Jump to content

Nice article on the pros and cons of advanced stats.....


jammer2
 Share

Recommended Posts

  One of my favourite writers Steve Simmonds of the Toronto Sun explores the pros and cons of advanced stats. I've promised myself I will immerse myself in these stats come September. It's not the whole story, but I do believe they are an important variable. As per usual, a thought provoking piece by Simmonds, who is my second favourite hockey writer next to Metzler. More often than not, I end up agreeing with him, he has made some strong arguments over the years. I just think he has a nice grip on the sport in general.....but he does come across as a dweeb on the hockey panelists show The Reporters on TSN.

 

 http://www.torontosun.com/2014/05/20/why-hockeys-trendy-advanced-stats-are-a-numbers-game

 

 I know a few people here are intrigued by advanced stats, anyone have a site they can recommend to begin my quest for knowledge?

Link to comment
Share on other sites

  One of my favourite writers Steve Simmonds of the Toronto Sun explores the pros and cons of advanced stats. I've promised myself I will immerse myself in these stats come September. It's not the whole story, but I do believe they are an important variable. As per usual, a thought provoking piece by Simmonds, who is my second favourite hockey writer next to Metzler. More often than not, I end up agreeing with him, he has made some strong arguments over the years. I just think he has a nice grip on the sport in general.....but he does come across as a dweeb on the hockey panelists show The Reporters on TSN.

 

 http://www.torontosun.com/2014/05/20/why-hockeys-trendy-advanced-stats-are-a-numbers-game

 

 I know a few people here are intrigued by advanced stats, anyone have a site they can recommend to begin my quest for knowledge?

http://www.extraskater.com/

 

Advanced stats can be useful I suppose. but sometimes, they blur the picture. Braun was out best defenseman after Vlasic went down, but being partnered with Stuart 100% of the time, his stats look 2nd worst on the team.

Link to comment
Share on other sites

I don't know what the hell Steve Simmons is talking about here, and it's pretty clear that he doesn't either.

 

 


When I asked the stats man about the discrepancy between what we’d seen and what the numbers showed, he answered: “Sample size.”

 

Not a single stats man would offer such an answer, since it's completely non-sensical within the context of the story Simmons is sharing. It was one game!

 

 


In Game 1 of last year’s playoff series between the Leafs and the Boston Bruins, Toronto was badly outplayed. Only one Leafs player seemed capable of competing at that level — James van Riemsdyk. So, curious after the game, I asked my stats friend who had the best numbers for Toronto.

It so happened van Riemsdyk had them, but his numbers were just a percentage point better than Phil Kessel, who I thought had a dreadful game. Again, I asked: “How can the numbers be reliable, when two players can have such varying games and end with similar statistics?”

“Sample size,” I was told.

 

Again, "sample size" doesn't even make sense as a response within the context of the conversation. But, even more importantly, Simmons is grotesquely incorrect about the game in question. The Leafs Corsi numbers for the game he's moaning about:

 

Screen-Shot-2014-05-20-at-8.40.44-PM.png

 

Sample size be damned, the numbers show that JVR was taken apart piece by piece at ES, but not as badly as Kessel. Yet, in Simmons' story, VanRiemsdyk and Kessel both had good Corsi numbers. Either Simmons is lying/making up ****, or he just has no idea what he's talking about.

 

The most painful part of the article is where Simmons goes on at length, describing plays, with pucks bouncing, being cleared, all with no stats which can be clearly shown, yet all are things which happen and that players can affect. And there, where his (probably fictional) stats guy should be saying:

 

SAMPLE SIZE!

 

As the sample size increases, and given enough time, looking at which team drove the puck north more often and generated more shots while he was on the ice, where he started his shifts, and the competition he played against, these plays begin to coalesce into a better picture of his ability.

 

 


The Maple Leafs were among the worst Corsi and Fenwick teams (the best known of the advanced statistics) in the NHL this season. When they collapsed, the stats mavens were almost gleeful. They knew it was coming. They called it. The Leafs were their convenient poster-boy for the changing way to interpret hockey. And an easy target.

 

That's my favorite part of the article right there, where he says that anybody could see the Leafs were a bad hockey team that would come down to earth. Certainly, Mr Simmons wasn't fooled by the amount they were winning against the house early on:

 

 


"Good thing the Leafs don't play in the CHL. The CORSI hockey league. They're doing just fine in NHL, though."

 

https://twitter.com/simmonssteve/statuses/395375707821309952

 

Oops.

Edited by JR Ewing
Link to comment
Share on other sites

@JR Ewing Well, to be fair to Steve, he indicates he does not know a lot about the topic, but wants to learn more. He's in the same boat as me. I intend to makes these stats a part of my repertoire moving forward, they can be informative and shed light on area's that don't show up in a box score.

Link to comment
Share on other sites

http://www.extraskater.com/

 

Advanced stats can be useful I suppose. but sometimes, they blur the picture. Braun was out best defenseman after Vlasic went down, but being partnered with Stuart 100% of the time, his stats look 2nd worst on the team.

 

The stats also show that Braun had a brutal assignment: toughest Zone Starts as well as Quality of Competition. With the teams GF% only being half the CF% while he was on the ice, there's some bad luck there, too. When the Zone Shift Adjusted numbers come out, it would be fun to have a peek at them.

Link to comment
Share on other sites

The stats also show that Braun had a brutal assignment: toughest Zone Starts as well as Quality of Competition. With the teams GF% only being half the CF% while he was on the ice, there's some bad luck there, too. When the Zone Shift Adjusted numbers come out, it would be fun to have a peek at them.

Daryll, any chance you could do some educating to me on advanced stats?

 

I read them all the time, but do not quite understand them, at least, anywhere near as well as you. Corsi, Fenwick, etc

 

All I know is, a ton of sharks fans were going blah blah on Braun's numbers = bad personal performance and I had to wonder WTF they were smoking. The poor kid was stuck with Brad Stuart. Most of the good posters and shark fans called them full of crap

Link to comment
Share on other sites

I think that's a good idea, actually. Laymen's terms, then...

 

Corsi - Put simply, it's the shot differential while a player is on the ice. It tells us which players are (and aren't) pushing the puck north while on the ice. This includes not only shots, but missed shots (including those which hit the post), and shots which are blocked. It's a proxy for possession, since the only way to accumulate these counter numbers is to have the puck. There is a consistent and strong link between teams having a high Corsi and a higher win%. You can buck the odds for awhile, even a whole season sometimes, but there's only so long it can last before you plummet back to earth.

 

Fenwick - This is basically Corsi, but with blocked shots removed from the equation. I find it less instructive than Corsi. Yes, shot blocking is a skill, but it still means the other team possessed the puck. What would you rather your team did? Not possess the puck a lot, but block plenty of shots, or would you rather they had the puck a lot, press the play and get more shots blocked by way of it. I know what I'd prefer.

 

Corsi For% - This is how many shot-attempts a team had relative to the total amount of shot-attempts generated by both teams in that game. So, if there were 100 shots attempts between both teams, with each club having 50 a piece, then each had a Corsi% of 50. Even 5 percentage points is a LOT here.

 

GF% - This is the same idea as Cosi%, but applied to goals scored. Generally speaking, these two numbers should be somewhat close to each other, or else luck is playing into the equation more than anything else, usually. After all, picture a guy who doesn't really have the puck a lot, but his team scores a lot when he's on the ice. Let's use a player as an example:

 

CF% - 46.9%

GF% - 57.7%

 

If you guessed Nathan MacKinnon, you were right. He's a hell of a prospect, but we still have a kid who spends majority of the time in his own end, yet the Avs cashed in for 57.7% of the goals while he was on the ice. Do we have some ideas as to why that is? Yeah, actually:

 

Even-Strength Save% while MacKinnon was on the ice: .953

 

WHEN (not if) that freakish ES SV% comes back down to earth, the goals against will pile up, and the Avs will follow. For pool players: watch out for players whose GF% greatly exceeds their CF%:

 

Screen-Shot-2014-03-25-at-3.00.57-PM.png

 

Look at those numbers fall the next year. On the flip-side finding players with GF% numbers much lower than their CF% represent a chance to buy low, as he'll probably see dramatic "improvement".

 

 

Relative Corsi (CorsiRel) - Shows how a player did compared to his teammates. Even a horrible team can have some effective players, and vice versa. If you play for a powerhouse where most of your teammates have Corsi numbers well into the positives, and you're barely above zero, you've done poorly relative to them.

 

But... We run into.

 

Zone Starts - Which is really more accurately put it were called Offensive Zone Starts. This has a huge impact on a player's Corsi numbers. If you have a lot of offensive zone starts, you have a greater opportunity to pile up shot attempts compared to how many you may expect to have against you. The opposite holds true for those players who start next to their net a lot.

Edited by JR Ewing
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...