Leafs - Sens - Habs 2013-14

[hockeylover]

But nobody is claiming the Leafs are "just" the team that lost 8 straight... they're the team that lost 8 straight but they're also the team that won 6 straight at one point. They aren't a 0 point team. They aren't a 164 point team. They aren't a 95 team... they're a 85ish point team.

Really? There are plenty of people acting like the team in front of us right now is the "Real Leafs."

 

[Habspatrol]

Really? There are plenty of people acting like the team in front of us right now is the "Real Leafs."

Sorry... I guess I can't speak for everyone. I'm sure there are plenty of morons. That said, I can't see even the biggest moron out there thinking that the Leafs are an 0-82 team... or even a 30-52 team.

If I'm saying "this is the real Leafs" I'm saying they're a bubble team. Like 85-90ish points.

 

[hockeylover]

Sorry... I guess I can't speak for everyone. I'm sure there are plenty of morons. That said, I can't see even the biggest moron out there thinking that the Leafs are an 0-82 team... or even a 30-52 team.

If I'm saying "this is the real Leafs" I'm saying they're a bubble team. Like 85-90ish points.

I don't have an issue with that. Before the season started, I had them in a wild card spot.

There are enough people that will say the times we played well were just luck and this is the regression to what they should be aka they are all geniuses and saw it coming. And you know who those people are.

 

[BKerr]

That's all very good, but PDO is based upon the presumption that any variations in team SV% & SH% from the leagues dead average is based on luck. There are a number of teams that have been average or better over a long period of time...and on the other side of the ledger, the same teams have tended to dwell among the bottom group as well over similar time periods (New Jersey must be very, very unlucky). If it was pure luck, we would see very wild swings from year to year regularly. Anyone who calls SH% "luck" doesn't understand probability. If SH% was based on "luck", the odds of a team being above average in luck, for multiple years running, is unlikely. The odds of multiple teams doing it is astronomical.

So sure, shooting 12% isn't sustainable, but that's ultimately, a meaningless statement. What matters is your % relative to your competition, and multiple teams have shown (Toronto, Pittsburgh, Tampa, Washington, Chicago) that staying significantly above average, over a long period of time (3-4 seasons) is absolutely sustainable.

As for SV%, it's ridiculous on it's face to call it luck. The NYR have been in the top 5 in team SV% in 5 of the last 6 seasons, top 10 in all 6. Anyone who wants to call that luck, and predict a return to the mean, is more that welcome to look ridiculous.

No PDO isn't all luck..... Being a 101... or even a 102... or something slightly above 100 is sustainable, when we start talking about being above 104.0 as the leafs were earlier in the year (ie December), you have to see that some of that is luck.

Where luck factors into the equation is when your goaltending is at .930, and your shooting is at 11.8... at the same time, leading to a PDO of close to 110 or higher, you are probably experiencing a fair amount of luck and should expect some regression to the mean.

Its going to be rare for any team to be at 100 right on. As i said there is some elements of skill in both shooting and goaltending, but to get to the range of 4% or more off the league average, its just not sustainable.

Puck possession is far more sustainable than "taking quality shots" and "preventing quality shots but allowing ones from outside" that Carlyle has been trying to sell.

 

[BKerr]

Really? There are plenty of people acting like the team in front of us right now is the "Real Leafs."

No the "real leafs" are not the team that lost 8 in a row. No one is saying that.

Regression to the mean is that their overall results.... Their 82 games... is around 85-90 points.

No one is trying to say the Leafs are at the level of the Buffalo Sabres. Thats not what "regression to the mean" implies.

 

[hockeylover]

No the "real leafs" are not the team that lost 8 in a row. No one is saying that.

Regression to the mean is that their overall results.... Their 82 games... is around 85-90 points.

No one is trying to say the Leafs are at the level of the Buffalo Sabres. Thats not what "regression to the mean" implies.

No one is saying that?

Or just not you?

 

[BKerr]

I haven't seen anyone saying that the Leafs are as bad as the Sabres, or are the worst team in the NHL.... but if someone has.... I will stand corrected.

 

[MindzEye]

No PDO isn't all luck.....

You might want to have a talk with about 95% of the hockey statistics community. They wrongfully use the word "luck" to describe variance in team SH% & SV% because they don't understand what is really going on there. Hockey analytics are such ridiculously lazy math, that that crowd has fallen into the trap of explaining away anything that doesn't fit neatly into their very basic math formulas.

Being a 101... or even a 102... or something slightly above 100 is sustainable, when we start talking about being above 104.0 as the leafs were earlier in the year (ie December), you have to see that some of that is luck.

I don't want to get into a semantics argument, but "luck" doesn't exist. Probability exists. Certain interactions lead to higher probabilities of certain outcomes. PDO is a measurement of probability, and skill can absolutely effect those interactions.

Where luck factors into the equation is when your goaltending is at .930, and your shooting is at 11.8... at the same time, leading to a PDO of close to 110 or higher, you are probably experiencing a fair amount of luck and should expect some regression to the mean.

First of all...a PDO 10% higher than average? With a .930sv% and 11.8%? I call bullshit. That's a 1045 max.

Yes, but the expected regression should be to the mean. You're no more likely to be a 1100 than you are a 900, and if you are a 1100, regression to the mean, means performing at something much closer to 1000 from that point onward.

Its going to be rare for any team to be at 100 right on. As i said there is some elements of skill in both shooting and goaltending, but to get to the range of 4% or more off the league average, its just not sustainable.

This is based more on the composition of a team though and not on some mystical levelling effect of "luck". If you constructed a team that received elite .930+ goaltending, and had high talent level forwards shooting 10%+, you're going to be around a 1030 or higher every year. That's not luck at work, that's elite levels of skill at work. There will be variance in performance, even among elite athletes, but that's a group of players that will out perform average, every year.

Puck possession is far more sustainable than "taking quality shots" and "preventing quality shots but allowing ones from outside" that Carlyle has been trying to sell.

You're preaching to the choir, though blindly believing in possession as the cure all is as wrong as Randy's nonsense.

 

[MindzEye]

I haven't seen anyone saying that the Leafs are as bad as the Sabres, or are the worst team in the NHL.... but if someone has.... I will stand corrected.

Corsi & Fenwick both suggest that they are.

 

[Habspatrol]

I've thought Corsi and Fenwick were complete shit from day one.

 

[BKerr]

You might want to have a talk with about 95% of the hockey statistics community. They wrongfully use the word "luck" to describe variance in team SH% & SV% because they don't understand what is really going on there. Hockey analytics are such ridiculously lazy math, that that crowd has fallen into the trap of explaining away anything that doesn't fit neatly into their very basic math formulas.

Well don't put me in with that crowd.

Hockey analytics are useful tools.... tools that can be used to augment traditional eyeball scouting. They are not the be all and end all and I've never said they are. They are useful though.

You'll never find me in either the pro-analytics (SB Nation style)or the anti-analytics (Steve Simmons style) crowd. I don't believe things are so black and white. I think there are some very useful metrics out there. Metrics that when used, with context, can be very informative. But context is key.

I don't want to get into a semantics argument, but "luck" doesn't exist. Probability exists. Certain interactions lead to higher probabilities of certain outcomes. PDO is a measurement of probability, and skill can absolutely effect those interactions.

First of all...a PDO 10% higher than average? With a .930sv% and 11.8%? I call bullshit. That's a 1045 max.

Yes I had a typo... as you can see above I said, "when we start talking about being above 104.0 as the leafs were earlier in the year (ie December), you have to see that some of that is luck."

I know it should have been 104.8, not 110.

Yes, but the expected regression should be to the mean. You're no more likely to be a 1100 than you are a 900, and if you are a 1100, regression to the mean, means performing at something much closer to 1000 from that point onward.

Yes, and look at the Leafs record from December onward. A team that from that point forward was a 500 hockey club, propped up by an unrealistic number of shootout victories and eventually even that collapsed on them.

The first thing to go was the goaltending dropping from 940 levels (first with Reimer, then with Bernier, who while still playing well and above average for an NHL goalie, wasn't playing the mind numbingly insane he was in october november) .... then we saw shooting percentages completely fall of the map in recent weeks.

This is based more on the composition of a team though and not on some mystical levelling effect of "luck". If you constructed a team that received elite .930+ goaltending, and had high talent level forwards shooting 10%+, you're going to be around a 1030 or higher every year. That's not luck at work, that's elite levels of skill at work. There will be variance in performance, even among elite athletes, but that's a group of players that will out perform average, every year.

Theoretically could you assemble a team with enough high level talent forwards shooting 10%+? Yes you probably could. In today's NHL with a salary cap and the draft set up to favor weaker teams, and all the other mechanisms for parity? Good luck with that.

You're preaching to the choir, though blindly believing in possession as the cure all is as wrong as Randy's nonsense.

Cure all? No... But a more possession system would have got this roster into the playoffs IMO.

Winning the cup? No, they didn't have the talent for that.

Its more than just Randy... its roster construction too.

 

[BKerr]

Corsi & Fenwick both suggest that they are.

Well if you rely only on corsi and fenwick, and don't look at anything else..... and blindly follow stats without proper context, then you may fall into the opposite trap of those who simply looked at the standings at the end of February and believed the Leafs were a top 10 team in the NHL (ie Nazem Kadri's comments).

They were never a top 10 team, they always had flaws that would be exposed in time.

they are also not nearly as bad as their corsi and fenwick as their system does allow some of their shots to be low quality, and they have an elite scorer in Phil Kessel, and some good ones in Lupul, JVR, and Kadri, along with a good goalie in Bernier. The cupboard of NHL ready talent is not close to bare, which is what the Sabres have had most of the year (basically just ryan miller, and he's now gone).

 

[MindzEye]

I've thought Corsi and Fenwick were complete shit from day one.

I don't think that they're shit. They're valuable pieces of information which can tell you how much a team drives play, maintains possession, etc. Possession is important. If you possess the puck more often, you'll draw more penalties than you'll take (on average), you'll score more goals than you give up (again...on average). Where I think the analytics zealots get it wrong though is when they explain away PDO as luck (when the individual aspects of PDO, SH% & SV% are shown by numerous teams to be above average sustainably...which suggests that they're at least partially, and probably more than just partially, skill driven) and proclaim that possession is all that matters.

I don't think it's accidental that the teams most would agree have high levels of individual skill are high in PDO, and teams most would agree don't have a lot of skill, have very low PDO.

 

[MindzEye]

Well don't put me in with that crowd.

Hockey analytics are useful tools.... tools that can be used to augment traditional eyeball scouting. They are not the be all and end all and I've never said they are. They are useful though.

You'll never find me in either the pro-analytics (SB Nation style)or the anti-analytics (Steve Simmons style) crowd. I don't believe things are so black and white. I think there are some very useful metrics out there. Metrics that when used, with context, can be very informative. But context is key.

Fair enough, you have to understand that we're a bit sensitive around here due to the abuse the team has taken at the hands of the pro analytics crowd this season. 4 points out of 24 isn't "regression" to the mean.

Yes I had a typo... as you can see above I said, "when we start talking about being above 104.0 as the leafs were earlier in the year (ie December), you have to see that some of that is luck."

I know it should have been 104.8, not 110.

Fair enough. Though I still disagree with the assertion that it's top down "luck" that is unsustainable. If you take the more logical direction of looking at a team from it's constituent parts upward, you can pretty easily ascertain the sustainability of a team's performance. Even here, nobody expected Bernier/Reimer to keep up a .930 all season, and there were numerous posts about us needing more shots on net, even at the height of our SH%. Again though, we were a little weary from dealing with people who don't understand what regression to the mean, actually means.

Yes, and look at the Leafs record from December onward. A team that from that point forward was a 500 hockey club, propped up by an unrealistic number of shootout victories and eventually even that collapsed on them.

The first thing to go was the goaltending dropping from 940 levels (first with Reimer, then with Bernier, who while still playing well and above average for an NHL goalie, wasn't playing the mind numbingly insane he was in october november) .... then we saw shooting percentages completely fall of the map in recent weeks.

But again, the first was no more likely that the 2nd (especially considering the fact that what occurred the last month was far less likely statistically than what had occurred previously, over a large sample)

Theoretically could you assemble a team with enough high level talent forwards shooting 10%+? Yes you probably could. In today's NHL with a salary cap and the draft set up to favor weaker teams, and all the other mechanisms for parity? Good luck with that.

Pittsburgh's been 9.8% or higher in 6 of the last 8 seasons, with one of the two seasons they didn't do it, had Crosby & Malkin out for the majority of the season.

Cure all? No... But a more possession system would have got this roster into the playoffs IMO.

Winning the cup? No, they didn't have the talent for that.

You seem to have a reasonable view of the use of the numbers, you're in the minority then when it comes to those who use them.

Its more than just Randy... its roster construction too.

Was it roster construction in Anaheim too? They saw an across the board 5-7% increase in Corsi% when Boudreau replaced Randy, much like Toronto has seen a similar across the board drop in individual Corsi% since Randy took over.

Are there roster construction issues? To a degree, sure. But system is a massive issue here, according to both the numbers and the eye test.

 

[BKerr]

Fair enough, you have to understand that we're a bit sensitive around here due to the abuse the team has taken at the hands of the pro analytics crowd this season. 4 points out of 24 isn't "regression" to the mean.





Fair enough. Though I still disagree with the assertion that it's top down "luck" that is unsustainable. If you take the more logical direction of looking at a team from it's constituent parts upward, you can pretty easily ascertain the sustainability of a team's performance. Even here, nobody expected Bernier/Reimer to keep up a .930 all season, and there were numerous posts about us needing more shots on net, even at the height of our SH%. Again though, we were a little weary from dealing with people who don't understand what regression to the mean, actually means.




But again, the first was no more likely that the 2nd (especially considering the fact that what occurred the last month was far less likely statistically than what had occurred previously, over a large sample)






Pittsburgh's been 9.8% or higher in 6 of the last 8 seasons, with one of the two seasons they didn't do it, had Crosby & Malkin out for the majority of the season.






You seem to have a reasonable view of the use of the numbers, you're in the minority then when it comes to those who use them.



Was it roster construction in Anaheim too? They saw an across the board 5-7% increase in Corsi% when Boudreau replaced Randy, much like Toronto has seen a similar across the board drop in individual Corsi% since Randy took over.

Are there roster construction issues? To a degree, sure. But system is a massive issue here, according to both the numbers and the eye test.

Like I said a system change might give you that 5% in corsi. Is that a cup contender? No but its likely a playoff team with bernier playing decently (what wilson never had was a goalie).

But its still roster construction as I dont believe they would be a true cup contender either. Switching randy out helps... but is not the cure all. I think we both agree to that.

That said the team can be better with a better system.

 

[BKerr]

Its also not just regression that led to this collapse. There is more to it.

The offence sputtered. I believe kessel and jvr have run out of gas. You have a compressed schedule plus two guys who played in the olympics. Plus a coach who only rolls three lines. That has to lead to fatigue and bumps and bruises.

Then you have Phaneuf playing too many minutes and hes run out of gas as well.

Add.to it the fact that carlyles system of lack of possession fatigues as well... (chasing the puck is harder than possessing it) and youve got a very human reason for why the teams overall play is worse too.

 

[Pronger84]

Its also not just regression that led to this collapse. There is more to it.

The offence sputtered. I believe kessel and jvr have run out of gas. You have a compressed schedule plus two guys who played in the olympics. Plus a coach who only rolls three lines. That has to lead to fatigue and bumps and bruises.

Then you have Phaneuf playing too many minutes and hes run out of gas as well.

Add.to it the fact that carlyles system of lack of possession fatigues as well... (chasing the puck is harder than possessing it) and youve got a very human reason for why the teams overall play is worse too.

You make some good points about the offense, the Kessel line sputtered down the stretch and it very well could be because of the oylmpics. However, Randy actually cut down Phaneufs icetime down to 24 mins a game, so really Phaneuf cannot use the " I was tired excuse" to justify things, same as every one else not named Kessel/JVR.

Want my opinion? The collapse occoured just after the cali trip, I think the Leafs got cocky with taking out 2/3 of the best teams in the League then they come back to play a mediocre team in Washington who dominated them, and once they got bit in the ass it just snowballed from there... plus it didn't help that Bernier (one of their sole reasons for being in a playoff spot) went down with a groin injury.

There was a lot of things that cost the Leafs the season, a dried up offence played a role but wasn't the main factor IMO for their demise.

 

[Habspatrol]

I don't think that they're shit. They're valuable pieces of information which can tell you how much a team drives play, maintains possession, etc. Possession is important. If you possess the puck more often, you'll draw more penalties than you'll take (on average), you'll score more goals than you give up (again...on average). Where I think the analytics zealots get it wrong though is when they explain away PDO as luck (when the individual aspects of PDO, SH% & SV% are shown by numerous teams to be above average sustainably...which suggests that they're at least partially, and probably more than just partially, skill driven) and proclaim that possession is all that matters.

I don't think it's accidental that the teams most would agree have high levels of individual skill are high in PDO, and teams most would agree don't have a lot of skill, have very low PDO.

Sure, they tell you about possession... but I don't think they tell us anything more than regular old shots for and against.

 

[BKerr]

Remember its not just minutes... its also the fact that his system with keeping the other team to the outside and counterattack is more tiring than having the puck. It also leads to more blocked shots... which take a toll. Either via injury (missing games) or being hurt (playing thrugh it but less than 100%.

Something is not right. The Dion Ive seen in he last three weeks is not the same player I saw earlier in he year. My eyeball test tells me that much. Something happened to him... either he's fatigued.... playong while hurt.... or doing a massive choke job. I go with a combination of 1 and 2 and dont really buy into number 3

 

[MindzEye]

Sure, they tell you about possession... but I don't think they tell us anything more than regular old shots for and against.

There's a few neat ways to slice and dice it that are useful (Close, 5 on 5, etc) but I prefer overall % (special teams matter....) so I don't necessarily disagree. When used what I would consider to be in proper context, think of them as a bit of a refined shots for/against.