Yes, in the end the result is still the same. The coaches match up their best defensive players with other teams top offensive players. The coaches use the QoC to set their lines, so how can it not be pertinent to the conversation. A second pair D-man who does not see the top end competition is NOT better than the top pair D-man who does, just because his analytics seem slightly better on paper. This whole argument makes no sense. The players who see the highest TOI and play against the top competition are your best and most indispensable players.

For the initial point that James made, can anyone say that if Gardiner was utilized in the exact same situations Rielly is, his numbers would be anywhere close to what they are? I personally doubt it.
- Beergu

This is 100% wrong. If Gardiner has better stats than Rielly over 60 games, he is a better player regardless of quality of competition. End of story. No debate. Final.

Indisputable fact at this point and opinions don't enter into it.

For a real life example, take Jake Gardiner and Morgan Rielly. Rielly gets top pairing minutes and has been the Leafs #1 defenseman this season. Jake Gardiner gets second pairing minutes. Yet, all stats based analysts will tell you that Gardiner is the superior player. Others argue, saying "sure, Gardiner has better stats, but he faces easier competition."
Very simple actually!
Rielly has a ZSO% of -6.96 to Gardiners +2.84
What this means is that Jake Gardiner while getting one less "O" zone start had almost 100 less "D" zone starts then Rielly
Oh! and he is 21 years old to Gardiners 25
And since you hate this so much...............I will also add that obviously Mike Babcock trusts the 21 year old Rielly more then he does the 25 year old Gardiner
And could you find any newer articles about this? One is two years old and the other four! I get that the theory is the same, But considering that analytics has come so far in such a short time, Four years is a long time to consider an article still relevant. Of course unless it is groundbreaking. Which this article is not!

This is 100% wrong. If Gardiner has better stats than Rielly over 60 games, he is a better player regardless of quality of competition. End of story. No debate. Final.

Indisputable fact at this point and opinions don't enter into it.
- James_Tanner

He doesn't though!
Rielly- 15 points
Gardiner- 13 points

For a real life example, take Jake Gardiner and Morgan Rielly. Rielly gets top pairing minutes and has been the Leafs #1 defenseman this season. Jake Gardiner gets second pairing minutes. Yet, all stats based analysts will tell you that Gardiner is the superior player. Others argue, saying "sure, Gardiner has better stats, but he faces easier competition."
Very simple actually!
Rielly has a ZSO% of -6.96 to Gardiners +2.84
What this means is that Jake Gardiner while getting one less "O" zone start had almost 100 less "D" zone starts then Rielly
Oh! and he is 21 years old to Gardiners 25
And since you hate this so much...............I will also add that obviously Mike Babcock trusts the 21 year old Rielly more then he does the 25 year old Gardiner
And could you find any newer articles about this? One is two years old and the other four! I get that the theory is the same, But considering that analytics has come so far in such a short time, Four years is a long time to consider an article still relevant. Of course unless it is groundbreaking. Which this article is not!
- camfor

We've been through this with zone starts before - conclusively and beyond a shadow of a doubt they don't make a long term difference any more than quality of competition does.

Or rather, they aren't causation - you earn your zone starts. They are a factor in why you're good, not how you're good.

He doesn't though!
Rielly- 15 points
Gardiner- 13 points
- camfor

And points are not the way to evaluate players.


Old Mr. Camfor, you're clearly just being a contrarian and you wouldn't agree if Tyler Dellow or Eric Tulsky showed up at your house in person and explained it to you like you were five.

There are so many things we could argue about that there isn't irrefutable proof about, but this isn't one of them.

We've been through this with zone starts before - conclusively and beyond a shadow of a doubt they don't make a long term difference any more than quality of competition does.

Or rather, they aren't causation - you earn your zone starts. They are a factor in why you're good, not how you're good.
- James_Tanner

Strongly disagree!

And points are not the way to evaluate players.


Old Mr. Camfor, you're clearly just being a contrarian and you wouldn't agree if Tyler Dellow or Eric Tulsky showed up at your house in person and explained it to you like you were five.

There are so many things we could argue about that there isn't irrefutable proof about, but this isn't one of them.
- James_Tanner

Agree!

This is 100% wrong. If Gardiner has better stats than Rielly over 60 games, he is a better player regardless of quality of competition. End of story. No debate. Final.

Indisputable fact at this point and opinions don't enter into it.
- James_Tanner

This is exactly what is wrong with the debate.

What the hell does this have to do with the coyotes?

We've been through this with zone starts before - conclusively and beyond a shadow of a doubt they don't make a long term difference any more than quality of competition does.
- James_Tanner

It seems like you've read the intro to advanced analytics page at Maple Leafs Hot Stove article a few years ago (https://mapleleafshotstov...n-to-advanced-statistics/) and took it as gospel, refusing to acknowledge more recent analysis.

To wit: If zone starts don't mean anything, then why do all the analytics sites calculate zone starts adjusted by removing events from the first 10 seconds after every faceoff? Because "it has been shown that the majority of the benefit or penalty of a zone start occurs during the first 10 seconds" (David Johnson @ puckalytics.com). That seems to directly contradict your past statement that zone starts have "virtually no effect". Johnson also claims that zone starts and QoC aren't as good as other stats (which I agree with, for example quality of linemate), but he's also shown the aforementioned 10s period after a faceoff does confer an advantage / disadvantage, as well as shown that heavily skewed offensive or defensive zone assignments can alter a player's Corsi by 1-2%.

Not to mention McCurdy's recent October article on Hockey-Graphs.com in which they took a more in-depth look at zone starts than Johnson and the rest had, in which he concluded "Although for most players the overall effect is going to be small, the examples above clearly show that for some players it’s quite significant." - for example, the Sedins both directly benefited by more than 3 goals each due to offensive zone starts in both the 2011 and 2012 seasons, Sharp, Toews & Tavares more than 2 goals each from their offensive zone starts, etc. http://hockey-graphs.com/...l-differential/#more-4268 Even Johnson, the biggest advocate of the idea that zone starts don't matter when averaged out over time, admits that in small sample sizes (eg, a game) zone starts may have profound effects.

Likewise there are calculations that show additional benefits / disadvantages to zone starts - for example, depending on what methodology was used, it's been found that each offensive zone start at 5v5 results in anywhere from 0.18-0.6 shots on goal, or ~0.3 Fenwick events. This would undoubtedly become even more pronounced in PP or SH situations. The fact that there's still some considerable variation in findings between different methodologies implies that there is some considerable work still to be done at excluding confounding factors and selection bias. Likewise QoC stats seem largely focussed on 5v5 play, where they inevitably largely even out over time due to home & away line change advantages by coaches. However, I haven't seen anything about QoC and PP / SH time, which I suspect might display the opposite.

Similarly we've got more recent articles which throw into question your "proven facts" - for example, http://hockey-graphs.com/...be-as-simple-as-we-think/ , which notes "Not long ago, Eric Tulsky looked at competition factors at NHL Numbers. The hockey analytics community saw two changes come from this research: 1) Ice time became a preferred measure over shot metrics; 2) Impacts of “line matching” moved from being important, to overrated, and then eventually inconsequential. Then research stopped. Not much more had been done despite huge gaps in our knowledge [...] Competition is still a real variable." My career is guided by the scientific method and I understand the enormous value of statistics, but I think hockey fans should be taking hockey analytics with a grain of salt, because the field is so nascent and there is still so much to be discovered - even the analytics experts are telling us so.

We've been through this with zone starts before - conclusively and beyond a shadow of a doubt they don't make a long term difference any more than quality of competition does.

Or rather, they aren't causation - you earn your zone starts. They are a factor in why you're good, not how you're good.
- James_Tanner

Yes, but you have it exactly backwards regarding defensemen. Top defensemen earn the toughest assignments by design and this is exactly why you are forever touting secondary defensemen like Jake Gardiner or offensive defensemen Erik Karlsson over the Duncan Keiths and Shea Webers of the league. It is exactly why you think Gardiner=Rielly and Muzzin>Doughty.

This is exactly what is wrong with the debate.
- Garnie

Not really. In our current society guys like Trump and Ted Cruz exist exactly bc no one is ever wrong. Its like we collectively decided that all things are subjective and its ok to just ignore anyone who doesn't tell you what you want to hear.

A consensus is not nessesary. You might choose to ignore proof and facts, but you're still wrong. This isn't a debate. I've made a statement, provided proof and all experts on the subject agree. This isn't about being fair or anything else. There is a right answer here .

It seems like you've read the intro to advanced analytics page at Maple Leafs Hot Stove article a few years ago (https://mapleleafshotstov...n-to-advanced-statistics/) and took it as gospel, refusing to acknowledge more recent analysis.

To wit: If zone starts don't mean anything, then why do all the analytics sites calculate zone starts adjusted by removing events from the first 10 seconds after every faceoff? Because "it has been shown that the majority of the benefit or penalty of a zone start occurs during the first 10 seconds" (David Johnson @ puckalytics.com). That seems to directly contradict your past statement that zone starts have "virtually no effect". Johnson also claims that zone starts and QoC aren't as good as other stats (which I agree with, for example quality of linemate), but he's also shown the aforementioned 10s period after a faceoff does confer an advantage / disadvantage, as well as shown that heavily skewed offensive or defensive zone assignments can alter a player's Corsi by 1-2%.

Not to mention McCurdy's recent October article on Hockey-Graphs.com in which they took a more in-depth look at zone starts than Johnson and the rest had, in which he concluded "Although for most players the overall effect is going to be small, the examples above clearly show that for some players it’s quite significant." - for example, the Sedins both directly benefited by more than 3 goals each due to offensive zone starts in both the 2011 and 2012 seasons, Sharp, Toews & Tavares more than 2 goals each from their offensive zone starts, etc. http://hockey-graphs.com/...l-differential/#more-4268 Even Johnson, the biggest advocate of the idea that zone starts don't matter when averaged out over time, admits that in small sample sizes (eg, a game) zone starts may have profound effects.

Likewise there are calculations that show additional benefits / disadvantages to zone starts - for example, depending on what methodology was used, it's been found that each offensive zone start at 5v5 results in anywhere from 0.18-0.6 shots on goal, or ~0.3 Fenwick events. This would undoubtedly become even more pronounced in PP or SH situations. The fact that there's still some considerable variation in findings between different methodologies implies that there is some considerable work still to be done at excluding confounding factors and selection bias. Likewise QoC stats seem largely focussed on 5v5 play, where they inevitably largely even out over time due to home & away line change advantages by coaches. However, I haven't seen anything about QoC and PP / SH time, which I suspect might display the opposite.

Similarly we've got more recent articles which throw into question your "proven facts" - for example, http://hockey-graphs.com/...be-as-simple-as-we-think/ , which notes "Not long ago, Eric Tulsky looked at competition factors at NHL Numbers. The hockey analytics community saw two changes come from this research: 1) Ice time became a preferred measure over shot metrics; 2) Impacts of “line matching” moved from being important, to overrated, and then eventually inconsequential. Then research stopped. Not much more had been done despite huge gaps in our knowledge
- TheMaritimer[...] Competition is still a real variable." My career is guided by the scientific method and I understand the enormous value of statistics, but I think hockey fans should be taking hockey analytics with a grain of salt, because the field is so nascent and there is still so much to be discovered - even the analytics experts are telling us so.

So you basically just took 20 minutes and wrote a book to tell me that im basically right. Jeez thanks for the input.

One of the big issues with quantifying hockey is that it requires a tremendous sample size to be relevant. Anything can happen on a given face-off. Who is on the ice against and when and where all could play a factor on a play, or it couldn't. It's just too god damn fast and random sometimes.

It's not anti-intellectualism, it's healthy skepticism about methodologies that contain quite a bit of circular logic.

In this case, these shot attempt differentials do not occur in a vacuum. Virtually every team in the league matches lines when warranted and even if they don't have last change, they are preemptively putting their best out for defensive zone draws and quickly changing if possible when they do not have the desired matchup.

So if Rielly is facing competition with only a slightly better shot differential than Gardiner on a shift-ly basis, does it not stand to reason that those same top line players Rielly is facing are also facing every other team's best just about every shift, while the players Gardiner is facing are not?

Also, the same argument against QoC (the fluidity of the game) can be used to challenge the efficacy of all shot-based statistics.
- Snowblind

Not really. In our current society guys like Trump and Ted Cruz exist exactly bc no one is ever wrong. Its like we collectively decided that all things are subjective and its ok to just ignore anyone who doesn't tell you what you want to hear.

A consensus is not nessesary. You might choose to ignore proof and facts, but you're still wrong. This isn't a debate. I've made a statement, provided proof and all experts on the subject agree. This isn't about being fair or anything else. There is a right answer here .
- James_Tanner

So if just using Corsi or whatever you want to call it.

Gardiner has better stats ( Corsi ) and by the way you can see this with the eye test too.

Does QOC matter over 1 shift?

Serious question, I won't ever comment QOC again regardless of the answer. ( I know I've said that before lol )

thanks for providing the research. i hope my drunken comment of last night found the right balance of insult/poking fun. if it came off as total poopbaggery.... my bad. seriously.

I still strongly disagree until i follow up the links provided - simply because the quote analyzing QoC provided refers only to Corsi... which Is an overrated stat. Useful but overrated. Kadri randomly flinging pucks at the net like a zoo animal flinging feces does not make him a higher QoC to face. But if you are determining QoC based on shot or shot attempts it is severely flawed. Babs told him to shoot like this and thats a Leaf strategy fir ateam with low sniper abilities.

As for the gardiner and rielly ex. I will say that Gardiner has been much better this year. and Ive written him in as a long term top four D now. But... The game against the Sens he looked like the Gardiner of old. Each giveaway is unique, and one giveaway that results in a high quality scoring chance is equal to... so many other giveaways. speaking of high quality scorung chances... thats an advanced stat that should be valued far more than shooting metric stats.

keep up the good work JT, always entertaining and often educational.

ipad entry...

I think it's pretty obvious that the home team has a slight advantage. But in poker and in baseball, last to act is a massive advantage. I don't know the all time record for home teams, but I'm guessing because the coach will just adjust on the fly that this isn't even close the advantage you think it is. I assume - but don't actually know - that in baseball the home team wins way more often than it does in hockey.

Do the research and prove me wrong by all means though.
- James_Tanner

poker, last to act is so absurdly important it shouldnt even be compared to sports. Ive made more than 80 percent of my adult income(age 21 to 33) playing poker. Solid analogy though... In baseball the advantage might come from selecting pitching staff or relievers. Its the only sport(f criminal football) that completely segregrates offense from defense.

All i got on this one is... solid analogy. and this story. vegas 2007. during wsop.

this one time i played poker against chris moneymaker and ben affleck.(not a tourney) buyin was too high for my bankroll( i bougyt 2gs but to play proper i shouldve had 10gs) so i just sat there for hours and. eventually doubled up off moneymaker. KK vs 57 suited. all in preflop. he just wanted to bust me cuz i didnt play any hands. crazy thing is.. his 57 suited has about the same, and maybe skightly better chance to beat my KK as the Leafs chances to win the auston.

This is 100% wrong. If Gardiner has better stats than Rielly over 60 games, he is a better player regardless of quality of competition. End of story. No debate. Final.

Indisputable fact at this point and opinions don't enter into it.
- James_Tanner

So, just to be clear James, it is your stance that if Babcock utilized Gardiner in the exact same way that he does Rielly (TOI, opponents, zone starts, etc.), that his advanced stats wouldn't be negatively impacted at all, and his Corsi, etc. would still be better than Rielly's (playing Gardiner's current role)? That seems like terribly flawed logic.

poker, last to act is so absurdly important it shouldnt even be compared to sports. Ive made more than 80 percent of my adult income(age 21 to 33) playing poker. Solid analogy though... In baseball the advantage might come from selecting pitching staff or relievers. Its the only sport(f criminal football) that completely segregrates offense from defense.

All i got on this one is... solid analogy. and this story. vegas 2007. during wsop.

this one time i played poker against chris moneymaker and ben affleck.(not a tourney) buyin was too high for my bankroll( i bougyt 2gs but to play proper i shouldve had 10gs) so i just sat there for hours and. eventually doubled up off moneymaker. KK vs 57 suited. all in preflop. he just wanted to bust me cuz i didnt play any hands. crazy thing is.. his 57 suited has about the same, and maybe skightly better chance to beat my KK as the Leafs chances to win the auston.
- Allan Bester

Last week I got QQ, called an all-in against JK and the flop came JK8 I don't know what the odds were, but I hit a queen on the river and it was pretty sweet.

Reindeer Games is one of my favorite movies.

I think it'd be pretty sweet to play in a game for 2K I only ever play for twenty bucks!

After looking into it, I was surprised to learn the baseball home field advantage isn't as great as I'd thought.

So, just to be clear James, it is your stance that if Babcock utilized Gardiner in the exact same way that he does Rielly (TOI, opponents, zone starts, etc.), that his advanced stats wouldn't be negatively impacted at all, and his Corsi, etc. would still be better than Rielly's (playing Gardiner's current role)? That seems like terribly flawed logic.
- Beergu

The logic is perfect. As I said, as the guys I quoted says, and as I've repeated multiple times here in the comments section: if you play against bad competition you will do better and if a coach can get a good match-up, it will be beneficial.

There is a chart in the article I quoted - actually it's in both of them - that shows that a player's possession stats rise dramatically and consistently when they play bad players.

The point we are making is that, basically, in the NHL it is impossible in real-world situations to get an advantage in matchups enough of a time to make a difference.

So if you look at their stats, you'll see that despite being labeled "the top pair" Rielly and Gardiner's opponents average the same amount of ice time per game and have the same possession stats.

This means they play against the exact same competition.

They have slightly different usage but - based on the research and seeming consensus of the advanced stat community - that usage does not have enough impact to explain away the difference in their stats.

Gardiner has two less 5v5 points and 4 points higher of a Corsi than Rielly. That is a massively better performance. If he was deployed the same as Rielly he would still be better.

And that is backed by research and isn't my opinion.

Last week I got QQ, called an all-in against JK and the flop came JK8 I don't know what the odds were, but I hit a queen on the river and it was pretty sweet.

Reindeer Games is one of my favorite movies.

I think it'd be pretty sweet to play in a game for 2K I only ever play for twenty bucks!

After looking into it, I was surprised to learn the baseball home field advantage isn't as great as I'd thought.
- James_Tanner

There is little to no advantage gained by batting last. Psychologically maybe but that is immeasurable. As pertaining to this conversation, the only way your baseball analogy would apply would be if somehow the home team would face easier pitchers and thus face easier QoC like a home team in hockey can deploy certain lines in certain situations to gain an advantage.

The logic is perfect. As I said, as the guys I quoted says, and as I've repeated multiple times here in the comments section: if you play against bad competition you will do better and if a coach can get a good match-up, it will be beneficial.

There is a chart in the article I quoted - actually it's in both of them - that shows that a player's possession stats rise dramatically and consistently when they play bad players.

The point we are making is that, basically, in the NHL it is impossible in real-world situations to get an advantage in matchups enough of a time to make a difference.

So if you look at their stats, you'll see that despite being labeled "the top pair" Rielly and Gardiner's opponents average the same amount of ice time per game and have the same possession stats.

This means they play against the exact same competition.

They have slightly different usage but - based on the research and seeming consensus of the advanced stat community - that usage does not have enough impact to explain away the difference in their stats.

Gardiner has two less 5v5 points and 4 points higher of a Corsi than Rielly. That is a massively better performance. If he was deployed the same as Rielly he would still be better.

And that is backed by research and isn't my opinion.
- James_Tanner

ICYMI...this link was provided earlier in this thread

http://hockeyanalysis.com...3/04/12/home-vs-road-qoc/

Point of fact: You don't get this kind of variance in QoC if it is "impossible" in real life hockey situations.

Granted there are variables like game situation (if home team is down in the 3rd matchups goes out the window) and game play variables (multiple PP's or PK's with a limited number of players filling those minutes and thus possibly affecting line combos after due to fatigue). I fully admit that these variables exist and based on averages could otherwise skew QoC stats over long periods of time, but still even with those variables I have provided proof that QoC variations do exist between home and road. Like I originally pointed out, it is clear as day that most if not all teams use the last change on home ice to their fullest advantage possible when the game situation allows. They aren't doing this b/c it is fun, they are doing it b/c it can give the home team an advantage and sometimes a significant advantage.

There is little to no advantage gained by batting last. Psychologically maybe but that is immeasurable. As pertaining to this conversation, the only way your baseball analogy would apply would be if somehow the home team would face easier pitchers and thus face easier QoC like a home team in hockey can deploy certain lines in certain situations to gain an advantage.
- uf1910

I would think there is a significant advantage to being the home team in baseball in a tie game after 8 innings.

ICYMI...this link was provided earlier in this thread

http://hockeyanalysis.com...3/04/12/home-vs-road-qoc/

Point of fact: You don't get this kind of variance in QoC if it is "impossible" in real life hockey situations.

Granted there are variables like game situation (if home team is down in the 3rd matchups goes out the window) and game play variables (multiple PP's or PK's with a limited number of players filling those minutes and thus possibly affecting line combos after due to fatigue). I fully admit that these variables exist and based on averages could otherwise skew QoC stats over long periods of time, but still even with those variables I have provided proof that QoC variations do exist between home and road. Like I originally pointed out, it is clear as day that most if not all teams use the last change on home ice to their fullest advantage possible when the game situation allows. They aren't doing this b/c it is fun, they are doing it b/c it can give the home team an advantage and sometimes a significant advantage.
- uf1910

Not a lot of analytics enthusiasts' favorites on either column of that list. I wonder why...