Feb 03, 2013; Montreal, Quebec, Canada; Montreal Canadiens left wing Erik Cole (72) celebrates his goal against Ottawa Senators goalie Craig Anderson (41) with teammates defenseman Francis Bouillon (55) and center David Desharnais (51) during the first period at the Bell Centre. Mandatory Credit: Jean-Yves Ahern-USA TODAY Sports

Some early numbers on the Habs


The old adage says that “it’s better to be lucky than good”. In the context of hockey, that may well be true, but there’s something that’s one better: to be lucky and good. That’s been the story of the 6-2-0 Habs so far.

Hockey has a great deal of factors that are outside of the players’ control and a great deal more that are, but that players can’t reliably reproduce. No NHL team wins 3/4ths of its games over any kind of stretch without a generous helping of those transient factors, whether one calls them “luck”, “temporary peaks of performance”, or by another name. One of the advanced statistics that measures a big part of this is PDO, which is simply the shooting percentage and save percentage of a team added together. Teams have virtually no ability to reliably affect their shooting percentage in the long run. Teams do have the ability to affect their save percentage, largely due to goaltending talent, but the effects of goalies in the long run is rather small. Therefore, over a large sequence of games teams will tend to end up with PDOs very close to 100. (Here is a more elaborate explanation on PDO).

At five-on-five, the Habs’ team-wide PDO currently stands at 105.7. That’s an extremely high number, owing to both a very high shooting percentage and excellent even-strength goaltending prowess. The Habs’ shooters have collectively scored on of 10.3% of their shots, a very high number (the average is about 8.5%). Meanwhile Price and Budaj (but mostly Price) have combined for a whopping .954 save percentage at even-strength.

Now, the Habs’s shooters aren’t any better than anyone’s in the league (any differences between players are generally small and tend to get lost in the mass of a full team). And though Price is a well above average goaltender, he’s not superhuman, and even the best goalie in NHL history couldn’t post a .954 save percentage for long. Both those numbers will come down to a more reasonable level down the line. That will cost the Habs; for example if the shooting had been a well-above-average 9.3% rather than 10.3%, the Habs would have scored two fewer goals, and had the goaltending been an extraordinary .930 rather than a superhuman .954, the Habs would have allowed a bit more than 3 more goals. The Habs have scored 8 goals than they allowed so far 5-on-5; with those two reductions, that differential would have been reduced to a mere 3. That alone would have cost the Habs a win.

So the Habs have been lucky, and we should expect a regression in the future. This shouldn’t be surprising, really — the Habs aren’t likely to keep scoring over three goals a game, for example. That doesn’t mean they are going to come crashing down to the basement, however, because they have another thing going for them: they have been completely dominant in puck possession.

It doesn’t necessarily show when comparing straight shot totals or even shot attempts. Part of that is because the Habs have spent a great deal of time leading games, sometimes by quite a few goals. In the NHL teams generally play to the score. When a team is leading, it tends to take fewer shot attempts as it tries to defend the lead, but conversely it converts on those chances more often as they more often get odd-man rushes. Likewise, a team that trails will take more shot attempts as they play more aggressively, but convert on fewer of them due to the opposition’s focus on defense. Therefore,

And that’s where Montreal has shined. They have taken a whopping 57.1% of all 5-on-5 shot attempts (also known as Corsi events) with the score close, defined as either tied or within a goal in the first two periods. And among that lot, they have taken an even more impressive 60.8% shot attempts when the score was tied. In high-leverage situations, the Habs have simply dominated, outpossessing their opponents to an almost absurd degree. And these are some of the best overall measures of team quality we have, so the trend is certainly encouraging.

But.

There’s some context that needs to be brought up. First, and foremost, 8 games is not a whole lot of games, especially when pulling subsamples from it. The score-tied sample is especially small, covering only 143 events. It’s important not to read too much into so small a sample.

More importantly, this has been the docket for the Habs:

TOR@MTL, 2-1 loss (Toronto is currently 4-5-0)

FLA@MTL, 4-1 win (Florida is currently 3-5-0)

MTL@WSH, 4-1 win (Washington is currently 2-6-1)

NJ@MTL, 4-3 overtime win (New Jersey is currently 4-1-3)

WPG@MTL, 4-3 win (Winnipeg is currently 3-4-1)

MTL@OTT, 5-1 loss (Ottawa is currently 5-3-1)

BUF@MTL, 6-1 win (Buffalo is currently 3-5-1)

OTT@MTL, 2-1 win (Ottawa is currently 5-3-1)

That’s a lot of home games (six, in fact), which has allowed Therrien to deploy his preferred matchups much more easily. This has been important particularly on defense, where he had to deal with six games with only one truly strong pairing. Subban’s return will help greatly with that, giving him two strong pairings and allowing the coach to deploy one of Subban or Markov for 45 minutes a game or more (though I would suggest partnering Subban with Gorges rather than Bouillon). Nevertheless, the team’s been favored by a LOT of home cooking. Case in point: while the two-game sample is too tiny to draw much from it, the Habs have only gotten 41.8% of possession events on the road, when the score was within one!

More importantly, though, has been the quality of opposition that the Habs have faced. Of the eight games, three have been in teams in playoff positions, namely the Devils and the Senators. Only Ottawa is actually above .500 as of this writing, and those two games were far and away the toughest Montreal played, being outscored 6-3 and outpossessed both games. The Habs, in other words, have been feasting on weak teams, of which the East has no shortage. This isn’t all bad — the Habs aren’t making the schedules and good teams are supposed to feast on weak clubs the way Montreal has. But it does mean that we don’t have a clear picture of where the Habs actually stand, and it means that the gaudy possession figures quoted above are almost certainly overvaluing the Habs. There’s bound to be some regression there as well.

There’s plenty space for those figures to come down and still leave the Habs a pretty good team. The Habs aren’t a 60% possession club, but whether they are 53%, 51%, 50% or lower remains to be seen. The Habs are inevitably better than last year’s version. The personnel is better (in no small part due to health, though the resurgence of Rene Bourque plays a part). Therrien is a much more adept personnel-manager than Cunneyworth, and to my pleasant surprise, he’s also emphasizing puck possession, with the Habs generally preferring to enter the offensive zone and leave the defensive zone in possession of the puck. It’s not only a more entertaining option from Cunneyworth’s ceaseless chip-and-chase approach, it also promotes winning. The combination of better players and better tactics was inevitably going to lead to better results. The question that remains obscured is “how much better”. Montreal’s ability to beat the likes of Winnipeg, Toronto, Buffalo or Washington on possession doesn’t tell us much because everyone is beating those teams. Their comparative struggles against Ottawa are more worrying, as Ottawa is the only good possession team they have faced yet.

In that sense, the game versus Boston tomorrow represents an interesting test. It’s a home game, but Boston has been a very strong possession club for the last few seasons and represents one of the few genuinely strong teams the East has. Both teams will be rested, which is proving to be a fairly rare occurence in this shortened season. The game represents an interesting measuring stick, and from how well the Habs play we will be able to draw better conclusions about what manner of team the Habs are.

But not too much. Because if 8 games is a small sample, one game is even more so. It will take much more hockey before we can truly discern what kind of team the Habs are. Let’s hope they keep that “lucky and good” thing going!

 

Tags: Fancy Stats M Montreal Canadiens