Saturday, July 27, 2013

Ribeiro will be missed

One thing I've been meaning to do for awhile is look at how good a playmaker Mike Ribeiro is at even strength (along the lines of this post examining other top playmakers, like Sidney Crosby, Henrik Sedin, and Pavel Datsyuk).

If you look at Ribeiro's Behind the Net player card, you'll notice that he consistently has a very good on-ice shooting percentage. That does not mean that he is a good playmaker by itself, though: good players play with other good players, and good scorers can elevate shooting percentage though their own efforts (shooting skill, like Steven Stamkos and Ilya Kovalchuk, or getting to better locations, like Alexander Semin with his little toe-drag before every wrist shot). But Ribeiro in particular doesn't really shoot or score enough to elevate that on-ice shooting percentage himself, and he's usually been a second-line center, so it's likely that he does have some significant playmaking skill that elevates his on-ice shooting percentage to top-line level. The question is, how much, and how much will the Capitals suffer as a result.

To estimate his playmaking talent without going into shot locations, I first made a list of every skater who has spent at least 200 5v5 minutes with Ribeiro. (I only used players who played a full season on the same NHL team as Ribeiro, to avoid situations like James Neal going to Pittsburgh.) I compared the 5v5 shooting percentages with Ribeiro and without him using the tools at Hockey Analysis. (Ideally we look at every player who has taken a shot or missed shot and compare Fenwick shooting percentages, but that's more work than i want to do right now).

You can check out the full list in the spreadsheet here. Long story short, only Alex Ovechkin, Mike Green, James Neal, Michael Ryder, Karl Alzner, and Trevor Daley saw their individual shooting percentage fall with Ribeiro on the ice, and only Daley has more than 100 shots both with and without Ribeiro. (One hundred shots still isn't that much.)

For good measure, approximating this sample of shots as a binomial variable*, we can be pretty confident Ribeiro is a substantially above average playmaker. The players shot 9.2% with Ribeiro on about 1800 shots, compared to 8.2% on about 3000 shots without. A one point bump in shooting percentage is obviously an extra goal per 100 shots. To translate that to possession indicators, that's roughly equivalent to going from 50% Fenwick to 56% Fenwick, or a goal every seven games or so at 14.0 5v5 on-ice shots per game, Ribeiro's rate this past season per BtN.

*If you're unfamiliar with the terminology, this means approximating shots on goal as independent events, with each having an equal chance of going in. That's obviously not true, but it's much simpler and faster than the alternatives. The probability I'm using here is the total shooting percentage of Ribeiro's frequent teammates without Ribeiro on the ice.

Because I limited the sample arbitrarily, I probably introduced some biases into the sample. As a group, forwards take more shots per unit ice time than defensemen, so I probably cut off some forwards who would be as relevant to this sample as the defensemen. That means that the average shooting percentage I'm using as the mean in the binomial calculations may be a bit high—after all, forwards shoot for a better percentage than defensemen. But that would only make Ribeiro look even better than he does now.

For a point of comparison, here are the estimated playmaking bumps from the article linked above (which, it should be noted, using linemates' total shooting percentage, not merely shooting percentage without the player in question):

H. Sedin: 9.6% to 10.8%
Crosby: 8.5% to 10.6%
Malkin: 9.1% to 10.3%
Ovechkin: 9.1% to 9.9%
Backstrom: 9.4% to 10.1%

(And in case you were wondering, Alexander Semin looks like a pretty good playmaker, too.)

Even though those differences look significant, keep in mind that it takes six goals for a marginal win, and these guys are on the ice for only a few hundred 5v5 shots a season. I'd peg Ribeiro's playmaking talent at worth about one win a year.

And two more to note, Ribeiro's replacements:

Perreault: 6.0% with (~450 shots), 7.4% without (~2000 shots)
Laich (11-12 only): 8.4% with (275 shots), 6.2% without (~600 shots)

(The reason I use only 2011-12 for Laich is because that's the only year he played center full-time. For a bigger picture, one could look at shots with Laich on ice without one of the Caps' full-time centers: Backstrom, Fedorov, Morrison, Gordon, Arnott, etc)

There's roughly a 30% chance Perreault looks that bad on luck alone—less than half a standard deviation away from the mean—and about 20% Laich looks that good based on luck alone (again, assuming a bunch of things about individual shots). For a better estimate, you could regress the shooting percentages to league average based on the number of shots. (We should be in the other cases, too, but their samples are so big that—I hope—the difference would be minimal.) With that many on-ice shots, we can't really estimate a significant playmaking impact for either of these players, which means that, they need to be significantly better than Ribeiro possession-wise to adequately replace him at five-aside.

With this in mind, here's my take on the 2C situation:

1) The (5on4) power play, with Ribeiro on the ice, shot over 25%. That's not sustainable—only Philadelphia has broken 17% shooting at 5-on-4 in a full season since 07-08. With or without Ribeiro, a power play drop off is inevitable.

2) Ribeiro's value is closely linked to the power play (where he scored over half his points). Power play regression = Ribeiro regression. When the power play is worse next year, there will be a lot of unnecessary angst over it, when a substantial part of that is due to worse-than-awesome puck luck.

3) Ribeiro has been a legitimately strong player at 5v4 for years, so losing him will lead to a legitimately worse power play, all else equal. But I don't think value of an elite power play threat is nearly as big as it's made out to be—the estimates I've seen peg special teams talent as around 10-15% of team success moving forward, about a third or a fourth of even strength shot differential. There is no adequate replacement to offset that much puck luck, and it's not worth crediting Ribeiro with that puck luck. Long story short, the Capitals lost one really good power play player, and one player does not a power play make.

4) Ribeiro isn't a good Fenwick player, so part of his playmaking skills at 5v5 go toward offsetting his negative shot differential. That being said, he's probably not as bad as the numbers indicated this season (-10.3 Corsi Rel) because he spent a lot of time with guys who don't "drive possession" (e.g. Brouwer) or weren't "driving possession" at the time (e.g. Ovechkin). Give him good linemates and he'll come close to breaking even possession-wise. That makes him worth one win a year from his even-strength playmaking, and a few extra goals from the power play. But as a 33-year-old, that's as good as it will get, and it could fall off sharply any year.

5) The Capitals have a potential replacement in Mathieu Perreault who is a far better possession player (career 54.1% Corsi) but hasn't proven himself to be so much better than the possession advantage completely offsets the playmaking advantage. (Again, we're looking for at least 56% Corsi or Fenwick if Ribeiro can break even in Corsi with reasonably good linemates, and plus, Perreault has played more sheltered minutes.) That being said, Perreault also helps his own cause with personal shooting percentage more than Ribeiro (although he doesn't shoot more on a per-minute basis: at 5v5, 5.3 shots/60 for Perreault, 5.8 for Ribeiro).

6) A player like Perreault moving up the lineup creates a hole in the bottom six. Move someone else to 3C (like Laich) and you've created another hole. Eventually you get to the point where Ribeiro is better than a replacement, so the Caps could be a worse team today than they were in April. Not much worse—and we haven't even considered better health from Brooks Laich and a full season from Martin Erat, which could more than compensate for Ribeiro's loss—but as much as I bang the "Ribeiro is overrated" drum, I don't think the team is better off at even strength without him, all else equal. In other words, this isn't addition by subtraction.

7) Grabovski, who is decent on the power play and quite good at even strength, was the best solution at 2C on July 5, and still is today. Go get him, GMGM.

2 comments:

  1. They are luxurious re-defined. Cheap Taps UK could be the inadequate bit in your jigsaw problem of one's bathroom. Fundamentally receiving a Waterfall Taps person must not be adequate, it's necessary that you simply know what sort of Bathroom Sink Taps appliances fit your Taps Sale. That will not solely necessarily mean matching them while using tiles and ground of the lavatory, on the Bathroom Accessory hand coordinating his or her being compatible with your living Kitchen Taps with regard to normal water pressure and a few different www.uktaps.co.uk.

    ReplyDelete
  2. There are various Taps Sale accessories and with the passage of time, these Basin Taps are being improved. Nowadays the most Bathroom Sink Taps used kitchen faucets are the one with sensors. The Bathtub Taps with sensor have some significant features making them an appropriate choice for any Antique Kitchen Taps. The kitchen faucets with sensor run on a switch function. This switch is for hot and Automatic Taps.

    ReplyDelete