Sell me on Analytics... then sell me on ours

[Pookie]

I wonder if ppl in here know that Montreal also runs their team on analytics ...

I am fine with saying analytics matters.

Let’s focus on part 2. How has Dubas shown that he is any good at interpreting them?

 

[Aus34Matt]

I am fine with saying analytics matters.

Let’s focus on part 2. How has Dubas shown that he is any good at interpreting them?

From the results we've seen so far I'm gonna go with fake news.

 

[AllDay28]

I am fine with saying analytics matters.

Let’s focus on part 2. How has Dubas shown that he is any good at interpreting them?

did the leafs finish 1st in the North? How did they achieve that?

 

[Donnie740]

did the leafs finish 1st in the North? How did they achieve that?

Regular season hockey and playoff hockey are two completely different games, my friend.

You may not realize it, but the games are much more intense in the playoffs - - tighter checking, harder forechecking.

 

[Polaris1010]

So here's my challenge:

1. convince me that investing in analytics are important.
2. If so, convince me that the current leader of this analytics movement is good at his job. What other resources does he need to make analytics useful?

I do not believe analytics is applicable to hockey.

Even if analytics were applicable to a sport, building a championship team has no reliable "analytics" formula.

The human mind has watched many games, becoming part of human experience, and the road to a championship is always the same. It never really changes, every sport has their formula.

The game may change, but the formula stays the same.

For an example of the game may change but the formula stays the same, is to win a cup, the team needs a top flight defenceman, in the old days he was one prototype, nowadays the prototype he is fast and good at passing.

 

[Polaris1010]

Brady went to 10 Superbowls and won 7 of them.

What separates Brady from everyone else, is his mind.

In the combine, he was the worst in history to the combine. A record that is unlikely to be broken.

What Brady is, he is smart, competitive, tough, and cool under pressure.

Brady always wins the bad weather games, because he just brings all his intangibles for the bad weather game and will win it.

Does bad weather provide "analytics?"

No, bad weather should make the "analytics" null and void for those games, but Brady always wins the bad weather games.

What wins under high stress situations, is the intangibles that Brady has.

Obviously, the Leafs have none of that.

 

[Duke16]

I'm sure it's been mentioned in this thread and I don't feel like scoring through, but no analytics backed up the many decisions the Leafs have made over the past year (Simmonds, Bogosian, Foligno acquisition cost).

If anything, the Leafs probably overcorrected in a way that should have pleased the more traditional crowd and the result was the same... and everyone is pissed as they should be.

 

[ottomaddox]

Did Dubas use analytics to evaluate Keefe?

 

[Polaris1010]

Analytics in baseball is different than the other sports.

Baseball is about who is pitching and who is batting.

The analytics people said that the home run was decisive. So teams loaded up on sluggers.

This season, it all different. Teams have loaded up their pitching staffs, their starters are reliable junk ballers, who can throw curve balls and slides, or the like to say, spin the ball.

If the pitcher spins the ball, it is harder for the slugger to square it up to hit it out of the park.

The Yankees offense and Twins offense suck this season because the entire pitching philosophy league wide changed this season.


The fundamental problem to analytics for hockey, is that the correlations, or cause and effect, can be rather weak or even backwards.

Is possession needed for scoring? Or that teams that score tend to have higher possession numbers because they are more talented with the puck?

In baseball, when someone hits it out of the park, that is scoring. That is proof beyond any doubt because that is the rules.

Possession in hockey? Who knows what it really means.

Did the Leafs wins the possession battle against the Canadiens, and lose on the scoreboard?

(rant over)

 

[FraumBallard]

They are a tremendous tool to aid in valuations, but they are an aid, not the number 1 tool. Many times that directly conflict or even refute what is visually apparent.

The issue with us, is u think our GM is like many of our posters, where either they follow the analytics without watching games. Or, he does watch the games, but doesn’t understand the nuances of what he sees on the ice nearly as much as he understands graphs and charts. We have a lot posters with the very same issue.

Well said.
Agree 100%.

 

[Gallagbi]

Analytics in baseball is different than the other sports.

Baseball is about who is pitching and who is batting.

The analytics people said that the home run was decisive. So teams loaded up on sluggers.

This season, it all different. Teams have loaded up their pitching staffs, their starters are reliable junk ballers, who can throw curve balls and slides, or the like to say, spin the ball.

If the pitcher spins the ball, it is harder for the slugger to square it up to hit it out of the park.

The Yankees offense and Twins offense suck this season because the entire pitching philosophy league wide changed this season.


The fundamental problem to analytics for hockey, is that the correlations, or cause and effect, can be rather weak or even backwards.

Is possession needed for scoring? Or that teams that score tend to have higher possession numbers because they are more talented with the puck?

In baseball, when someone hits it out of the park, that is scoring. That is proof beyond any doubt because that is the rules.

Possession in hockey? Who knows what it really means.

Did the Leafs wins the possession battle against the Canadiens, and lose on the scoreboard?

(rant over)

Do people really consider hockey analytics to be simply possession at this point?

 

[Prominence]

It stopped being about possession about 4 years ago due to this thing called score effects where if a team (not the leafs obviously) is trailing, they will generate more shots because they work harder and if they don’t, the coach will bench them or the gm will trade them to set an example (hockey culture thing).

 

[Prominence]

Analytics in baseball is different than the other sports.

Baseball is about who is pitching and who is batting.

The analytics people said that the home run was decisive. So teams loaded up on sluggers.

This season, it all different. Teams have loaded up their pitching staffs, their starters are reliable junk ballers, who can throw curve balls and slides, or the like to say, spin the ball.

If the pitcher spins the ball, it is harder for the slugger to square it up to hit it out of the park.

The Yankees offense and Twins offense suck this season because the entire pitching philosophy league wide changed this season.


The fundamental problem to analytics for hockey, is that the correlations, or cause and effect, can be rather weak or even backwards.

Is possession needed for scoring? Or that teams that score tend to have higher possession numbers because they are more talented with the puck?

In baseball, when someone hits it out of the park, that is scoring. That is proof beyond any doubt because that is the rules.

Possession in hockey? Who knows what it really means.

Did the Leafs wins the possession battle against the Canadiens, and lose on the scoreboard?

(rant over)

I will preface this by saying that i don’t follow baseball, but i’m pretty sure you’re wrong. I thought baseball analytics is based around runs where the key metric is obp. I believe baseball culture is fixated on win shares and WAR. I don’t think it’s home runs because sluggers tend to get paid the most money and baseball analytics is about finding inefficiencies (cheap players who can generate runs). Also, if teams have loaded on pitchers, i believe it’s due to covid precipitating the league to shorten double header games to 7 innings, making each inning more valuable.

 

[Donnie740]

It stopped being about possession about 4 years ago due to this thing called score effects where if a team (not the leafs obviously) is trailing, they will generate more shots because they work harder and if they don’t, the coach will bench them or the gm will trade them to set an example (hockey culture thing).

So “puck possession” is passé now - - is it all about “high danger chances” now?

Who knew a shot from close range in the slot was more effective than a shot from outside the blue line.

What a breakthrough! This could revolutionize the game of hockey.

 

[Donnie740]

Analytics in baseball is different than the other sports.

Baseball is about who is pitching and who is batting.

The analytics people said that the home run was decisive. So teams loaded up on sluggers.

This season, it all different. Teams have loaded up their pitching staffs, their starters are reliable junk ballers, who can throw curve balls and slides, or the like to say, spin the ball.

If the pitcher spins the ball, it is harder for the slugger to square it up to hit it out of the park.

The Yankees offense and Twins offense suck this season because the entire pitching philosophy league wide changed this season.


The fundamental problem to analytics for hockey, is that the correlations, or cause and effect, can be rather weak or even backwards.

Is possession needed for scoring? Or that teams that score tend to have higher possession numbers because they are more talented with the puck?

In baseball, when someone hits it out of the park, that is scoring. That is proof beyond any doubt because that is the rules.

Possession in hockey? Who knows what it really means.

Did the Leafs wins the possession battle against the Canadiens, and lose on the scoreboard?

(rant over)

Excellent comments. You and I are on the same page about analytics being laughably overrated in hockey.

Analytics are far more applicable in baseball because there's way fewer variables and especially because it's a non contact sport.

In physical contact sports like hockey and football success depends largely on toughness and determination. As I explained to an analytics apologist yesterday, it's IMPOSSIBLE to quantify those qualities.

 

[Donnie740]

I will preface this by saying that i don’t follow baseball, but i’m pretty sure you’re wrong. I thought baseball analytics is based around runs where the key metric is obp. I believe baseball culture is fixated on win shares and WAR. I don’t think it’s home runs because sluggers tend to get paid the most money and baseball analytics is about finding inefficiencies (cheap players who can generate runs). Also, if teams have loaded on pitchers, i believe it’s due to covid precipitating the league to shorten double header games to 7 innings, making each inning more valuable.

So you admittedly don't follow baseball, but you want to try and tell someone their baseball related comments are wrong?

 

[Bomber0104]

When was the last time you heard a Corsi argument?

Whoosh

 

[17 Clark]

did the leafs finish 1st in the North? How did they achieve that?

That’s because none of the other teams tried The division is garbage and they only start a try when it counts do you know the exact time the Mapleleaf stop playing do you think they should be rewarded it’s not children playing they act like it it’s the National Hockey League you want to be paid a premium then play like it inexcusable not sweeping it under the curb this time

 

[57 Years No Cup]

Analytics are a good supplement to a management and coaching staff that is well versed in hockey culture and playoff hockey to gain the edge.

But I often wonder if we’re seeing a Leafs game plan built to generate stats totals like number of passes attempted and % completed, total time with puck on stick and clock time in the offensive zone which results in good metrics while actually not particularly dangerous.

Sometimes you just need a chip and chase down the boards and a hard board battle and recovery but we curl back for the extra 3 steam boats and complete 2 extra touches of the puck which looks good on paper until that puck is turned over with momentum going the other way.

BINGO! Stats for stats sake.

 

[Bomber0104]

Analytics are a good supplement to a management and coaching staff that is well versed in hockey culture and playoff hockey to gain the edge.

But I often wonder if we’re seeing a Leafs game plan built to generate stats totals like number of passes attempted and % completed, total time with puck on stick and clock time in the offensive zone which results in good metrics while actually not particularly dangerous.

Sometimes you just need a chip and chase down the boards and a hard board battle and recovery but we curl back for the extra 3 steam boats and complete 2 extra touches of the puck which looks good on paper until that puck is turned over with momentum going the other way.

Correct.

Years ago, Boston seemingly had no problem dumping the puck in the corner, handing possession back over to Jake Gardiner, and absolutely hammering him to the point he was visibly uncomfortable handling the puck for the rest of the series.

On paper that's "bad"

In real life Boston wins the series after a game 7 meltdown by Jake Gardiner.

 

[Pi]

Leafs went away from the analytics this year by getting guys like Simmonds, Thornton, Foligno etc.

You want to talk about teams that are all in on analytics? That’s the Bruins, Avs, Canes etc.

I would fire Dubas and hire Eric Tulsky if we could but I bet that the next hire we have is an over the hill vet like Rutherford so we’ll go from one extreme to another and effectively f*** our entire team up.

 

[sparxx87]

Leafs went away from the analytics this year by getting guys like Simmonds, Thornton, Foligno etc.

You want to talk about teams that are all in on analytics? That’s the Bruins, Avs, Canes etc.

I would fire Dubas and hire Eric Tulsky if we could but I bet that the next hire we have is an over the hill vet like Rutherford so we’ll go from one extreme to another and effectively f*** our entire team up.

You mean like Shanahan already did with Lamoriello to Dubas?

Shanahan needs to go, he’s responsible for this mess and everything that’s happened since he hired an inexperienced and unqualified individual to run the team.

New president can choose his GM.

 

[Gallagbi]

Whoosh

I completely agree.

And to build. The Trotz article talks specifically about Corsi and how it could negatively impact true possession and scoring chances. Oddly enough, the advanced stats we see today are trying to measure that, specifically scoring chances. We as fans get the basics, if I'm paying, I'm looking for the next steps.

I'd also expect the team talks with players about their metrics if they're attempting to use them and see value, so not sure if be worried about a player/agent knowing. I would hope they do. What would be the point of tracking if you're not trying to influence behaviours?

 

[Mickey Marner]

Should also acknowledge that the analytics teams care about and the analytics fans/bloggers care about aren't the same. Like, a team doesn't care about RAPM, WAR, GAR etc. because it doesn't convey any useful information to them. If I say player B has a WAR of 2.6 it doesn't tell you anything that would help with team building or evaluation. Fans care about these numbers because their interest is in trying to find out who the best players that they're not watching regularly are. If they were building a team they would have a different purview.

 

[WaW]

Not a Leafs fan, but these discussions are always interesting to me. I think analytics are a useful tool, but not a be-all end-all in hockey, like they can be in other sports. I think OPs title should be "sell me on how heavily we are relying on analytics"

Why analytics work by themselves in baseball is obvious. The 1 v 1 transaction of the pitcher vs the batter is a constant that always occurs in the exact same spot, with and with virtually similar barriers outside of these two players in any given play.

I would also argue they are far more useful in football and soccer than in hockey. These are sports played on massive fields of play with so much more exploitable space than what exists on the ice surface in hockey, and due to the number of players and the various different roles these players all have compared to roles on a hockey team, the individual transactions are far more diverse, and alot can happen both on and off the ball than what can happen in hockey.

Alot of the most popular hockey analytics are loosely based off of soccer analytics, and I think the realities of hockey limit their effectiveness (however that isn't to say they aren't useful). The ice surface is too small relative to the sheer number of bodies on it at any given time. You need to include refs in that, because as much as they try to stay out of the way, having 4 extra bodies out there at all times limits the possibilities of utilizing space even further. Additionally the individual transactions that can take place in a hockey game, most notably those which occur off the puck, are equally limited due to the size of the surface and the saturation of consumed space. 16 bodies on such a small surface, where the diversity of roles (basically C, W, D, G) is less than the other two sports I mentioned here, significantly limits the possibilities of exploiting the same, or similar data that is effectively exploited in these other sports.

Again though, I still see it's usefulness in certain situations and there are some advanced metrics that I do think are extremely helpful, especially in judging players who are tough to distinguish simply by looking at traditional stats (e.g. bottom 6 forwards, middle and bottom pairing dmen)