Ongoing Stats and Analytical Discussion Thread: Battle of the Defense

[devilsblood]

My post was about how poor word choice has sidetracked discussions about this topic. You respond, using my new word choice as an opportunity for a pun, full stop, nothing else to your post. But it did get me to stop discussing this topic with you, so I guess I have to thank you for hollowing out a nice divot with a pillow in that brick wall.

I also picked only bits and pieces of sentences from within the argument and strong them together into a new sentence. I thought the tactic was so obviously ridiculous that we may all get a chuckle from it. Really wasn't looking to step on toes.

 

[NJDevs26]

Relying on analytics lol. I don't think managements takes this stuff that serious

you should tell that to Josh Harris & co. because the Devils (& sixers) are definitely heavily engaged in that department.

This is something I want to know from the game thread, and I'd rather have the discussion in the advanced stats thread. Why if the Devils ARE into analytics do they make so many decisions that are seemingly anti-analytics? John Moore's continued use and overuse being at the top of that list. It defies logic, I'm forced to conclude either they don't use analytics as much as they claim, or they have a completely different analytics system than the one most people here cite.

 

[Feed Me A Stray Cat]

This is something I want to know from the game thread, and I'd rather have the discussion in the advanced stats thread. Why if the Devils ARE into analytics do they make so many decisions that are seemingly anti-analytics? John Moore's continued use and overuse being at the top of that list. It defies logic, I'm forced to conclude either they don't use analytics as much as they claim, or they have a completely different analytics system than the one most people here cite.

I've wondered this myself.

The publicly available stats for hockey analytics are pretty poor. It's basically all related to shots attempted and looking at things on a per 60 minute basis. That definitely adds more texture than simply relying on boxcar stats, but in the end I would hope someone of Sunny Mehta's ilk would have more sophisticated numbers.

The real breakthrough in publicly available hockey analytics will be when the NHL starts tracking actual puck possession and passing. I suspect this is what the NHL analytics departments are already tracking themselves. Things like:

-Team puck possession when a player is on the ice
-Individual player puck possession
-Pass completion rate
-Number of passes attempted
-Number of passes attempted in high danger areas
-Number of shots attempted in high danger areas
-Number of passes leading directly to shots
-Number of passes leading directly to shots in high danger areas
-Zone entries
-Zone exits

Those are things we really care about. Corsi is an okay approximation for them, because being good at all of those things should lead to more shots, but it's only an approximation.

I remember reading a year or two ago that the Devils were tracking high danger scoring chances. It was either DeBoer or Hynes who talked about how the team was doing well in that regard despite a losing streak.

 

[Feed Me A Stray Cat]

To illustrate a point about how pervasive luck can be, take a look at this rudimentary excel spreadsheet I put together.
http://www.filehosting.org/file/details/642551/Book1.xlsx

30 teams who each play 82 games. A win or loss is decided by a coin flip. Pure chance. The logical assumption is that every team will end up with 41 wins, right?

No. Some teams end up with 49 wins. Others with 35 (press f9 to run a new scenario). Despite these teams being of equal ability. How is that possible? Because a sample of 82 trials is not enough to tease out the tail ends of the distributions for a population of 30.

Also take a look at how many streaks there are. By pure chance, some teams will go on seven game winning and losing streaks.

This is not a perfect analogy to real life. NHL teams don't all have a 50% chance of winning an individual game, and those odds aren't fixed throughout the season.

But what this should help prove to people is that it is possible - and quite common - for teams and players to experience statistical abnormalities across a full season.

 

[Kurt Cobain]

This is something I want to know from the game thread, and I'd rather have the discussion in the advanced stats thread. Why if the Devils ARE into analytics do they make so many decisions that are seemingly anti-analytics? John Moore's continued use and overuse being at the top of that list. It defies logic, I'm forced to conclude either they don't use analytics as much as they claim, or they have a completely different analytics system than the one most people here cite.

I believe every organization relies on this stuff much less the fans who obsessed with them. It's not just us. I also find people relying on this stuff and not just critically anazlying what they see on the ice to be so incredibly foolish.

 

[Feed Me A Stray Cat]

I believe every organization relies on this stuff much less the fans who obsessed with them. It's not just us. I also find people relying on this stuff and not just critically anazlying what they see on the ice to be so incredibly foolish.

I disagree. Most teams have established analytics departments at this point. Also look at the influx of new, young GMs and assistant GMs who have analytics backgrounds.

It's a process. It took baseball a good 15 years before everyone got on board, and now almost every team is pretty open about how much they use analytics. There were plenty of people calling it a fad in the mid-2000s.

Hockey is a different sport than baseball, and harder to analyze statistically, but in 5-10 years I suspect we'll be seeing stats like Corsi Rel used just as much as +/-.

 

[SteveCangialosi123]

I believe every organization relies on this stuff much less the fans who obsessed with them. It's not just us. I also find people relying on this stuff and not just critically anazlying what they see on the ice to be so incredibly foolish.

You've come to the conclusion that Ben Lovejoy is a good defenseman. It may be time to start looking at analytics. But the funny thing is that the eye test is arguably even worse than his advanced stats which are anemic.

 

[Kurt Cobain]

Of course I know you disagree lol. Just because teams have an analytics department that could just be a guy and a laptop lol.

 

[Kurt Cobain]

You've come to the conclusion that Ben Lovejoy is a good defenseman. It may be time to start looking at analytics. But the funny thing is that the eye test is arguably even worse than his advanced stats which are anemic.

I said he's not as bad your all trashing him as never said he was special. Work on your reading skills buddy.i

Lol I never even said he was good lol I just said he does his job and he's way overplayed.

 

[devilsblood]

To illustrate a point about how pervasive luck can be, take a look at this rudimentary excel spreadsheet I put together.
http://www.filehosting.org/file/details/642551/Book1.xlsx

30 teams who each play 82 games. A win or loss is decided by a coin flip. Pure chance. The logical assumption is that every team will end up with 41 wins, right?

No. Some teams end up with 49 wins. Others with 35 (press f9 to run a new scenario). Despite these teams being of equal ability. How is that possible? Because a sample of 82 trials is not enough to tease out the tail ends of the distributions for a population of 30.

Also take a look at how many streaks there are. By pure chance, some teams will go on seven game winning and losing streaks.

This is not a perfect analogy to real life. NHL teams don't all have a 50% chance of winning an individual game, and those odds aren't fixed throughout the season.

But what this should help prove to people is that it is possible - and quite common - for teams and players to experience statistical abnormalities across a full season.

Absolutely not.

 

[Feed Me A Stray Cat]

Of course I know you disagree lol. Just because teams have an analytics department that could just be a guy and a laptop lol.

Based on the rapid progress and dissemination of stats in sports like baseball, basketball and soccer, I think it's safe to say it's going and will continue to go that way in hockey.

Winning in hockey creates a lot of money. Profit driven owners aren't going to turn their noses at receiving more and better information that could lead to more money.

 

[Feed Me A Stray Cat]

Absolutely not.

Then you're agreeing that luck plays a large factor in outcomes over a full season?

 

[devilsblood]

I disagree. Most teams have established analytics departments at this point. Also look at the influx of new, young GMs and assistant GMs who have analytics backgrounds.

It's a process. It took baseball a good 15 years before everyone got on board, and now almost every team is pretty open about how much they use analytics. There were plenty of people calling it a fad in the mid-2000s.

Hockey is a different sport than baseball, and harder to analyze statistically, but in 5-10 years I suspect we'll be seeing stats like Corsi Rel used just as much as +/-.

Sports have always been about stats. But in the past there were more surface oriented stats. Analytics are not revolutionary really, they are taking what has always been done, but are now doing it at a much deeper level.

The rise of OPS and WHiP, with the devaluation of batting average in baseball are prime examples.

 

[Triumph]

Let's imagine that we have a league of 30 teams who use 'the eye test' - this means traditional scouting, qualitative ranking of players (rating players' speed, shot, etc. on a 1-10 scale, for example), and 5 stats - games played, goals, assists, +/-, ice time.

How does a team figure to get ahead of the competition by using the same exact methods everyone else is? By just being better at those things? Eye-testing harder? I mean sure, by virtue of lottery balls falling, careers ended by injury, and wives who don't like living in a particular city, some teams are going to end up better than one another regardless of the scouts. And sure, you don't need analytics to tell you that Taylor Hall is better than Adam Larsson. But without something different than other teams, you are just spinning the wheel of fortune any time you make a decision, hoping you avoid Bankrupt.

 

[devilsblood]

Then you're agreeing that luck plays a large factor in outcomes over a full season?

How does me challenging your premise lead you to think I agree with your conclusion?

 

[Kurt Cobain]

Based on the rapid progress and dissemination of stats in sports like baseball, basketball and soccer, I think it's safe to say it's going and will continue to go that way in hockey.

Winning in hockey creates a lot of money. Profit driven owners aren't going to turn their noses at receiving more and better information that could lead to more money.

I don't know what this has to do with my comment, but I still stand by my previous statments make up no more that 25% of player evaluation of even. I don't believe signing or trades are made based on them, if anything they're used by coaches to help players figure out aspects of their game

 

[R8Devs]

I don't know what this has to do with my comment, but I still stand by my previous statments make up no more that 25% of player evaluation of even. I don't believe signing or trades are made based on them, if anything they're used by coaches to help players figure out aspects of their game

http://www.espn.com/blog/craig-custance/insider/post?id=8534

 

[devilsblood]

Let's imagine that we have a league of 30 teams who use 'the eye test' - this means traditional scouting, qualitative ranking of players (rating players' speed, shot, etc. on a 1-10 scale, for example), and 5 stats - games played, goals, assists, +/-, ice time.

How does a team figure to get ahead of the competition by using the same exact methods everyone else is? By just being better at those things? Eye-testing harder? I mean sure, by virtue of lottery balls falling, careers ended by injury, and wives who don't like living in a particular city, some teams are going to end up better than one another regardless of the scouts. And sure, you don't need analytics to tell you that Taylor Hall is better than Adam Larsson. But without something different than other teams, you are just spinning the wheel of fortune any time you make a decision, hoping you avoid Bankrupt.

Having better scouts?

 

[Kurt Cobain]

http://www.espn.com/blog/craig-custance/insider/post?id=8534

what this is? We traded for hall because of analytics not because goal scoring was needed. If that's what your trying to say.

 

[R8Devs]

you literally said the only thing that comes out of analytics is that they base lineups off them. and then you made up some number saying that's how much it factors into their decision making. that article shows that Ray gives merit to it. i'm not saying the trade was entirely based of off it but it did play a role. Also based on how the Sixers are run I really doubt Josh Harris would hire a GM that didn't show credence to analytics and their importance.

 

[Kurt Cobain]

dude you literally said the only thing that comes out of analytics is that they base lineups of them. and then you made up some number saying that's how much it factors into their decision making. that article shows that Ray gives merit to it.

We traded for hall because of his offensive abilities. They didn't go through every roster find the analytics they liked best for a particular player and it happened to be hall. That's what I'm saying and that's a fact.

 

[R8Devs]

We traded for hall because of his offensive abilities. They didn't go through every roster find the analytics they liked best for a particular player and it happened to be hall. That's what I'm saying and that's a fact.

I didn't say that they did that. It says right in the article that ray requested Hall's statistics. It's not like they didn't know of Taylor Hall as a player but they still wanted a breakdown.

 

[devilsblood]

I'm pretty certain they could build a robot that could make a coin land on heads every single time.

Physics.

But by hand we don't have that control. Which is not to say that physics are not in play, just that we are not aware of what the result will be. But the physics are 100% determinate of whether that thing lands heads or tails. Luck plays zero role.

 

[Feed Me A Stray Cat]

I'm pretty certain they could build a robot that could make a coin land on heads every single time.

Physics.

But by hand we don't have that control. Which is not to say that physics are not in play, just that we are not aware of what the result will be. But the physics are 100% determinate of whether that thing lands heads or tails. Luck plays zero role.

What are you talking about?

I think you might have missed the point of my post.

 

[Feed Me A Stray Cat]

Having better scouts?

What makes a better scout? I think most scouts in the NHL probably have fairly equal abilities. Some might be worse than others, but I don't think there's a Wayne Gretzky of the scouting world out there.

Besides, look at how NHL teams perform at the draft. There is almost no consistency to it. The Devils scouting was amazing in the 90s. Then it sucked in the 00s. And then got a bit better in the early to mid 10s.

Did Conte suddenly lose his magical scouting powers and then regain them?