Around the league part 2

[King'sPawn]

What context do you need to add save percentage and shooting percentage? Good teams generally have high PDO, bad teams generally have low PDO. The only way to outdo PDO is to regularly outshoot your opponent.

Edit: You're right if you apply PDO to individual players. It has to be applied in the context of the team's overall quality. However, PDO can be applied to team stats easily with no additional context.

Well, yeah. Good teams get more saves and score more goals, but it doesn't say how. Is a goalie standing on his head with the team bleeding shots against, while being opportunistic with odd man rushes? That was the strategy of the Coyotes for many years. Aside from making the third round in 2012, would you ever consider them a consistently good team?

This also stems from my opinion that no single stat can quantify the quality of a team. And we went through this for years when Corsi alone was used to say how good a team was, and the 2012-14 Kings were Corsi darlings.

 

[Johnny Utah]

I expected PLD to play more like JT Miller.

Also, Ducks have been scratching Max Jones. If I was Blake he is a player I would be watching closer to the deadline.

 

[Maynard]

The Oilers will turn it around very soon IMO.

Agreed. They can turn it around and win 7 of their next 9 (or whatever), get back in the mix and be right there to piss us off in the post season.

 

[Raccoon Jesus]

Weren't they insane down the stretch last season, too? Never underestimate their ability to get hot if McDavid is healthy and on.

But there are too many teams to jump so unless one of Van, VGK, LAK has a really cold streak, Edmonton is already chasing a wild card spot.

 

[King'sPawn]

I expected PLD to play more like JT Miller.

Why? He's historically inconsistent with his engagement.

Weren't they insane down the stretch last season, too? Never underestimate their ability to get hot if McDavid is healthy and on.

Yeah, I didn't recall on my own, but others called this out.

 

[lumbergh]

Let's make it simple--do you believe that the Canucks are a team that will beat the NHL record for shooting % by 3 whole percentage points

AND

That they will set the single-season goaltending save percentage record as well?

No I don't, but I don't know that they won't. I wrote before that their PDO is absurdly high, but that's what it is, as of now.

I'll ask again. Can you define luck in hockey without PDO? What is puck luck? Pretty simple question.

 

[johnjm22]

The Oilers had 45 regulation wins last year. Only Boston had more.

Edmonton probably wins the cup last year had Skinner not fallen apart in the playoffs.

 

[Raccoon Jesus]

No I don't, but I don't know that they won't. I wrote before that their PDO is absurdly high, but that's what it is, as of now.

I'll ask again. Can you define luck in hockey without PDO? What is puck luck? Pretty simple question.

okay, and to what degree of certainty do you feel that way?

If you don't believe that they have the best shooters and goaltender in history, why not?

Especially since what you discussed here is LITERALLY regression:

"gain, PDO does not equal luck. Regression towards 1 is true if the events that drive shooting percentage and save percentage are random. In reality good teams generally have high PDOs above 1 and bad teams generaly have low PDOs below 1. You can check this season after season. No regression." And then you say you don't believe the Canucks will stay that way because their absurdly high numbers are an outlier. So you're really resisting "1", but not "within a certain amount of 1 is reasonable."

That's why we call that element of PDO luck. If you want to call it 'outperforming historical records by several standard deviations' you can. I'll simplify it and just say it's luck, particularly as it's consistently happening against the run of play. But otherwise, sure, it's an intangible.

 

[lumbergh]

Well, yeah. Good teams get more saves and score more goals, but it doesn't say how. Is a goalie standing on his head with the team bleeding shots against, while being opportunistic with odd man rushes? That was the strategy of the Coyotes for many years. Aside from making the third round in 2012, would you ever consider them a consistently good team?

This also stems from my opinion that no single stat can quantify the quality of a team. And we went through this for years when Corsi alone was used to say how good a team was, and the 2012-14 Kings were Corsi darlings.

Now we're getting somewhere.

2012-14 Kings were Corsi darlings. I recall reading somewhere that Sutter thought that he could win games if his team took more than 51% of the shots or something like that. It worked for a few seasons, and then the league changed. He tried to make up for a consistently low PDO by trying to outshoot the other team, but that only goes so far. It turns out that PDO is a better predictor of a team's winning percentage than Corsi For %. Sutter's teams started losing by 2016 despite outshooting the opponents because they couldn't make up for the low skill level on the team.

You can look at PDO over the past 8 years and it is the single best correlation to points percentage, better than Corsi For %, Fenwick For %, Shots For %. Only as of the past two years has xGF% caught up to PDO. That tells you how powerful a stat it is. Stat-heads have been working for a decade to model the game as well as a simple PDO stat does and only in the last two years have they reached parity.

 

[King'sPawn]

Now we're getting somewhere.

2012-14 Kings were Corsi darlings. I recall reading somewhere that Sutter thought that he could win games if his team took more than 51% of the shots or something like that. It worked for a few seasons, and then the league changed. He tried to make up for a consistently low PDO by trying to outshoot the other team, but that only goes so far. It turns out that PDO is a better predictor of a team's winning percentage than Corsi For %. Sutter's teams started losing by 2016 despite outshooting the opponents because they couldn't make up for the low skill level on the team.

You can look at PDO over the past 8 years and it is the single best correlation to points percentage, better than Corsi For %, Fenwick For %, Shots For %. Only as of the past two years has xGF% caught up to PDO. That tells you how powerful a stat it is. Stat-heads have been working for a decade to model the game as well as a simple PDO stat does and only in the last two years have they reached parity.

What's the explanation for New Jersey, the third-best team in the Eastern conference, being ranked 27th in PDO? And Montreal, who's out of a playoff spot, from having the fourth best PDO?

At least according to this: NHL Team PDO Leaders 2023-24? | Team Rankings

My main argument isn't the validity of PDO. It has applications and uses. And it's beautiful in its simplicity and easy accessibility while also removing subjectivity.

But it's still flawed BECAUSE the objective numbers don't tell the whole story. And I just think using it as a predictor alone can lead to faulty conclusions.

 

[Johnny Utah]

I thought Skinner is and was a solid goalie especially against us in the playoffs. The Oilers should have just signed a veteran backup at 1-2 million to back up him up. Never understood the Campbell contract.

They could have used that money for a player in the bottom 6 or another D.

 

[lumbergh]

What's the explanation for New Jersey, the third-best team in the Eastern conference, being ranked 27th in PDO? And Montreal, who's out of a playoff spot, from having the fourth best PDO?

At least according to this: NHL Team PDO Leaders 2023-24? | Team Rankings

My main argument isn't the validity of PDO. It has applications and uses. And it's beautiful in its simplicity and easy accessibility while also removing subjectivity.

But it's still flawed BECAUSE the objective numbers don't tell the whole story. And I just think using it as a predictor alone can lead to faulty conclusions.

I have New Jersey just above water at 1.003, all strengths according to Natural Stat Trick:

The numbers you are quoting use even strength PDO, which is also fine. Point still stands; teams doing well on PDO are winning. The exceptions just tell you that New Jersey is winning on special teams. Montréal, on the other hand, is losing points due to special teams.

 

[Raccoon Jesus]

Now we're getting somewhere.

2012-14 Kings were Corsi darlings. I recall reading somewhere that Sutter thought that he could win games if his team took more than 51% of the shots or something like that. It worked for a few seasons, and then the league changed. He tried to make up for a consistently low PDO by trying to outshoot the other team, but that only goes so far. It turns out that PDO is a better predictor of a team's winning percentage than Corsi For %. Sutter's teams started losing by 2016 despite outshooting the opponents because they couldn't make up for the low skill level on the team.

You can look at PDO over the past 8 years and it is the single best correlation to points percentage, better than Corsi For %, Fenwick For %, Shots For %. Only as of the past two years has xGF% caught up to PDO. That tells you how powerful a stat it is. Stat-heads have been working for a decade to model the game as well as a simple PDO stat does and only in the last two years have they reached parity.

Wait, I thought you were suggesting it WASNT predictive, only that it's about where you've been?

you can't assign predictive value AFTER the season because those are the results. That's like saying 'teams that ended the season with high +/- had winning records'.

Since we're going with what PDOs are now--here are the predicted final standings. Let me know how strongly you agree

West
-----
1. Vancouver
2. Vegas
3. Arizona
4. Anaheim
5. LAK
6. Dallas
7. Minnesota
8. St Louis

East
-----
1. Montreal
2. Boston
3. Detroit
4. New York Islanders
5. Ottawa
6. Pittsburgh
7. Tampa Bay
8. NYR


This also suggests the avalanche are actually a 29th place team, NJD is actually 27th.

This is a severe misuse of the stat.

Edit: all strengths PDO would also find Carolina in the dumps and the Avs still in the bottom tier.


In short, PDO has precious little predictive value, even YOU agreed that outliers will regress, and this stubborn adherence to the semantics surrounding the term 'luck' as it relates to outliers is clouding your judgment.

I appreciate the discussion and I don't intend this to sound as aggressively as it may appear but we're just dealing in 'facts' so I'm being blunt.

 

[DAkings20]

 

[lexlavender]

PDO does not regress to 1 over time. I just wrote earlier that the Boston Bruins had a PDO of 1.04 for 2022-23. It's because they're good. PDO just tells you they were really good.

Again, PDO does not equal luck. Regression towards 1 is true if the events that drive shooting percentage and save percentage are random. In reality good teams generally have high PDOs above 1 and bad teams generaly have low PDOs below 1. You can check this season after season. No regression.

I challenge you or any of the advanced stats cogniscenti to define luck without using PDO. What is puck luck? I'll wait.

Over time is not a single season, over time is in deep time over infinite iterations. It is a statistical average

 

[Ghetty Green]

Not very often you hear a team's captain criticize the fans.

Hes laying the groundwork for his inevitable departure

 

[Raccoon Jesus]

Over time is not a single season, over time is in deep time over infinite iterations. It is a statistical average

Can't even count anymore how many times a team has gone on an absolute PDO bender and their fans think they've finally figured it out only for everything to come crashing down the next season.

 

[Herby]

Not very often you hear a team's captain criticize the fans.

I like him a lot as a player, but give me a break, suck it up and play better. This is the second straight year the whole has underperformed the sum of the parts.

 

[Chazz Reinhold]

Can't even count anymore how many times a team has gone on an absolute PDO bender and their fans think they've finally figured it out only for everything to come crashing down the next season.

This is always good to pull out when this issue pops up: Percentage Luck in Hockey, Explained

The top five teams experienced what stats people sometimes call “PDO benders,” where they scored and won games at a rate wildly out of whack with their shot share (SF%). For example, the 2015-16 Panthers shocked the hockey world by winning the Atlantic division. They scored 56% of the goals at 5v5 and won 47 games despite being about even in terms of shots, in large part because of an inflated 102.1 PDO. The next season, despite only playing slightly worse, their percentages regressed and then some, leaving them below .500 and costing Gerard Gallant his job. PDO told us that they were not as good a team as they seemed in 2015-16, and not as bad as they seemed in 2016-17. The bottom four teams prove that the same can be true in reverse; the 2014-15 Coyotes had abysmal percentage luck and rode it all the way to a third overall pick. The next season they were actually worse in terms of shot share, but because their PDO regressed upwards they finished just below .500.

The traditional line with PDO is that it will “regress to the mean”, i.e. 100. But as we saw above, there is a skill component involved, even if it is slight. Based on the trendlines above, I calculated using a simple linear regression what an “expected” year two percentage performance might look like for each team-level save percentage, shooting percentage, and PDO:

We don’t only use PDO to predict regressions from one season to another. Teams go on small PDO benders during the regular season, and those often lead people to get unrealistic ideas about how good their favourite team is - or even lose coaches their jobs. Here’s the same calculation from above, but instead of comparing Year 1 and Year 2 it compares a team’s December 31st PDO to where it’s settled by the end of the season. It still regresses to the mean, but a bit more softly (since there’s ~41 rather than 82 games to do it).

 

[johnjm22]

Senators have a coaching problem IMO.

They should be able to get more out of their roster than they do.

After the Kings/Sens game, Chychrun made some comments about how well structed the Kings are. It felt like he was implying his team (the Senators) isn't very well structured in comparison.

 

[Choralone]

I like him a lot as a player, but give me a break, suck it up and play better. This is the second straight year the whole has underperformed the sum of the parts.

It's easy to give an answer that doesn't make you look like a schmuck. Most players do; it goes something like: "The fans weren't happy with our performance and neither were we. They're frustrated and we're frustrated." Etc.

 

[ScoreZeGoals]

This is McDavid's 8th year in the league, he's as old as Kopi was when the Kings won their second cup. How much longer until he wants out?

 

[Bandit]

This is McDavid's 8th year in the league, he's as old as Kopi was when the Kings won their second cup. How much longer until he wants out?

Who can possibly afford to take him?

 

[Jericho111091]

Shouldn't we have some sympathy for him? After all, it's probably not easy going through life with a face that looks like a rat.

He's like if a rat had horse teeth

 

[Steve Zissou]

He's like if a rat had horse teeth

Chatting about Couture, are we?