The Canadiens are bleeding goals this year at an alarming rate, or at least in the last two games. While Carey Price has hardly been much help, it's pretty obvious the defense has a few holes in it so far. But who has really been struggling, and who looks worse than they've actually been?
This early in the season it's very hard to evaluate players because four games is a pretty irrelevant sample size. However I believe it is worthwhile to look into some advanced stats just to challenge common perceptions. For the purposes of this exercise I'll be leaving out Chris Campoli and Jaroslav Spacek because a) they're not playing, and b) they didn't play enough to learn anything from their stats.
All statistics are from Gabe Desjardins' behindthenet.ca
Four games into the year and already fans are screaming "sophomore slump". Even reasonable fans. But is it warranted? Has Subban been that bad? Right away when you look at Subban's quality of competition it's evident how difficult his job is. With a score of .481 he's facing competition nearly twice as tough as the next defenseman on the team, as Gill scores .278.
Against that competition Subban has still managed the highest Corsi score among all Habs defensemen at 24.77 and the highest relative Corsi at 24.8. To illustrate how effective Subban is at carrying the play, the Canadiens' Corsi score while Subban is on the bench is zero. You read that right, while Subban isn't playing, the Habs produce the same amount of shots as their competition. While he IS playing, they dominate. Not exactly the numbers of a sophomore slump.
Where Subban is struggling however, is his zone start/finish scores. He's started 60.5% of his shifts in the offensive zone, but ended only 45.7% in the offensive zone. That's a startling reversal of play last year than had him start more often in the defensive zone and finish in the offensive zone. However every defenseman on the Canadiens as an offensive zone start percentage of at least 50% which might tell us that the Habs are spending far less time in their defensive zone overall than they did last year.
Where Subban is getting victimized so far this year this year seems to have a lot more to do with luck than anything else. While he's on the ice the team is massively out-shooting the opposition, but he's still getting outscored. While part of this is due to a few brain farts on Subban's part, giving up high quality scoring chances while trying to dangle, a large part of the blame also lies with Carey Price, who just hasn't been very good so far. Price's save percentage while Subban is on the ice is just .864. When you combine that with the bad luck Subban has experienced offensively, reflected by a team shooting percentage of 5.56, it looks like a big ball of bad luck. To put these numbers into perspective, Subban also had the worst on ice save percentage of all the Habs defensemen last year, but it was .911, nearly 5 percentage points higher than this year. The combination of team shooting percentage and on ice save percentage gives us a measure called PDO. Over a season a player's PDO usually edges close to 1000, which is basically breaking even. Last year for example, Subban has the team's lowest PDO with 977, while Spacek had the highest at 1005. Subban's PDO this year (919) will undoubtedly move closer to the mean of 1000 as the sample size gets larger. It's extremely unlikely that Subban continues to experience bad luck to this extent on both ends for an extended period, so I wouldn't be worried about him.
Luckily for Habs fans, Subban's bad luck has been partially negated by Gorges being an absolute stud on the back end. I was a little worried coming into the season that he might not be the same rugged Gorges we've all come to love, but he's the same guy, only a better skater and showing some offensive tools he previously hasn't used. While Gill plays with Subban on the first pairing, Gorges is left to carry the second pairing, mostly with Diaz so far this year. Gorges was the Habs' top defenseman during the preseason, and so far this season he's performed better than anyone else as well. His quality of competition is a little low for him at .048, but it's important to remember how a small sample can play with the results. I doubt Subban's Qualcomp will remain that much higher than the rest of the team as the season goes on for example.
Gorges has managed a positive Corsi rating, third best among Habs defenders at 6.44, however his relative Corsi to the rest of the team is -3.6. The reason for this largely being the extent to which Subban carries the offensive play. Gorges has put up a good Corsi number, but the team is still out-shooting the competition at a higher rate when he's on the bench because of Subban. This would be why context is so important in statistics.
Where Gorges is most dominant happens to be where Subban has been weakest. Gorges has started 51.4% of his shifts in the offensive zone, and finished 56% of his shifts in the offensive zone. That tells you how effectively he's been moving the puck out of Montreal's zone and pushing play towards the opponent. This is probably the reason that Gorges has been getting the most even strength ice time of all Canadiens defenders.
Gorges has also been the beneficiary of some good luck. The team's shooting percentage while he's on the ice is a staggering 15.38%, while he's received average goaltending (well, average for Carey Price) with Price putting up a .923 save percentage while Gorges is playing. While the save percentage Price is giving Gorges is clearly sustainable considering it was his average last season, the on ice shooting percentage is not. Gorges' PDO so far is 1077, which we should expect to regress. The Habs aren't the 1980's Oilers unfortunately, so that scoring will eventually even out throughout the lineup.
A lot of people have been on Gill this young season. I'll admit that I'm one of them, but I don't blame Gill one bit. The problem is that he's being asked to do far too much. It's clear that the team wanted to keep the Gill - Subban pairing from last year in tact because it was so effective, but Gill is another year older and usually struggles a bit to start the season. Personally I think it's time to split them up and give Gill a slightly easier job. He's simply not cut out for another full season of the kind of competition we need Subban to face, simply a .278 Qualcomp score for Gill is too high, even with injuries. We need to put him in a position where he can succeed.
In spite of taking most of his shifts with Subban, Gill's Corsi numbers aren't that impressive. He sits fourth on the Habs blueline at 5.54, while his relative Corsi is -4.5. This isn't really surprising consideringjob isn't to bring the offense, but to shut down the opposing team, but playing with Subban should yield more impressive results.
Similar to Subban, Gill has been started in the offensive zone more often than last year at 57.6%, and finished in the offensive zone far less than is desirable at just 46.2%.
Gill has suffered the same bad luck that Subban has with team shooting percentage and save percentage while he's on the ice, just not to the same extent. Gill has received a save percentage from Price of just .875, while the team has converted on just 7.41% of their shots. Gill's PDO comes out to 949.
When looking at the stats I expected Emelin to come out looking better than guys like Diaz and Weber, but I was surprised to see that he doesn't just look better, he looks damn good. Shockingly Emelin has been trusted enough by Jacques Martin to have the third highest quality of competition among defensemen at .105. It's especially significant considering his D partner, Weber is in the negatives at -.067.
Even more impressive is Emelin's Corsi stats, as he ranks second among the defensemen in raw Corsi at 15.61, as well as relative Corsi at 14.3. Emelin is the only defenseman on the team who's relative Corsi isn't pushed into the negatives by Subban's ridiculous numbers.
Emelin joins Josh Gorges as the only two defensemen who's offensive zone start/finish score doesn't end up negative. Emelin starts and ends his shifts in the offensive zone at the exact same rate of 55.6% of the time.
Surprisingly even though Emelin has the second most impressive numbers of all the Habs defensemen, he's had some horrendous luck. Even worse than Subban's actually. The team's shooting percentage while Emelin is on the ice is a paltry 4.76%. Meanwhile he's getting killed by Price putting up a horrendous save percentage of just .846! Emelin ends up with the worst PDO of all the Habs defensemen. Expect this number to get closer to 1000 very quickly, provided Price improves.
Another surprise from Emelin so far is his discipline. The biggest fear coming into the season was that he was going to regularly get caught out of position going for big hits, or get penalized for doing something stupid. So far it seems he's really bought in to Martin's system. Add to that that he's drawn 3.9 penalties per 60 minutes of ice time and his future is incredibly encouraging. I think Emelin is ready to be tested a bit with further responsibility.
Weber is an interesting case. Of the defensemen currently in the lineup, he's clearly sheltered more than the rest. He also clearly has less trust from Martin than the other 5, deservedly or not. As you can see from Chris Boucher's analysis today, Weber is probably our worst option on the penalty kill as well. However, he's not nearly as sheltered as he was last season thus far, with a Qualcomp score of -.067 in comparison to last year's -.103.
Unsurprisingly, Weber ranks fifth among Habs defenders in Corsi scores. While he's out there we're still out-shooting the competition with his 4.64 rating, but as with most players he's relative Corsi gets pushed into the negatives at -5.7. None of this is surprising for a 3rd pairing defenseman thus far, however out-shooting the competition consistently is a good sign.
Perhaps one reason why Weber is looking so good is because of all Habs defensemen currently playing he's been starting in the offensive zone most often at 61.3% of his shifts. That doesn't tell the whole story however because he's not bleeding out in the zone finishing department, still managing to end his shift in the offensive zone 57.6% of the time.
Here's where we get back into that luck thing again. While Weber is on the ice, the team shooting percentage is a ridiculous 15.38%. Unsustainable but impressive. Even more unsustainable is that Weber hasn't been on the ice for a goal against at even strength, giving him an on ice save percentage of 1.000. Those crazy numbers put his PDO at 1154. Weber is off to a hot start, and it might be a good idea to play him a bit more until he cools off and his luck runs out.
Diaz is a pretty smart player. He's savvy enough to compensate for his lack of size, even though he's not as physically strong as the similarly small Weber. But I believe he's being played out of his depth. Diaz is still being sheltered a bit with a -.015 Qualcomp rating, but I have a feeling that is slightly affected by the first game against Toronto when he wasn't being asked to play on the second pairing with Gorges.
Among the Habs defensemen, Diaz is the only one with a negative Corsi score. What makes this even more troubling is that it's not just a mild negative, it's pretty severe. His score is a terrible -16.20. Even worse than that, considering the rest of the team handily out-shoots the opposition while he's on the bench, his relative Corsi is -37.2, which is the 14th worst relative Corsi in the entire NHL, ninth worst among players who've played three or more games. Of course part of this is an adjustment period to the smaller rink and a faster game, but Diaz could use some sheltering.
Maybe a bit of an excuse for Diaz's Corsi numbers lie in the fact that of all the Habs defensemen he's started in the offensive zone the least at just 50% of his shifts. Unsurprisingly however, he still ends his shifts in the offensive zone less often than any other defenseman.
Luckily for him, his horrible numbers have been negated so far by some luck. The Canadiens have managed a solid 9.09 shooting percentage while Diaz is on the ice, and Price has put up an excellent .939 save percentage over that time, putting Diaz's PDO at 1030. This obscures the fact that Diaz is really treading water so far in the NHL. I'm not saying he won't be a good player or doesn't show promise, just that he could use a slower adjustment period. He certainly shouldn't be playing 19 minutes per game.
So what can be done to stabilize the defense a bit while this adjustment period takes place for all these inexperienced players? I think we saw a little bit of the plan happen in overtime last night as Martin juggled the pairings a little bit. For the short term, I would say the most likely way for the Habs to experience success defensively is for Martin to deploy the following pairings at even strength:
- Emelin - Subban
- Gorges - Weber
- Gill - Diaz
These pairings keep an experienced player on with each young player, put our two strongest even strength Corsi guys against the opponent's top lines, they give Gorges a player who's currently playing hot, and they allow Gill to play against weaker competition at even strength while Diaz adjusts to the North American game. At this point trying to give Emelin some extra responsibility is low risk and high reward. If he can't handle it you revert to your previous pairings, if he can, you've discovered you have another top 4 defenseman earlier than you thought you would.