Skip to content

Leafs collapse should be cautionary tale for young Avs

Mar 31, 2014, 1:09 PM EDT

Tyson Barrie, Semyon Varlamov AP

The collapse of the Toronto Maple Leafs — a collapse that was widely predicted by hockey’s burgeoning analytics community — should be a wake-up call for anyone who still doubts the predictive value of shot-based metrics like Corsi, and what those statistics say about a team.

No, Corsi doesn’t tell the whole story; if you’re looking for a be-all-end-all hockey statistic, you’re not going to find one. But those who continue to ignore what’s already out there, well, frankly, we’re not sure how anyone can continue to ignore these stats.

Consider: the top three teams in Corsi (score close, 5-on-5) are Los Angeles, Chicago, and Boston. The bottom three are Buffalo, Toronto, and Edmonton.

This is not some giant coincidence. Good teams possess the puck more than they don’t, and that’s all these shot-based metrics are trying to measure. There’s really nothing “advanced” about these stats.

A lack of puck possession can be overcome, at least in smaller sample sizes, by factors like goaltending, special teams, and high shooting percentages. However, as the Leafs are discovering, relying on those things can be a dangerous game to play.

In the Globe and Mail, this is what they’re writing about the Leafs, losers of eight straight in regulation:

Fixing what ails this franchise will not be easy. Doing it while in continued denial about its faults will be impossible.

There has been an unearned hubris around this group for a while, going back well before last season’s half season run, and it permeates the organization from the players to the staff, management and ownership.

And this is what they’re writing in the National Post:

Hockey is fighting its way through the early days of analytics, which long ago washed over baseball and are transforming basketball.

And in hockey, the basic puck possession stats — Corsi, which measures shots attempts for and against, and Fenwick, which excludes blocked shots — have often been met with derision, not least by the Leafs themselves.

For an up-and-coming team like the Colorado Avalanche, what’s happened to Toronto should be a cautionary tale. We say this because the Avs are currently 26th in Corsi, above only Buffalo, Toronto, Edmonton, and Calgary. And that’s generally not the kind of company you want to keep.

For all the coach-of-the-year buzz that Patrick Roy is receiving, if not for his starting goaltender, Semyon Varlamov, there’s just no way Colorado would have the sixth-most points in the NHL. Varlamov has made 1,715 saves this season, the most of any netminder in the league. If the Hart Trophy were actually given to the “player judged most valuable to his team,” he would have a strong case for winning it.

Now, this isn’t to blame the Avs for having a good goalie. Nor can you say their 10.3 shooting percentage, the second highest in the NHL, is just luck, what with all the young offensive talent they boast up front. When those guys get the puck in good spots, they’re going to bury a lot of shots.

It’s simply a question of, can you count on these things to continue at their current rate into the future?

History says no.

Last season, the Ottawa Senators led the league in 5-on-5 save percentage with the score close. They made the playoffs and Paul MacLean was named coach of the year.

The season before, the Florida Panthers led the league in 5-on-5 save percentage with the score close. They made the playoffs and many felt that Kevin Dineen, since fired, was snubbed big-time in the coach-of-the-year nominations.

On Saturday, Colorado beat the Sharks, 3-2, to clinch a playoff spot.

The Avs were outshot, 49-22.

  1. davekillercarlson - Mar 31, 2014 at 1:37 PM

    Tell us how you really feel, Jason.

  2. stepanup - Mar 31, 2014 at 1:41 PM

    Nicely done JB. Finally something with some opinion, with statistics to support it.

  3. comeonnowguys - Mar 31, 2014 at 1:50 PM

    I like the Avs, have for a couple of years where you could see them start to put the pieces in place, and they are a nice balance of youth and speed and defense.

    That said, it felt like most of those Hawks-Avs games this year were a lot the like Sharks games describe above.

    If you can pull it off, more power to you. However, if V can’t keep this pace up, it could end badly.

  4. molsonexport - Mar 31, 2014 at 2:13 PM

    One and done.

  5. guitarhunterdude - Mar 31, 2014 at 2:20 PM

    Solid article, nothing groundbreaking here. I expect Varlamov to remain hot through the playoffs, which could make the Avs a second-round or conference-final team. Next year, I think there’ll be some inevitable regression, both in terms of team shooting percentage and Varlamov’s save percentage. I don’t think they’re quite set up long-term (they obviously need better defensemen and better forwards throughout the bottom six), but they’re a scary team this year.

  6. blomfeld - Mar 31, 2014 at 3:15 PM

    “Relying on goaltending can be a dangerous game to play” ? …

    The short answer to that is yes & no, just as it is with most things in life. Obviously ‘having’ elite goal tending is critical in terms of winning the SC, which of course is completely different from ‘relying’ on elite goal tending to do the same. Therefore the bottom line is that supreme victory is a sum of ‘multiple’ required inputs, with goal tending only being just ‘one’ of them.

    Example A) … where X=Quick, Y=Doughty and Z=Carter, then:

    3x + (2z − y)/x + 125y − (x + y)/(z +2) x 82 “regulation” games = Stanley Cup ! :)

    • patthehockeyfan - Mar 31, 2014 at 4:21 PM

      Jeez, blom! I wish you would have just stuck with the short answer!

  7. hosewater2 - Mar 31, 2014 at 3:36 PM

    …or the Avs success could support the null hypothesis that advanced analytics cannot predict future success.

    It’s easy to predict failure, and you don’t need advanced analytics to do that.

    Who else is ready for the pseudo-journalists/bloggers to stop with all this business of claiming responsibility for predicting the Leafs’ late season slump? The same prediction has not held true for the Avs, so a reasonable person would say you’re 50/50 on accuracy, you know like flipping a coin.

    • hosewater2 - Mar 31, 2014 at 4:14 PM

      And oh wait, this will get better: if the Avs lose in the first round it will be because of Corsi/Fenwick, not because they lost their leading scorer, right? Let’s hear it for Mom’s basement!

    • killhowe - Mar 31, 2014 at 5:55 PM

      Yes advanced stats can be difficult to understand, but your ability to overlook them is laughable. Let me simplify JB’s article. Lack of Possession = Shots Against = Goals. Varlamov is playing great right now, but to expect your team to win when you are giving up 40-50 shots a game is unrealistic. Ask the Leafs how thats working out for them.

    • no - Mar 31, 2014 at 9:33 PM

      Yeah that’s not the null hypothesis. Cute, though.

  8. petermanpeter - Mar 31, 2014 at 4:29 PM

    go to the official Avalanche site. Watch the video of Roy post practice this morning. This says everything you need to know about how they play. It is “scoring chances” that count, not shots.

    • Jason Brough - Mar 31, 2014 at 4:43 PM

      Listen to Leafs VP Dave Poulin. He says the exact same thing:
      http://iphone.tsn.ca/tsnpodcasts/LL_DavePoulin_Mar31.mp3
      Doesn’t mean he’s right.

      • petermanpeter - Mar 31, 2014 at 7:00 PM

        If you want to compare the Avs to the Leafs, you’ve got to go way beyond shots. The Leafs are a hot mess up and down that organization and their problems in the front office and behind the bench have been well documented by Mirtle and others. Everything Roy does is by design; he’s able to get the most out of each and every player on the roster. The Avs pass the eye test, and have recently been praised by Ken Hitchcock on their style of play. Getting outshot may be a similarity between these two squads, but that’s where the comparison ends.

    • no - Mar 31, 2014 at 9:30 PM

      yeah because “scoring chances” are quantifiable, and aren’t just casual lay observation with no discernible consequence

  9. AppealToReason - Mar 31, 2014 at 7:13 PM

    “Advanced stats seem to be reliable long term and predictive of outcomes, unless its my favourite team!”

    No one remembers when the Avs basically did this exact same thing on the back of an unbelievable season from Anderson a couple of years ago?

  10. no - Mar 31, 2014 at 9:28 PM

    Wow people are real afraid of numbers

    “n-no!!! they look good and i like them!!!”

    cool. i’ll be watching my team get past the first round. patrick roy can eat one for dinner.

  11. dbfan4ever - Mar 31, 2014 at 9:33 PM

    “Lack of puck possession can be overcome in small sample size” Does an entire season count as a “small sample size”?! Can’t wait for the AVS to continue to make this author look like a fool. Hey Jason did ever hear the expression “there is more than one way to skin a cat?” I guess I need to explain to you what that means. It means it is possible that there is more than one method for winning. Roy is going to continue to baffle the “experts” and hockey know-it-alls. What do the metrics have to say about a team pulling the goalie with 4 to 5 mins left? Roy has done it several times and has had great success!

  12. xx1111 - Apr 1, 2014 at 12:02 PM

    LOL

    Who cares if the Avs are one and done.. we have one of the most exciting teams in the NHL. Our core age is like 24 years old. Most so called “experts” predicted the Avs to finish last this season. This is great experience for this young club. We can go into the offseason and build up the defensive side.

    We have enough offense but we really need 2 shut down defenders to pair with them.

    But i wouldnt count the Avs out from advancing to the 2nd rd in the playoffs either. They dont quit.

Top 10 NHL Player Searches
  1. D. Alfredsson (1957)
  2. B. Bishop (1909)
  3. M. Fleury (1738)
  4. J. Schultz (1491)
  5. C. Anderson (1466)
  1. D. Krejci (1456)
  2. J. Boychuk (1451)
  3. E. Staal (1378)
  4. R. Lehner (1278)
  5. D. Setoguchi (1263)