and missed shots were done per minute, then expand that back out to get an idea of the number of shots and missed shots per 20 minutes the player took (or would conceivably take).
This is where it gets dicey, and I'm open for suggestion. I debated for a while just how much a player should be docked for missed shots. Initially, I was going to make it a one-for-one, where missed shots completely subtract the value of a shot. Then I reconsidered, because a missed shot should not subtract the total value of a shot. For one, a missed shot could have been close and was missed for a tip-in, which puts it much closer to value. For two, a missed shot's negative value should not equal a shot's positive value, because it doesn't automatically result in something that would be equally negative, such as an opponent's shot on goal. So for those reasons, I cut the missed shots per 20 in half and subtracted it from shots per 20. And there's your ShV/20.
As mentioned before, I'm completely open for debate on this metric; it's possible missed shots could detract even less, or they could detract even more.
Now, to get back to the fun, here's how the NHL defensemen did last year. Over 234 eligible defensemen, the average ShV/20 was .96. As usual, I have the minimum games set at 20 to bench the gimps, goons, and rooks, and I'm digging on continuing to include games played and average time-on-ice:
NHL Top 5, 2008-09
1. Rob Blake, Cold Fish, 2.23 --- 73 GP, 21:16 ATOI
2. Mike Green, The Bandwagon, 2.23 --- 68 GP, 25:45 ATOI
3. Sheldon Souray, Slippery Slope, 2.00 --- 81 GP, 24:50 ATOI
4. Dion Phaneuf, Stampeders, 2.00 --- 80 GP, 26:31 ATOI
5. Shea Weber, Pre-Daters, 1.97 --- 81 GP, 23:58 ATOI
NHL Bottom 5, 2008-09
230. Kurt Sauer, Scrotes, .23 --- 68 GP, 20:37 ATOI
231. Sean O'Donnell, Queens, .22 --- 82 GP, 20:29 ATOI
232. Mike Lundin, Shockers, .22 --- 25 GP, 16:38 ATOI
233. Ryan Parent, Pumpkins, .21 --- 31 GP, 18:12 ATOI
234. Jay Leach, Lucifer, .14 --- 24 GP, 14:50 ATOI
A little bit of a Shawshank Redemption for Old Man Blake, who was in the NHL Bottom 5 for DV/20; it gives you a great idea of how he's being used (PP, primarily a shooter), and if you've ever seen him in a game, it's justified. Quite awkward on skates at this point, you completely forget about it when Blake winds up and lets it fly. Souray likewise redeemed himself, from bottom 5 to top 5. It really gives you an impression of these guy's abilities, and things I took for granted (Blake and Souray, good hitters, must be good defensively, right?) are looking a bit different.
You can definitely note that the top players are guys who get a lot of ice time (and PP time), but that fact does not erase the main point that they are creating goal opportunities on the ice, and with a higher percentage than others. In this metric, Mike Green has great value, and that's undeniable. You want to tell me he has no shot value? He's 155th in DV/20.
As for the bottom guys, a lot of part-timers, and Parent has been bit by the injury bug early in his career. O'Donnell is a mainstay at the bottom of this metric, and that's not unusual...
But I'm starting to think that these values are not helping O'Donnell much, as he was 192nd in DV/20. Slowing down a bit?
As we can see, ShV/20 will (and always will) favor the big shooters, but by taking into account misses we can shift down the guys who are spraying them all over. Chris Pronger, for instance, finished 8th in shots attempted, but 50th in ShV/20 (with a 1.30) for this reason.
And thankfully, it looks like Ludwig is going to leave me alone. True story, I once met him in Mel's Sports Store in Rhinelander, Wisconsin (he and I are from the Northwoods). Very big person, looks like he's in a biker gang...
Okay, I better leave. See you next time, when I get into the full Defensemen Player Value.