Thursday, October 08, 2009

Game 2: Uneven Strength

[N.B.: As I live on the opposite side of the planet from virtually all NHL games, my comments will routinely be lagged a day or two. In the future I hope to keep the intervals shorter. In the meantime… well, there are plenty of more current blogs out there.]

There are two kinds of hockey people: people who like even strength, and people who like power plays. Even strength people are the sensible ones, the ones who care about objective value and replicable results, the accountants and scientists, who love practical shoes, bulk groceries, and responsible two-way players on a value contracts. Power play people, on the other hand, go in for drama and artistry, myths and intangibles, and like their wine expensive, their cats Persian, and their players insouciant, one-dimensional divas.

Okay, that’s bullshit, just as every phrase that begins with ‘there are two kinds of people…’ is bullshit. But there is some faint truth in it. Imagine, for a moment, a known power play specialist. Pick anyone you like, forward or defenseman, old or young, friend or foe, so long as it’s someone who gets a high percentage of his points with a man or two advantage. Now, I don’t know you, but I can bet your feelings about this dude follow one of two trajectories: either you think he’s a virtuoso who can do amazing things with a little extra space, or you think he’s overrated and probably overpaid, and much less talented than he appears.

The core issue is that power play ice time is sweet, sweet ice time, and it’s built into the game. One of hockey’s special features is the idea that rule violations are routine and unremarkable. There aren’t any real ‘punishments’ in hockey, none of the standard penalties are considered especially bad in the moral sense, and there’s no burning desire to purge them from the game or humiliate the perpetrators. Rather, it’s a compensatory regime, wherein the perceived unfairness of a given action is atoned by revisiting a greater injustice on the offending team. Every team, every game, no matter how disciplined or tender-hearted, expects to take some penalties.

Power play time is therefore guaranteed. In fact, rather a lot of power play time is guaranteed. In the last season (i.e. 2008-2009, from which are drawn all subsequent numbers as well), total PP opportunities ranged from 307 (Devils) to 374 (Canadiens), meaning most teams average something like four power plays in a game. And there is not a team yet constructed that doesn’t benefit from this advantage. Every team, without exception, gets 2-3 goals per hour of ES play time (the range was 2.0 (Avs, Isles Kings) to 2.9 (Wings, Pens, Flames) last season). Even the worst power play (NYR, anyone?) is nearly twice as effective per minute. A really good PP, an excellent PP, can be more than four times as effective as the typical ES scoring rate (Capitals, 9.9 GF/60, 5 on 4).

Look at it another way: if the best PP% is 25.5%, and the most PP opportunities one might get is 374, that’s 95 goals. Worst case scenario, 12.7% of 307, 38 goals. The real numbers don’t fall too far from that: the most PP goals (all advantages) last season was 91, the fewest, 41. The difference between a holy PP and a hellish one is perhaps 50 goals, between the middling-sad and middling-happy varieties, 20-and-change. And unlike ES markers, there’s significantly less trade-off in power play offense- that is, an all-out pursuit of the goal doesn’t carry as much risk of allowing goals against.

For players, we’re talking about the glory minutes of a game, forgiving of mistakes, indulgent to the ego, and quite likely to pad the numbers. It’s certainly easier to be an offensive terror on the PP than at ES; on the other hand, the expectations are significantly higher. The question isn’t whether a PP specialist is as good a hockey player intrinsically as an ES star. If we could leave aside the quasi-ethical proposition that it is ‘better’ to be good at ES than good on the PP, we might come to face the more significant issue: what proportion of team success is won on the PP.

In most cases, the two facets of the game fall in the same range- that is, teams who are good at ES tend to have good-ish PP rates, teams that blow at ES have worse ones. But sometimes the PP makes a major difference. The Sharks last season, while admirably stingy in their defense, had a fairly depressing 131 ES goals, putting them in the same range of ES scoring prowess as the Lightning and the Coyotes (yeah, I’m only considering goals-for at the moment, and yeah, I do realize that the whole GA thing is important also. That’s another post for another time). However, 87 PP goals, 8.6 per 60, was enough to put them in the upper tier of the NHL and buzz them, for a time, as one of the best teams in the League. Conversely, the Blue Jackets were just about as effective at ES scoring as the Capitals, but their bottom-feeding PP ensured a significantly less impressive outcome.

All this is a fancy, numerically-sub-literate way of trying to persuade you, dear remaining reader, of one thing: the power play is important. It’s not extra-credit points, it’s not free goals, it’s not the marshmallow topping on the sweet potato casserole that is ES play. It’s the greatest opportunity hockey offers, where the difference between good and bad might be 25% of your overall goals. The difference between winning and losing. The difference between the postseason and golf.

In Buffalo, the Canadiens played a game with nineteen penalties, nearly half the thing uneven. They won, of course, but by a hair, in OT, after scrabbling out a grinding, miserable horror of a game. For a couple of seasons, there was in Montreal a superior power play, the sort of thing one might lust for and linger over, the kind that alchemically turned the careless hooks of rookies and the ill-considered cross-checks of goons into standings points. We weren’t enough at ES to support those results, but year after year we let our PP specialists go, figuring them overpriced and unnecessary. I wonder, now, if they really were. That power play, that excellent one, might only have been worth another 18 goals, maybe twenty, nothing more than the extra goal every four games suggested by the averages, but games like this it would steal easily. With Markov out for so long one might as well forget his existence (it’ll just seem that much more special when he magically appears in February!), the selection of especially useful PP players has dwindled beyond dwindling, and that is cause for concern. A good PP is not necessarily as difficult to find or as expensive to purchase as good ES. It’s less mysterious, less given to chance, and the agents of its success are more easily identified. It might be the simplest way to make a team better.


Ms. Conduct said...

What of people who love the Penalty Kill? :)

E said...

an eccentric, masochistic minority!

Ms. Conduct said...

Guilty as charged. :) Figures I'm a goalie, too...

Jeff J said...

Hilarious opening.

"...what proportion of team success is won on the PP[?]"

Going by goals, two thirds of the game is EV and one third is special teams. Really, EV should account for more than two thirds because PP opportunities result from EV play.

I think most of the special teams game is driven by the PP. If you were to replace your best PP guys with your worst and your best PK guys with your worst, I'm pretty sure your PP would see the bigger drop-off in effectiveness.

If you peg the EV/PP/PK importance breakdown at 70/15/15, then 1/5 of a standard deviation improvement EV is almost equivalent to a full standard deviation improvement on the PP.

o The number of PPs has been dropping since the *New NHL* was introduced in 05/06.
o PP performance is unpredictable. The smaller sample size makes for streakiness. A player's true value is harder to pin down.
o The fraction of ST play diminishes in the playoffs.

"[The PP] might be the simplest way to make a team better."

This is really the heart of the matter, and it's the reason M-A Bergeron has a job right now. Realistically, it depends on circumstances and holes in your roster, but by and large your statement is true.

Take two point-per-game forwards: player A scored 35 EV points and 47 PP points, while player B put up 58/24. Who do you choose to build your team around?

The correct answer is B, because it's easier to make up 23 PP points with the rest of your lineup than 23 EV points. PP replacement level is way higher than EV replacement level.

For the curious: Player A is 07/08 Kovalev, player B is 07/08 Gaborik.

MathMan said...

Besides, Cammalleri has been a Kovalev-level PP performer. The Habs have let go their PP specialists, but unlike last year, they went in and imported replacements that have the possibility of matching the value of the old guys. Having Markov would help (oh, would it ever, ever help) but the roster does have guys that traditionally are good on the PP.

It's just that a successful PP like those the Habs have been graced with requires tactical cohesion that the barely acquainted Habs are as yet incapable of. I think it will come.