NHL Analytics, My View...

[CreaseAndAssist]

NHL Analytics, My View

 

Just some thoughts on analytics and its place in the game today.  Enjoy and discuss or just read and think about it!  

[yave1964]

@CreaseAndAssist

 

  I think it was Bill James back in the eighties who said that a lot of people use stats the way that a drunk uses a light post, for support not actual illumination. Our resident Bill James stat geek @JR Ewing would probably know for certain.

 

  My thoughts on analytics, they are fine as a tool but NOT an end all be all, period. They are a great way to look at the game within the game, they support my notion that Connor McDavid is not only not the best plyer in the game he isn't one of the top ten, just the games best scorer who is a raw pimply faced kid with enormous upside who still is nowhere near his peak  in almost every other aspect. If you use scoring as the end all be all of determining who the best is, fine go with McDavid, If you use analytics to look at his poor faceoff pct in key situations, his defense actually slipping this year, that is fine too. 

 

To me the three variables are:

The eye test, watching the game itself and deciding for yourself if someone is any good.

The old reliable basic stats.

New age stuff like Corsi and Fenwick.

 

  Personaly I like to look at it all. Everything. Scoring, defensive efficiency, percentage of time in each zone, usage, ect…. I throw it in a blender and see how it looks. I know a lot of stat geeks live and die by Corsi and the rest of the new wave stats, if you do that but don't look at the eye test and smell test as well you are just a numbers cruncher who is missing the gameitself. Like I mentioned above, a drunk uses a light post for support and not illumination. But everything counts, it is just a matter of how much import you wish to put on each category.

[BluPuk]

LOL. I use the standings - you know, wins, losses and that other one - which escapes me at the moment.... 

[CreaseAndAssist]

15 minutes ago, BluPuk said:

LOL. I use the standings - you know, wins, losses and that other one - which escapes me at the moment.... 

 

At the end of the day, those are the only stats that matter. 

 

13 hours ago, yave1964 said:

@CreaseAndAssist

 

  I think it was Bill James back in the eighties who said that a lot of people use stats the way that a drunk uses a light post, for support not actual illumination. Our resident Bill James stat geek @JR Ewing would probably know for certain.

 

  My thoughts on analytics, they are fine as a tool but NOT an end all be all, period. They are a great way to look at the game within the game, they support my notion that Connor McDavid is not only not the best plyer in the game he isn't one of the top ten, just the games best scorer who is a raw pimply faced kid with enormous upside who still is nowhere near his peak  in almost every other aspect. If you use scoring as the end all be all of determining who the best is, fine go with McDavid, If you use analytics to look at his poor faceoff pct in key situations, his defense actually slipping this year, that is fine too. 

 

To me the three variables are:

The eye test, watching the game itself and deciding for yourself if someone is any good.

The old reliable basic stats.

New age stuff like Corsi and Fenwick.

 

  Personaly I like to look at it all. Everything. Scoring, defensive efficiency, percentage of time in each zone, usage, ect…. I throw it in a blender and see how it looks. I know a lot of stat geeks live and die by Corsi and the rest of the new wave stats, if you do that but don't look at the eye test and smell test as well you are just a numbers cruncher who is missing the gameitself. Like I mentioned above, a drunk uses a light post for support and not illumination. But everything counts, it is just a matter of how much import you wish to put on each category.

 

I don't put a ton of stock into most possession metrics because they don't assess quality.  I think players can benefit from inflated stats based on opportunity and ice time.  Like Valiquette once described, a defenseman had 7 shots on goal.  Good game?  Well, when you consider the defenseman shot the puck almost exclusively from the point without the benefit from the screen.  Valiquette's take is that player should not be congratulated on putting up good possession numbers but rather scolded for more or less playing catch with the goalie and wasting a bunch of potential high quality scoring chances.    

[JR Ewing]

@yave1964 Yeah, that was James.

---

 

"When someone looks at me and earnestly says, “I know what I saw,” I am fond of replying, “No you don’t.” You have a distorted and constructed memory of a distorted and constructed perception, both of which are subservient to whatever narrative your brain is operating under."

https://theness.com/neurologicablog/index.php/sleep-and-false-memory/

 

-Relying purely on the eyeball test leaves us prone to all of the biases that we bring along, not to mention our famously faulty and self-serving memories. Humans are VERY prone to confirmation bias, and I think that it can be pretty easy to confuse perceived effort with effectiveness.

 

-Relying purely on the numbers may be more objective, but is prone to errors of its own kind. The numbers may not measure what we really think they do, and can also be loaded with their distortions and illusions. I remember having a painful argument with another Oilers fan, where he was telling me that Kris Russell is terrific defensively because he has a lot of blocked shots, when that very same stat is evidence that he's not very good defensively.

i) The other team has the puck a LOT when he's out there, and

ii) His gap control is poor, leaving too much room for scoring chances.

 

-My own approach is to combine the two and to ask as many pointed questions as I can in order to try come up with the best answer possible.

 

-ANYBODY who says they have all of the answers while using only approach is simply wrong; like trying guess a house's square footage by looking at a photo of the front of the building. They may have strong knowledge in a particular area, but choose to limit their overall ability to learn more about the game by refusing to accept anything from the other, and often refuse to see some of the drawback which exist in their preferred approach.

 

It's seem funny to need to say this, but specifically with regard to advanced stats: the fight is over. All 31 teams have their own analytics department, and have for years. Some have better tools, others not so much, but they're all using it in conjunction with traditional scouting. It's only the outside, from our couches, where people question its value.

 

[CreaseAndAssist]

8 hours ago, JR Ewing said:

@yave1964 Yeah, that was James.

---

 

"When someone looks at me and earnestly says, “I know what I saw,” I am fond of replying, “No you don’t.” You have a distorted and constructed memory of a distorted and constructed perception, both of which are subservient to whatever narrative your brain is operating under."

https://theness.com/neurologicablog/index.php/sleep-and-false-memory/

 

-Relying purely on the eyeball test leaves us prone to all of the biases that we bring along, not to mention our famously faulty and self-serving memories. Humans are VERY prone to confirmation bias, and I think that it can be pretty easy to confuse perceived effort with effectiveness.

 

-Relying purely on the numbers may be more objective, but is prone to errors of its own kind. The numbers may not measure what we really think they do, and can also be loaded with their distortions and illusions. I remember having a painful argument with another Oilers fan, where he was telling me that Kris Russell is terrific defensively because he has a lot of blocked shots, when that very same stat is evidence that he's not very good defensively.

i) The other team has the puck a LOT when he's out there, and

ii) His gap control is poor, leaving too much room for scoring chances.

 

-My own approach is to combine the two and to ask as many pointed questions as I can in order to try come up with the best answer possible.

 

-ANYBODY who says they have all of the answers while using only approach is simply wrong; like trying guess a house's square footage by looking at a photo of the front of the building. They may have strong knowledge in a particular area, but choose to limit their overall ability to learn more about the game by refusing to accept anything from the other, and often refuse to see some of the drawback which exist in their preferred approach.

 

It's seem funny to need to say this, but specifically with regard to advanced stats: the fight is over. All 31 teams have their own analytics department, and have for years. Some have better tools, others not so much, but they're all using it in conjunction with traditional scouting. It's only the outside, from our couches, where people question its value.

 

 

I know some people, who write like I do that only occasionally watch the games, and mostly follow the analytics.  I think that's crazy.  

[JR Ewing]

1 hour ago, CreaseAndAssist said:

 

I know some people, who write like I do that only occasionally watch the games, and mostly follow the analytics.  I think that's crazy.  

 

Yeah, I agree. I've never understood the need to limit ourselves.

 

[rottenrefs]

9 hours ago, CreaseAndAssist said:

 

I know some people, who write like I do that only occasionally watch the games, and mostly follow the analytics.  I think that's crazy.  

KFAN is the worst. They narrate about the game, bring in Nanny and never watch a game. Then wonder why more people aren't on board.

[yave1964]

@JR Ewing

 

I am going to have to respectfully disagree with you in terms of the eye/smell test, I think it has more relevance than you do.

 

  I am going to take some names at random, lets say Patrik Kane several times in a game blows past the defense and takes deadly aim on Bobrovsky who absolutely stones him, Kane at the end of the day is a minus one and held off the score sheet, did he have a bad game? To my way of thinking no he did not.

  Or lets say that defenseman Lindholm does a terrific job of boxing out Laine preventing him from a decent scoring chance and then on the same play his partner Manson coughs up the puck to Wheeler who fires it in, Lindholm at the end of the day IMHO still played a very solid game shutting down Laine but suffers as a minus player due to his partner, or if both defenseman play extremely well but Gibson is off that day and gives up a couple of softies, their stats suffer. 

 Or lets say that Ryan O'Reilly is on his usual game and wins faceoff after faceoff in the offensive zone but Taresenko keeps missing, O'Reilly's metrics will be off as far as offensive zone scoring chances go. Sometimes I can see with my eye what the stats cannot.

 

  Now that said, these are individual game scenarios that I am pointing out, sometimes individual shift scenarios that will skew game stats unfairly but I do agree that in almost all cases by seasons end 'puck luck' tends to even out statistically speaking. My only point being that I believe that if you sincerely follow hockey and watch the game, seriously WATCH it you can truly see things that the numbers do not especially within the context of an individual game. I agree that teams have finally started taking numbers seriously as they should and have their own numbers crunchers as they should have decades ago, but because of the human element they have scouts as well. IMHO it takes a mix of everything to properly 'see' a game.

[yave1964]

A case in point from an AHL game that I saw a few years ago.

 

 Teemu Pulkinnen was playing wing for the Iowa Wild against the Cleveland Monsters and he literally never went into the defensive zone, he hung around the top of the blue line every shift, no matter how deep the puck went in behind his goalie, he hung around the top of the blue line waiting for a breakout pass. I was blown away, I have never seen more indifferent defense from a player at any level in my entire life ever. I sat there with my wife, Pulkinnen had been an AHL star with my Griffins for several years and I had stated that his slapper would not translate to the NHL (which was true) but that statistically his defense was not as bad as the scouts were claiming because his metrics always looked good. This was the first game ever that I studied him, my mind was blown at his indifference as his teammates worked their tails off and he stood around not breaking a sweat hoping for a breakout pass that he could take down the wing and fire one of his patented slappers on net. The chance never came, IMHO mostly because the Wild whenever the puck entered the Monster zone were essentially playing a man down and were lucky to even clear the puck.

  At the end of the day Michalek in net had a shutout against the Monsters. Did that mean Pulkinnen had a good game? God no, at least from where I was sitting. Maybe by his hovering around the blueline a defender had to hang back a bit and not pinch in on the play but he could not have broken a sweat in the game. He was not awful, he was indifferent. I saw that with my own eyes and was angry at him (although rooting for the Monsters). Like I said, if you know how to watch a game, you can pick up on things that the metrics will not.

[JR Ewing]

4 hours ago, yave1964 said:

@JR Ewing

 

I am going to have to respectfully disagree with you in terms of the eye/smell test, I think it has more relevance than you do.

 

Actually, it sounds like we agree entirely. The eye test is completely relevant, especially if people try their best not to watch just as a fan/employee of their team. Good scouts help a team and bad ones hurt. Good analytics help a team and bad ones hurt.

 

[CreaseAndAssist]

On 8/22/2018 at 5:04 AM, rottenrefs said:

KFAN is the worst. They narrate about the game, bring in Nanny and never watch a game. Then wonder why more people aren't on board.

 

I feel pretty confident Brandon Mileski and Pat Michelletti watch most of the games.  Lou...probably not.  However, the guys I'm talking about are fellow bloggers.  Some who I know, watch maybe once in every 5-6 games, but mostly just follow it via the analytics.  

[rottenrefs]

8 minutes ago, CreaseAndAssist said:

 

I feel pretty confident Brandon Mileski and Pat Michelletti watch most of the games.  Lou...probably not.  However, the guys I'm talking about are fellow bloggers.  Some who I know, watch maybe once in every 5-6 games, but mostly just follow it via the analytics.  

I'm not exactly sure when they're on (air) but almost every time after a game whichever DJ's commenting on those games concur they didn't watch the game but give their, what I call 'Wild staff fed' biased input that feeds an inaccurate synopsis of just about every game.

[CreaseAndAssist]

On 8/25/2018 at 10:36 AM, rottenrefs said:

I'm not exactly sure when they're on (air) but almost every time after a game whichever DJ's commenting on those games concur they didn't watch the game but give their, what I call 'Wild staff fed' biased input that feeds an inaccurate synopsis of just about every game.

 

I don't really agree with that.  If Kevin Falness is running the post-game show, I'd agree...its a total Wild propaganda fest.  I will never listen if he's running Beyond the Pond or the post-game fanline.  But if Mileski is running it, I think he calls it as he sees it.  Pat, certainly homerish at times...but he calls things out too.  

 

Back to analytics, I heard an interview with Islanders GM Lou Lamiorello talking about the validity of the advanced stats.  He said there are major differences between arenas and that you have to be careful with getting too carried away with those numbers because they can be greatly distorted depending on where.  He more or less said that the stats sites don't really tabulate their own and rely on the 31 different arenas to do the job and as I just stated, they are significantly different across the league so it kind of takes a bite out of the accuracy of it all.  Lamiorello stated he has his own people tabulate their stats for accuracy reasons and they can differ significantly from the arena collected ones.  So for people who swear by Corsica etc who use the league's #'s, they may be wise not to take it as gospel.  

[rottenrefs]

13 hours ago, CreaseAndAssist said:

 

I don't really agree with that.  If Kevin Falness is running the post-game show, I'd agree...its a total Wild propaganda fest.  I will never listen if he's running Beyond the Pond or the post-game fanline.  But if Mileski is running it, I think he calls it as he sees it.  Pat, certainly homerish at times...but he calls things out too.  

You're fortunate then. Most the time I have KFAN dialed in from 5:15 am to 5:53, then from 9:30 am to 9:45, from noon to 12:30 and for about 15 minutes varying from 4:30 pm to 5 pm weekdays. Then on and off through the weekends.

Most the time they come across that it's almost like a badge of honor when they say they didn't watch the games.