How Screwed are the Flames?

[DJones]

Why is pdo expected to go to 100 eventually? It's two completely unrelated stats. Why would good goal tending have any correlation with shooting percent?

[Jay Random]

There is no correlation. By definition, if you add up all the save percentages and shooting percentages over the entire league, they average out to 100 per team. But then, if you add up all the wins and losses (ignoring pity points), they average out to 82 per team; and nobody expects teams to regress to 41 wins and 41 losses every year.

Where the regression happens is that the standard deviation of PDO over a full season is smaller than the standard deviation over short stretches of games. Similarly, a team can easily go 8-2-0 in a 10-game stretch, but only a handful of teams have ever had an .800 winning percentage over a full season. If a team puts up an especially high or low PDO for a full year, chances are there is more than luck involved.

[Enoch Root]

Originally Posted by DJones

Why is pdo expected to go to 100 eventually? It's two completely unrelated stats. Why would good goal tending have any correlation with shooting percent?

The two stats are essentially reciprocals of each other. The assumption is that each team should somehow get an equal dose of the two in combination.

It is basically saying that your position on two different distribution charts should average out. If your goaltending is better than average, then your shooting percentage should be below average in order to compensate.

Pretty ridiculous when you break it down.

However, laws of large numbers. Because the league aggregate adds up to 100, and because there are so many players and variables involved for each team, the actual results don't vary a great deal from the mean - they are so muddied, that every team looks brown.

As a result, data minors looking for something to talk about, have a perfect opportunity here: greater than 100 - lucky. Less than 100 - unlucky.

Four legs good...

[edslunch]

Originally Posted by Freeway

If talent is evenly distributed, PDO predicts that over an INFINITE number of games, every team will have 100 Corsi (e.g., shooting/save percentages will regress/progress to that point). There is no consensus over what a suitable sample size is in order to say somebody was "right," nor is there a consensus over whether NHL talent is evenly distributed (it probably isn't).

PDO predicts nothing. Every shot is either saved or not, so it's simple math that for the league overall, SV% = 100 - SH%. For the league as a whole those numbers are directly (inversely) related to each other and the league average PDO will always be exactly 100.

This simply does not apply to an individual team, no matter how infinite your sample size. Does having a good goalie make your shooters worse? Does having a lot of snipers make your goalie a sieve?? BS, there is zero correlation.

If every team had exactly the same talent then sure, you'd expect them all to score and save the same on aggregate and have a PDO of 100. But that is an absurd condition that clearly does not and never will exist.

The only thing PDO does is suggest if you are getting better than average goaltending and/or shooting. Guess what? Some teams do get better than average tending and/or shooting - they're called good teams. PDO in fact is a poor indicator even of that, since bad SV% cancels out good SH% and vice versa. PDO obscures the data rather that clarifying it. If you want to know if team is getting historically rare SV% or SH% look at SV% and SH% - PDO tells you #####

[drewtastic]

I don't know how anyone can say that Kris Russell is highly overrated...


Sent from my iPhone using Tapatalk

[Flames Fan, Ph.D.]

Originally Posted by Phil Russell

"Statistics don't bleed" is a saying used to describe the failings of statistics-obsession in the human sciences. You can't model something as complex as human behavior or, in this case, a hockey game. Statistics don't bleed, but the Flames do and that's why they win. Sacrifice, team-work, opportunism, confidence etc. Where do these things fit in the model? And if they don't fit, then you are assuming they have no impact on winning or losing a hockey game. Which is, of course, ridiculous.

Agree in general but I think the bolded part is not as complicated as it may seem.

I've seen crazy mathematical models these days to describe the actions of cells, so the power is there. Hockey is, generally speaking, a game of finite options that can be modeled and have probabilities associated with them (eg. pass, shoot, dump the puck, stick-handle, turnover). Therefore, I think it is a great opportunity for some comprehensive statistical modeling. I'm pro-analytics.

The gripe I have with the current analytics movement is that, by and large, they constantly jam the real game outcomes into the Corsi / Fenwick boxes. Rather, they need to work and find ways to modify Corsi / Fenwick so that they better reflect the real world.

[Robo]

Originally Posted by drewtastic

I don't know how anyone can say that Kris Russell is highly overrated...


Sent from my iPhone using Tapatalk

didn't read the article nor will I but did he seriously say that?

[Rerun]

What was our PDO in the Detroit win?

[Geeoff]

Originally Posted by Quincy Egg

Except we ended up being right about the Avalanche, and most of us realize what the definition of an adequate sample size is.

The Avs made the playoffs...

They lost Paul Stastny in the off-season and didn't replace him and McKinnon is having a bit of a sophomore slump. If they have a good year next year, does that make you wrong again?

[drewtastic]

Originally Posted by Robo

didn't read the article nor will I but did he seriously say that?

I had never read one of his articles before, so curiosity got the better of me. At least now I am fully aware of his spiteful negativity and will be able to avoid it.

This was the quote about Russell (a toss-away reference within his analysis of Gio & Brodie):

"...And by dint of the fact that they're the only actually good defensemen on the team (with all apologies to the stunningly overrated Kris Russell)..."

Edit: just saw this Duhatchek article on Twitter about Russell. Posting for comparison http://m.theglobeandmail.com/sports/...+Article+Links


Sent from my iPhone using Tapatalk

[Rerun]

[GranteedEV]

Originally Posted by Geeoff

The Avs made the playoffs...

They lost Paul Stastny in the off-season and didn't replace him and McKinnon is having a bit of a sophomore slump. If they have a good year next year, does that make you wrong again?

McKinnon is also sliding over from winger alongside Statsny to top 6 center. Far from an easy transition for a 19 year old.

[sa226]

So I kind of accidentally read this article. I didn't want to give him my "click" but it came up in my Google News feed and the headline looked familiar but I didn't realize that it was this article.

Anyway.

It was so blatantly obvious that the article is designed to get a rise out of Flames fans. The style of writing, the hidden and subtle "Brodie doesn't skate like Subban or Karlsson" The dig on Wideman, the dismissing of some pretty good players as "average at best" and on and on.

If he had more talent for writing he could be more effective with his textual diarrhea. This was just so in your face obvious that I didn't find it as irritating as I thought I might have.

The tears are delicious.

[sa226]

Dp

[Gaskal]

Originally Posted by GranteedEV

McKinnon is also sliding over from winger alongside Statsny to top 6 center. Far from an easy transition for a 19 year old.

Hey, if Mony can make it look easy...

[Beatle17]

[QUOTE=bigehn;5174380]Are these Flames for Real?
Full Disclaimer:
1. I am a huge Hockey fan and a huge Flames fan.
2. I am a believer in the usefulness of advance stats and I work in the analytics industry. I believe in the power of information.

Thanks for this well produced post. The breakdown of the chart was perfect, and for a guy like me who doesn't put too much stock in stats such as Corsi/Fenwick, this makes so much sense.

I would bet that teams use information like this way more than the stats fans talk about. Excellent.

[woob]

I admittedly don't know much about Flames Nation, so if this post is out to lunch, please tell me so. That being said, why is a guy who seems to have so much dislike, zero faith in, and really not that much positive to say about the Flames, writing for a website called Flames Nation?

Don't get me wrong, a good fan website SHOULD have some criticism of the home team, but I read a few more of his articles after reading the one the OP shared here, and he sounds like he'd be down cruising the Red Mile if the Flames got mathematically eliminated, shouting "I told you so!!!."

And I have a hard time believing someone would invest that much time and effort just to troll. Are all of the advance stats guys like this??

[RyZ]

It is guys like Lambert, Drance and Mudcrutch that keep analytics from gaining wide spread acceptance from most fans. It's like they cant see the forrest for the trees and replace wins, losses, goals and assists with numbers that may or may not kinda represent puck possesion of all things. Most fans are already skeptical as it is and then you have these guys acting like marketers and it completely turns a lot of people off.

[CroFlames]

Hartley after the Wings win basically dismissed advanced stats saying that they don't play a possession game.

Last time I looked, he makes more dough than all the advanced stats guys combined. Oh, and he also has his team in 3rd in the Pacific.

Do I believe pencil pushing advanced stats guys or a Stanley Cup winning coach who coached both pre and post lockout.

[drewtastic]

Originally Posted by CroFlames

Hartley after the Wings win basically dismissed advanced stats saying that they don't play a possession game.

He didn't REALLY dismiss stats altogether, he was more making fun of people who are 'mind-bottled' by the Flames' winning ability given their low puck-possession stats.

But he did say (a) that blocked shots helps offset potential low puck possession and (b) that he does review stats daily. These comments could also have been completely sarcastic, but I sensed some sincerity in them.

http://video.flames.nhl.com/videocen...72528&catid=11


Sent from my iPhone using Tapatalk