Habs vs Flyers 2008 retrospective: an initial look at puck possession

facebooktwitterreddit

There are two competing explanations about why the Flyers defeated the Habs in 2007-2008. The first one is that the Flyers won due to their size, physicality, and intimidation; they used this to control the Habs’ skill game and prevent them from obtaining many good chances. The other theory is that the Canadiens vastly outplayed and outchanced the Flyers everywhere but the crease but were done in by a combination of stellar goaltending from Biron, a slump from Price, and no small amout of bad puck luck.

Analyzing a small set of games via statistical analysis is always problematic, simply because of the sample size. In the span of four to seven games, transient performance factors and sheer luck have a disproportionate effect on the individual result of games. In a way, this is one of the reasons why statistics like Corsi (total shots at net, counting all goals, saved shots, misses, and blocks) are helpful; even in a small span of games, they may represent several hundred events, a sample far more satisfactory and more meaningful, in the statistical sense, than the five or so goals of the typical hockey game. At the very least, shot-based metrics can give us a strong idea of who carried the play and held possession of the puck the most.

Let’s examine, then, on a game-by-game basis, who held the puck most five-on-five, as represented by the various shot-based metrics. In this case I use a percent share of events, from the perspective of the Habs; in a head-to-head matchup like this one, the Flyers’ number will of course be the inverse (100% minus that number).

GameCorsi%Fenwick%Goals+Goals-PDO(Corsi)Sh%Sv%PDO(Shots)
150.43%48.15%220.9990.0710.9260.997
259.77%62.50%030.9140.0000.7860.786
360.87%57.45%020.9260.0000.8330.833
457.73%63.24%211.0110.0740.9411.015
556.03%53.66%350.9480.0970.8390.935
Total56.40%56.29%7130.9640.0540.8710.926
Losses58.27%58.89%5110.9520.0500.8510.901

These metrics are very telling. The Habs’ puck possession numbers at 5 on 5 were generally extremely strong, especially in games 2 to 5. Their worst game, they had a 56.03% Corsi and a 53.66% Fenwick — to put things into perspective, the latter number was equal to the 2011-2012 Los Angeles Kings, who did pretty well for themselves last year. Their puck possession advantadge during games 2 to 4 was outright dominant. Yet they could not capitalize; in games 2 and 3, notably, they were outscored 0-5 5-on-5, even though they dominated puck possession to the tune of 57% and up.

I isolated the statistics during games 2 to 5, the games that the Habs lost, to emphasize how strong their puck possession during those games. Ironically, the game where they had the worst possession by far was the one they won.

But wait, what about context? Corsi metrics, of course, are not all-encompassing. Score context, in particular, has a huge impact on those metrics; if a team is behind a lot, experience tells us that they will increase their puck possession yet degrade their percentages in an effort to catch up. And the Habs, losers of four of five games, spent a lot of time trying to make up a deficit! They spent a lot of time playing “catchup hockey”. Is it possible that their puck possession advantage was due to this, and that the Flyers dominated until they took a lead, then played more defensively?

To answer this question, let’s look at the same metrics, but this time separated by score situation from the perspective of the Habs:

ScoreCorsi%Fenwick%Goals+Goals-PDO(Corsi)Sh%Sv%PDO(Shots)
Down 350.00%44.44%001.0000.0001.0001.000
Down 258.70%61.54%211.0210.1330.8330.967
Down 154.31%52.24%240.9740.0400.9050.945
Tied57.23%61.16%250.9530.0450.8480.894
Up 162.50%55.88%120.9220.0830.8460.929
Up 250.00%50.00%010.8000.0000.7500.750
Tied 2-556.93%61.80%140.9450.0300.8400.870

As we look over those results, we see that score effects did not really play a major role in dictating the teams’ puck possession. When the score was tied, especially, the Habs did exceedingly well, holding a 57% share of Corsi event and doing even better in Fenwick at 61%. Their numbers actually got lower when down by a goal, but they still held themselves well above the 50% bar. In any case, it did not appear that the Habs’ puck possession was an artifact of trailing a lot; in fact their puck possession was better when tied and up than it was did while trailing, the reverse of what we’d normally expect!

Isolating the losses (the last row on the chart) really starts to give us a good idea of exactly why the Habs lost the games. Whenever one of these games was tied, the Habs utterly dominated puck possession… yet scored on a measly 3% of their shots, while the Flyers made 16% of their own shots. The regular season league average was in the vincinity of 8% — and the Habs had done better than the Flyers that year. Simply put, whenever the game was tied, the Habs could not buy a goal, either because they couldn’t shoot straight, were stopped by Biron, were victims of bad bounces, or all of the above. Meanwhile, everything the Flyers shot at the net went in, or so it seemed; their conversion rate was double the usual. This was how the Flyers so often took the lead — not by dominating play, but by “making their shots”. While unsustainable in the long-term, this is the sort of thing that will happen with some frequency in short periods, such as a five (or seven!) game series: one team will dominate scoring chances and puck possession, but will find conversion rates tilted against it so much that it loses repeatedly.

Finally, since we’re talking about overall stats for the series, a tidbit on physicality. While real time stats are very problematic, and subject to large amounts of arena bias, I thought the following chart (which covers all situations) was quite amusing… and the arena bias is reduced because both teams played the same number of games in each arena.

GameHab HitsFlyer HitsHits%
1382857.6%
2332854.1%
3261957.8%
4191950.0%
5382659.4%
Total15412056.2%

Remember that the team that has the puck more (in this case, the Habs) tends to hit less simply because they’re not chasing, the other team is. Hits are usually negatively correlated with puck possession. Yet the “bigger, more physical” Flyers, even though they consistently held the puck less than the Habs, were also outhit despite this.

All this points to an answer regarding our two explanations above. The Habs were not outpossessed, outshot, or outplayed, regardless of the game or the score situation; they maintained notably superior puck possessions at almost all times. And they were not even outhit! At 5 on 5, it appears that the Habs were the better team, at least outside the crease, holding the puck more consistently and outshooting their opponent, regardless of the Flyers’ supposed advantadge in size or physicality; if the Habs’ “skill” game was somehow hampered by the Flyers’ tactics, it certainly wasn’t in the area of puck possession.