Chicago Blackhawks: ALL the Analytics

[GKJ]

This is from Mark Lazarus, saw James Mirtle tweet it.

http://www.suntimes.com/sports/hock...-secret-formula-for-success.html#.U2LcNuZdXWp


I think the best part is that there are people still think this doesn't happen.

Click to read more...

 

[Puck possession wins]

This is from Mark Lazarus, saw James Mirtle tweet it.

http://www.suntimes.com/sports/hock...-secret-formula-for-success.html#.U2LcNuZdXWp


I think the best part is that there are people still think this doesn't happen.

My questiom, is should we call it Advanced stats or Advanced Analytics?

 

[Bear of Bad News]

Thanks for posting this - great article. I'm not surprised that Bowman's on top of this stuff.

One of the reasons (some) people slag advanced metrics around here is because they're not seeing the current top of the line things that teams (who obviously have much better access to data) are doing. Obviously, teams have a vested interest in not letting what they're up to become part of the public domain.

It's like pointing to a 1998 Ford Taurus and saying "See? Cars today are the worst."

 

[TheDevilMadeMe]

One of the reason people slag metrics is because pop proponents of metrics write nonsense like this:

To the advanced stats crowd, the Colorado Avalanche are the exception that proves the rule. All season long, they defied the math by having an abnormally high shooting percentage and an abnormally high save percentage. The stat known as PDO adds these two numbers together, and the average is about 100. So if a team has a very high or very low PDO, it should eventually regress to the mean of 100 over the course of the season. Sure enough, the first-place Avalanche (with a 102.2 PDO in the regular season that offset a subpar Corsi percentage) lost in the first round to the fourth-place Minnesota Wild after getting outshot by an average of 33-25 each game.

Like a gambler on a hot streak at the blackjack table, the Avalanche defied the math for months. But in the long run, the math almost always wins.

Like, seriously? The main reason Corsi and Fenwick are considered to have higher predictive abilities over actual goals scored is because the sample size of events is much, much higher than goals actually scored. And he's using a series where the winner was determined by a single goal in Game 7 OT to brag about how great they are? Madness. It's as bad as the proponents of such stats bragging about how 5 of 8 teams that won in the first round had better scored-adjusted Fenwick than their opponents. How is 5 of 8 statistically significant, especially when we were a game 7 OT goal away from it being 4 of 8?

Rant over. The rest of the article, the part that focuses on the Blackhawks' use of analytics is fascinating. I know one of the major complaints in forming coherent hockey analytics is that the raw data provided by the NHL is incomplete, so there is only so much mathematical analysis that can be done. Seems like individual teams like the Blackhawks actually are tracking a lot more (this isn't surprising), but they ain't sharing!

 

[Bear of Bad News]

Agreed - a lot of the people writing about this stuff do it a disservice. It's important to know what the metrics and analytics are capable of, but it's more important to know what they're not capable of.

Long-term, it doesn't even stand to reason (to me) that a PDO should regress to 1000 - the save percentage should regress to a goaltender's "true" save percentage, and the shooting percentage should regress similarly.

Back to the author's (flawed) point - if "heads" comes up on a fair coin ten times in succession, that doesn't mean that tails is "due", and it doesn't mean that on a going-forward basis, we should expect anything other than "heads" 50% of the time.

By the way, that quote has one of my pet peeves - an exception does not (cannot) "prove a rule". The two etymologies of that phrase that I've heard (and think have credibility) are (1) that "prove" used to be "probe", and that the exception probes (tests) the rule.

More likely, it's suggestive that the exception proves that the rule exists. For instance, if you see a sign saying "Free Parking on Sundays", it suggests that there is a (hidden) rule that parking on non-Sundays is not free.

Anyhow, it's been bastardized and is now misused in the same sense that "literally" is misused.

 

[No Fun Shogun]

Saw Down Goes Brown just tweet about it, and posted the article on the Hawks board. Definitely interesting read.

 

[GKJ]

One of the reason people slag metrics is because pop proponents of metrics write nonsense like this:



Like, seriously? The main reason Corsi and Fenwick are considered to have higher predictive abilities over actual goals scored is because the sample size of events is much, much higher than goals actually scored. And he's using a series where the winner was determined by a single goal in Game 7 OT to brag about how great they are? Madness. It's as bad as the proponents of such stats bragging about how 5 of 8 teams that won in the first round had better scored-adjusted Fenwick than their opponents. How is 5 of 8 statistically significant, especially when we were a game 7 OT goal away from it being 4 of 8?

I was rather fiercely opposed on this the other day. You have the numbers, and you have the team with a Vezina candidate who lost to a Wild team on their 3rd and 4th goalies and blew 4 leads in Game 7. That one didn't even seem to matter which side of the ledger you were on.

You score-adjusted Fenwick point - guilty. Better team did ultimately still win no matter how close it was.

 

[The Mighty Duck Man]

One of the reason people slag metrics is because pop proponents of metrics write nonsense like this:



Like, seriously? The main reason Corsi and Fenwick are considered to have higher predictive abilities over actual goals scored is because the sample size of events is much, much higher than goals actually scored. And he's using a series where the winner was determined by a single goal in Game 7 OT to brag about how great they are? Madness. It's as bad as the proponents of such stats bragging about how 5 of 8 teams that won in the first round had better scored-adjusted Fenwick than their opponents. How is 5 of 8 statistically significant, especially when we were a game 7 OT goal away from it being 4 of 8?

Rant over. The rest of the article, the part that focuses on the Blackhawks' use of analytics is fascinating. I know one of the major complaints in forming coherent hockey analytics is that the raw data provided by the NHL is incomplete, so there is only so much mathematical analysis that can be done. Seems like individual teams like the Blackhawks actually are tracking a lot more (this isn't surprising), but they ain't sharing!

It's become an unfortunate double standard that many(typically the people in it for the false sense of superiority) keep going to. If the Avs win that game, or even sweep the series, and it's all about how the luck continued and it'll eventually end. But the Wild won, and it's all about how the numbers were correct.

Otherwise, yes, fantastic article. I suspect most teams keep track of a lot of these things, including possession, and all in all we don't give the league enough credit for being ahead of the game. Well, maybe not the league itself, but many of it's teams. It may not be public, but I suspect teams are well ahead of where we are, as they should.

 

[Bear of Bad News]

Washington Post article on what Toronto and Chicago tell us about advanced analytics in the NHL:

http://www.washingtonpost.com/news/...-tell-us-about-advanced-analytics-in-the-nhl/

Seems somewhat ironic that an article on advanced statistics is using a relatively small sample to prove its point. Good article, though.

 

[The Mighty Duck Man]

Washington Post article on what Toronto and Chicago tell us about advanced analytics in the NHL:

http://www.washingtonpost.com/news/...-tell-us-about-advanced-analytics-in-the-nhl/

Seems somewhat ironic that an article on advanced statistics is using a relatively small sample to prove its point. Good article, though.

It's not a bad article, but it's very easy to cherrypick a top team using analytics and succeeding and pick a bad one not using them and failing. Many forget Jay Feaster was a pretty big proponent of advanced stats and analytics, and look where it landed him. I believe the Sabres were one of the teams often mentioned to be involved as well. But they don't talk about them, just the ones like the Blackhawks.

Oh, and can writers stop using the term "moneypuck"? It's a great way to show they've never read Moneyball and don't remotely understand the differing concepts.

 

[Caeldan]

Agreed - a lot of the people writing about this stuff do it a disservice. It's important to know what the metrics and analytics are capable of, but it's more important to know what they're not capable of.

Long-term, it doesn't even stand to reason (to me) that a PDO should regress to 1000 - the save percentage should regress to a goaltender's "true" save percentage, and the shooting percentage should regress similarly.

Back to the author's (flawed) point - if "heads" comes up on a fair coin ten times in succession, that doesn't mean that tails is "due", and it doesn't mean that on a going-forward basis, we should expect anything other than "heads" 50% of the time.

By the way, that quote has one of my pet peeves - an exception does not (cannot) "prove a rule". The two etymologies of that phrase that I've heard (and think have credibility) are (1) that "prove" used to be "probe", and that the exception probes (tests) the rule.

More likely, it's suggestive that the exception proves that the rule exists. For instance, if you see a sign saying "Free Parking on Sundays", it suggests that there is a (hidden) rule that parking on non-Sundays is not free.

Anyhow, it's been bastardized and is now misused in the same sense that "literally" is misused.

Well that just begs the question...
Okay I'll stop myself here

I've never actually looked closely at the exception that proves the rule comment, usually just use it tongue in cheek.

As far as the article goes... I find it interesting that they single out the leafs - didn't they have a coach recently for a while who was trying to apply advanced analytics? Ron Wilson wasn't it? Perhaps that's where their current push back comes from.

 

[BraveCanadian]

One of the reason people slag metrics is because pop proponents of metrics write nonsense like this:



Like, seriously? The main reason Corsi and Fenwick are considered to have higher predictive abilities over actual goals scored is because the sample size of events is much, much higher than goals actually scored. And he's using a series where the winner was determined by a single goal in Game 7 OT to brag about how great they are? Madness. It's as bad as the proponents of such stats bragging about how 5 of 8 teams that won in the first round had better scored-adjusted Fenwick than their opponents. How is 5 of 8 statistically significant, especially when we were a game 7 OT goal away from it being 4 of 8?

Rant over. The rest of the article, the part that focuses on the Blackhawks' use of analytics is fascinating. I know one of the major complaints in forming coherent hockey analytics is that the raw data provided by the NHL is incomplete, so there is only so much mathematical analysis that can be done. Seems like individual teams like the Blackhawks actually are tracking a lot more (this isn't surprising), but they ain't sharing!

Well said all around.

It would really be interesting to see what the NHL teams have identified as being important and worth tracking on their own.

 

[JurequeparJuraj]

deleted sry for bump (still a good read here: http://www.cbssports.com/nhl/eye-on...m-stan-bowman-says-analytics-give-team-leg-up )