thinkingman

I see this a lot. Or variations of it:

@dzuunmod No fear at all.Hockey isn’t baseball. Too much of what goes on cannot be statistically quantified. It’s stop start stop start.

— steve simmons (@simmonssteve) May 17, 2013

The lesson here isn’t that baseball can be perfectly quantified, with raw data input into large machines that pump out baseball-playing robots that perform effectively and consistently. And hockey is different because you don’t get clear batter-pitcher matchups. In the baseball world, the latest fad is to blame sabermetrics for interfering with players’ approaches at the plate. There are tonnes of people unconvinced that modern baseball statistics and analytics can tell you more about the game than traditional stats or your eyes alone. Clint Eastwood even starred in movies made about that.

So when I see the argument “hockey is not baseball, you can’t properly quantify it” I wonder just to what extent people have given that any sort of thought. Baseball is not an easily quantifiable game, and neither is hockey. Baseball statistics are only scratching the surface of things like pitch sequencing, fielding and pitcher-catcher synergy. Nothing is absolute. Nothing is in hockey, either. Contrariwise to what the TSN statistics guys think about Corsi…

An example of why Corsi is a flawed statistic…..Crosby: + 11 in Game 2, Krejci: -9. — THE STATS GUY (@TH2NSTATSGUY) June 4, 2013

@th2nstatsguy Corsi is a waste of time!!

— Steve Kouleas (@stevekouleas) June 4, 2013

…Corsi isn’t WAR. Corsi isn’t the ultimate number that basement-dwelling hockey observers have isolated as the definitive statistic that determines the quality of a player or a team. Corsi is a measure of puck possession. The more shot attempts a team gets, the more time they spend in the offensive zone. Puck-possession correlates with winning and predicts it. That is it. That is the basis of Corsi, which tallies up all attempted shots for both teams when a player is on the ice and turns it into a plus/minus-like number.

I speculate that the general complaint the TH2NSTATSGUY had with Corsi was that in Game 2, Krejci was excellent and Crosby was terrible, and his version of Corsi didn’t bear that out. Many reasons exist as to why this is the case. Perhaps, because Crosby was +7 in the third, it’s an indication that the Penguins earned more cheap zone time in the third period with the game anything but on the line. Here’s the full chart for a blowout loss. Single game Corsi in a blowout is one of the more useless uses of statistics in hockey.

Using the Bruins and Penguins series in an attempt to disprove the merits of zone time is like using a waterfall to disprove gravity. The Pittsburgh Penguins are a good hockey team. They have lots of good hockey players. They are, unlike the Boston Bruins, not particularly good at holding onto the puck and that may be their eventual downfall in the playoffs. They’re capable, and boasting two lottery players, have the shooting talent on two lines to be able to overcome slight puck-possession deficiencies. The issue is that on defence, they are quite slow and predictable and can be eaten up by cycling opponents.

None of that is really new, but the harmonious integration of “viewing” and “statistics” doesn’t appear to have happened quite yet. I can look at a spreadsheet, and I can tell you that the Los Angeles Kings have been fading in puck possession over the last three months. I would not be able to tell you why, for the life of me, because I don’t have experience in the Xs and Os, and coaching, and real on-ice tactics.

As for assertions that numbers are ridiculous and that no general manager would dare to use analytics as a tool, well, how many strawmen need to be knocked down?

***

On the Kings’ front, that team is not doing so well when it comes to moving the puck. Generally, since the Jeff Carter trade, the Kings have been regarded as the best team at having the puck, posting the highest Fenwick Close, Corsi Tied, and Corsi rates of all NHL teams.

In the regular season, their Fenwick Close—the rate of all shots and missed shots that took place in the Kings’ opponents’ ends—was 57.35%, which was higher than even the Detroit Red Wings in 2008.

Those Red Wings carried over that prowess to the playoffs, however, while the Kings have been beaten in possession by both St. Louis and San Jose, and now Chicago, with Jonathan Quick as the only reason the team won playoff rounds. Their Fenwick Close in the postseason dipped to 49.48%, according to Behind the Net, the lowest among Conference Finalists including Pittsburgh.

It doesn’t help that they played the Blues, Sharks and Hawks in the playoffs, but the team is notably slower and noticeably less efficient in the neutral zone than they were this season and last playoffs. Injuries could be a big factor here. Replacing Willie Mitchell with Robyn Regehr doesn’t sound like a good way to continue winning games.

Robert from Jewels From the Crown talked about the team’s zone entry habits, and I’d like to see this expanded on a little more:

@camcharron their greatest strength was neutral zone play this year. that’s totally gone *poof*.

— Robert JftC (@RobertJFTC) June 7, 2013

@camcharron dumping it in so much only works when you out-gain the zone a ton

— Robert JftC (@RobertJFTC) June 7, 2013

The only background knowledge you need here is that entering the zone with possession yields far more shots than simply dumping the puck in and chasing. The second is closer to Darryl Sutter’s coaching style.

You can’t pin all their problems on defence, it seems, unless Regehr’s inability to make a lead pass out of the zone is the sole reason the Kings don’t have enough numbers to get into the offensive zone with possession. This is the part where the “scouts eyes” can tell you what’s wrong with a particular team in a way that statistics can’t, just yet.

We know the Kings are worse, we just don’t know why. The efficient juggernaut that rolled through four teams to the Stanley Cup last year is quite different than the one banking on Quick to steal them games game after game.

Comments (6)

  1. “The more shot attempts a team gets, the more time they spend in the offensive zone.”

    Is there a reason to not just measure offensive zone time instead of shots?

    • The NHL keeps track of one for us, but not the other.

      Corsi is also used as an approximation of scoring chances (and it’s pretty good…not as good as it is an approximation of zone time, though), but no one has looked at the relationship between zone time and scoring chances.

  2. Great article. It’s interesting that what we’ve all slowly been seeing is now an intangible number.

    From a fan’s perspective (so don’t take me too seriously), we’ve been seeing the team slow down since really April. Part of me believes 17 games in 30 days (against mostly solid, playoff contention worthy teams) really sucked out a lot of their energy and made them a much poorer team going into April. Another part believes that they were struggling without the strong leadership skills on the blue line. As a defense first team, they were spending the whole year trying to figure out how to cope with a well seasoned veteran and a young veteran (Doughty) who had never been put in a leadership position before.

    Perhaps the explanation is quite simple: they’ve simply run out of energy. The more rested teams are going to be the more successful teams and they didn’t give themselves a whole lot of rest this post season.

  3. Excellent article… love the fact that you stress that BOTH approachs are necessary for success and true understanding.

    People who ONLY look at stats can easily be deceived by the raw data while people who REFUSE to look at stats can easily be deceived by their own biases and prejudices. It is only by combining the stats with what you see that you can truly better understand them both.

  4. As someone who has watched the Kings all year (basically every game) I can tell you it is a few things:

    1 – Jake Muzzin being young, and untested. He has a bad habit of hesitating with the puck, and/or shooting it into the legs of opposing players who are on the forecheck.

    2 – Willie Mitchell is better than Robyn Regehr, and its not close.

    3 – Missing Stoll in the SJ series, Richards in the Chi series.

    4 – Dwight King has regressed for some reason, and cant hold onto or pass the puck to save his life all of a sudden.

    5 – Something has been wrong with Kopitar since well before the playoffs. I don’t know if its an injury (personally, I suspect it is) nor something else, but he has not been himself for awhile. It has shown up in greater degrees as the Kings have faced better opponents in each round, but I remember sitting in the stands in late April watching him struggle to speed up, get the puck past defenders or make second moves with the puck he’s made his whole career.

  5. The biggest problem, as alluded to above, is Kopitar not driving possession. Throughout his career, on an annual basis, he’ll go through stretches of 30-35 games with less than 5 goals. He just flat out disappears. This usually happens in around November to January, which games-wise is about where we are at now in the season. I do think his minor knee sprain hasn’t helped things, but then again the Kings were the best in puck possession during the regular season. (Carter being in the best shape of his life and putting up almost 30 goals at an almost 60 goals pace is probably the best reason why.)

    It’s tough watching him when he goes into his annual slumps, and some have suggested that it’s a mental issue; that he doesn’t have the drive to dominate games. That, in his words, paraphrased, “No one has to be great if everyone is good.” But others are adamant that putting all of the Kings struggles on him is unfair, especially after a Cup-winning season.

    It’s an interesting and somewhat enigmatic situation that the Kings are in, but savvy Kings fans would have predicted this outcome as a possibility if you had asked them at the start of playoffs.

Leave a Reply

Your email address will not be published. Required fields are marked *