Hockey Confidential (10 page)

Corsi is the difference between the two. It can be expressed in a number of ways. We could say you are plus-four (the difference between your 20 and my 16) or I'm minus-four. But it's most often, and best, expressed as a percentage. That is, of the 36 total shots taken by both teams in the game, you had 55.5 per cent. I had 44.5 per cent. There's your Corsi.

So, obviously, any number over 50 per cent is better than a number under 50. It doesn't get much more basic than that.

“The thinking behind it,” Ferrari said, “is that the more you have the puck, the more likely you are to win.”

You might ask, why count shot attempts? Wouldn't counting goals make more sense?

Fair enough. A team's goal differential—the number of five-on-five goals it scores minus those given up—can be a valuable metric, but statistics tend to be more effective the larger the sample size. In an NHL season, teams usually score between 200 and 300 goals. They often take more than 2,400 shots, plus all those missed shots and blocked shots. So the sample size for attempted shots is so much greater than that for goals scored—maybe 10 times as great. Many goals are scored as a result of skill and artistry, but most involve a great amount of luck or random bounces. Corsi gives you a volume of events that helps to offset the high degree of luck when using a smaller sample size.

You may also ask, Why even bother tracking Corsi? What's the point of it?

Corsi's value is that it has been proven, time and again, to provide a reasonably accurate representation of puck possession. That makes total sense. The more shots you attempt, it stands to reason, the more you have the puck on your stick. Just to be sure, researchers have used a stopwatch to actually measure puck possession in specific games, and the corresponding possession ratios inevitably correlated to the Corsi numbers. This is not fiction.

Now, positive puck possession doesn't absolutely guarantee victory—as noted earlier, the 2013–14 Colorado Avalanche were 25th in the NHL (47.4 per cent) in Corsi rankings, but finished with the third most points in the entire league. But the
probability
of a team's success is greatly enhanced with positive puck possession. Intuitively, I think we can all agree that makes sense.

That's not exactly an advanced notion, if you think about it, although the non-believers remain dubious of Corsi's usefulness and/or validity as a predictive number because, well, the numbers don't
always
add up.

It is by no means perfect, not even close.

In 2013–14, for example, three of the NHL's top 10 Corsi teams—New Jersey at No. 4 (54.6 per cent), Ottawa at No. 8 (52.2 per cent) and Vancouver at No. 9 (52 per cent)—didn't make the playoffs. And, of course, there was the glaring case of the Avalanche, who became to the #fancystats skeptics what the Leafs were to the numbers crowd.

“Each [exception] is different,” Desjardins said.

The Devils, for example, went an incredible 0-for-13 in shootouts.

“I think we can agree that is just bad luck,” Desjardins said. “With average luck in the shootout, the Devils make the playoffs.”

This is also where that line about advanced stats not being that advanced falls apart. Guys like Desjardins and Dellow can, and will, explain in great and complicated detail the whys and wherefores of Ottawa and/or Vancouver's misleading Corsi numbers. If I let them do that here, trust me, you'd get as lost in the numbers as I did with logarithms in Grade 12.

That, of course, infuriates the anti-numbers advocates, who believe the #fancystats guys trumpet the validity of their numbers when they're proven to be “right,” but find ways rationalize the hell out of them when they're perceived to be “wrong.”

The reality is that the stats guys know there's no one magic number that is right all the time; there will always be outliers and exceptions. They will be the first to tell you luck, or sheer randomness, is a big part of hockey and no metric can actually measure or forecast all the bounces. They'll also note Corsi is a five-on-five, even-strength stat that doesn't take into account special teams or goaltenders or a host of other factors that help determine the outcomes of hockey games. Therefore, while using Corsi, the #fancystats disciples will often apply it in conjunction with other metrics. But that's not something non-numerically inclined fans can always get their heads around. It starts to get complicated, putting the “advanced” in advanced stats.

When all is said and done, though, whatever frailties, real or imagined, exist within Corsi, the #fancystats crowd will tell you the data makes them right a lot more than they're wrong, that the probability of a good Corsi equalling a good hockey team is far greater than not.

Of course, we could make all of this even more complicated. But we won't do that now, other than to note there are myriad derivatives from basic Corsi.

Corsi can be a team metric or it can be applied individually to any player. Also, you can factor in whether faceoffs originate in the offensive or defensive zone, the quality of opposition a player faces, even the quality of teammates he plays with. You can measure a player's Corsi when he's on the ice in relation to his team's Corsi when he's not on the ice (a metric called Corsi Relative). It goes on and on . . .

There is, however, one further aspect of Corsi that can't be glossed over: the “close” factor.

Most Corsi references you see are what are termed “close,” which simply means the shot attempt calculations are made based on when the score is either tied or the teams are separated by only one goal in the first two periods. What the stats guys quickly figured out is that, once a team is ahead or behind by two or more goals, there is either a conscious (through coach's orders) or unconscious (through players instinctively “protecting” a lead) decision to alter behaviour to play a more conservative, defensive-minded game. If there were no distinction between “close” games and the others, Corsi's integrity would be severely compromised because of a radical change in mindset affecting how the game is played, depending on the score.

This concept is known as “score effects.”

“There is an almost universal tendency for NHL teams to get a greater share of the shot attempts when they are behind than they do when they're tied,” Dellow said. “They also get a greater share of the shot attempts when they're tied than they do when they're leading. For example, the 2013–14 Bruins had a five-on-five Corsi percentage, when leading by one, of 50.3 per cent. With the score tied, their Corsi was 55.7. When the Bruins were trailing by a goal, their Corsi was 59.3.

“One of the funny things about this is the ‘If we just started the game like we played the third period' postgame quote, which so many fans of bad teams have heard. It's funny to hear these quotes from people who have spent years in the NHL, seeing this phenomenon play out over and over. They aren't going to start games like that unless they start with a one-goal deficit, which isn't going to help them win games.”

Still with us?

Good, that's as far as we'll go with Corsi.

Now, on to Fenwick, which is going to be really easy because everything you read about Corsi still applies. The only exception is that Fenwick doesn't include blocked shots. Otherwise, all the equations, standards and terms of reference are identical.

So, you ask, why bother with it if it's so closely aligned with Corsi?

Well, Matt Fenwick was of the opinion that, while Corsi was a very good proxy for puck possession, if you wanted a metric to more accurately reflect scoring chances, blocked shots should be removed from the equation. A blocked shot, after all, is not a scoring chance. Neither is a shot from the blue line, but that gets counted in Fenwick, as it does in Corsi. Again, as was the case with a researcher actually timing puck possession to ensure Corsi was a good proxy, the same thing has been done with scoring chances and Fenwick. They match, more or less. Fenwick is, therefore, considered a valid proxy for scoring chances. Again, this is not fiction.

The difference between the two for #fancystats aficionados may be no different than your preference for Coke or Pepsi.

Corsi has perhaps better brand recognition. Because Corsi tracks more events, it tends to illustrate puck-possession trends more quickly than Fenwick. Stats guys will tell you that, in the short run, Corsi may be a preferable metric. But because Fenwick is perceived as a little more accurate than Corsi as a proxy for scoring chances, over the long haul, many feel it's superior to Corsi. My sense is the #fancystats cognoscenti tend to slightly prefer Fenwick, but will utilize both, since multiple studies have shown puck possession and scoring chances to be linked anyway. For the most part, Corsi and Fenwick tend to mirror each other pretty well, and on those rare occasions of marked differences, the nod seems to go to Fenwick.

As an aside, from a strictly neutral observer trying to grasp #fancystats, having two metrics so close in nature is more confusing than helpful and, therefore, hinders the “marketing” of new-age numbers. But that's just me.

Which brings us, finally, to PDO.

“A brilliant number,” Desjardins said of PDO.

Hockey's advanced number crunchers use all three metrics in concert with each other, as well as other, more complicated calculations, but you get the sense that if they could have only one of them, it would be PDO.

That's because it's perceived to be the most “predictive” of the Big Three.

“Saying that guys or teams with a really high PDO will regress is sort of the ultimate shooting fish in a barrel of hockey analytics,” Dellow said.

Here's how PDO works, or at least what intrigued Brian King enough to come up with the initial concept.

King figured if you take a team's shooting percentage with a given player on the ice (his “on-ice shooting percentage,” not to be confused with his individual shooting percentage, which is calculated only on his own shots and goals) and add it to his team's five-on-five save percentage when he is on the ice, there should be a total number that represents the baseline for any player who is neither extremely lucky or unlucky. As he said, “for shits and giggles,” he chose 100 per cent, which turned out to be pretty much perfect. When King calculated the PDOs for individual Oiler players at the time, he discovered those with PDOs of greater than 101 tended to get contract extensions, while those below 99 tended to get shipped out of town.

The more research that was done on PDO, the more it became clear that players or teams (PDO works well for both) with exceptionally high or low numbers were often experiencing really good luck or wallowing in the misery of bad luck. It was determined that the outliers would, over time, regress to meet somewhere in the middle.

Five-on-five team shooting percentage, for example, tends to run between 6 and 10 per cent, but the average is around 8 per cent. Five-on-five team save percentage generally ranges from .900 to .940, but the average is around .920. So, add the on-ice shooting percentage of .08 to the save percentage of .92, and there you have it: your midpoint of one—or 100, depending on how you want to present it. Players and teams can post PDOs much higher or lower than 100, but over time, that's roughly where they'll end up.

King couldn't possibly have known what an incredibly predictive tool PDO would become, both for players and team. Non-believers in #fancystats often don't like PDO because its core concept can really rain on a hockey fan's parade.

Take Edmonton's Jordan Eberle, for example.

In 2011–12, his second NHL season, Eberle had a breakout year, scoring 34 goals and 76 points in 78 games. The widespread sense at the time was that it was a hint of things to come, that Eberle might be on the cusp of becoming an elite-level point producer who was only going to get better, to put up bigger numbers. The sky was the limit. It was exciting, so hopeful and optimistic.

Dellow—like Ferrari, a passionate Oiler fan—looked at Eberle's underlying numbers (as measured by PDO) from that season and didn't like what he saw. The Oilers shot 12.7 per cent in five-on-five situations when Eberle was on the ice and had a .909 save percentage, making Eberle's PDO 103.7. If Oiler fan Dellow didn't believe so firmly in the power of PDO, perhaps he could have joined the cheery chorus predicting Eberle was on the launching pad to become a perennial point-a-game player (or better).

But he didn't. Or couldn't. The numbers wouldn't allow him, and the numbers, Dellow believed, don't lie. He wasn't shy about saying so, either. He didn't say Eberle wasn't—or couldn't become—a good hockey player; he just said Eberle's numbers that year were much more likely to be an aberration than the norm and that Eberle was likely to score less, not more, in coming seasons. A PDO of 103.7, Dellow told anyone who would listen, was unsustainable.

And it was.

Eberle scored 16 goals and 37 points the following year in the lockout-shortened 48-game season; he had 28 goals and 65 points in 80 games in 2013–14. His PDOs for those two seasons regressed from 103.7 to 98.9 and 100.5.

The same principles of PDO apply to teams.

That was, in large part, the rationale for the infamous #fancystats Maple Leaf Prophecy of 2013–14. Toronto had a PDO of 103.0 in 2012–13, which was viewed as unsustainable. Sure enough, in 2013–14, it dropped to 101.2.

That mean of 100.0, for most teams, is like a magnet. The teams with PDOs significantly above or below 100.0 tend to regress towards it. Again, though, as with Corsi and Fenwick, it's not inviolable. You can find exceptions to the rule. Pittsburgh posted back-to-back PDO seasons of 102.2 in 2007–08 and 2008–09 (regressing to 99.5 in 2009–10). Anaheim went from 101.7 in 2012–13 to 102.3 in 2013–14. Teams with chronically poor goaltending (bad save-percentage teams) can get mired in the range between 97.5 and 99.5 and not drift towards 100. Teams with really extraordinary goaltending can stay above 100 longer than they should.