The Uncovered Truth About PDO and SH%

[CantHaveTkachev]

the 2017-18 TB Lightning were the last team in the NHL with the highest PDO to win 2 rounds in the playoffs (1.023)...they lost in the ECF
and that team was a consistently "good team"...not an "out of nowhere" team like the Canucks

if the Canucks PDO of 1.046 holds, it would be the highest since at least 2008 (couldn't find stats before that)

also, 7 of the top 21 players in shooting % are Canucks...no other team has more than 2 (DAL and WSH)...this never never been done before

 

[Breakers]

Performance Discrediting Optics

I think the last one reached over 50 plus pages

 

[andora]

I think the last one reached over 50 plus pages

Prevalent Dick-measuring Opinions

 

[TS Quint]

Is anyone else sick of advanced stats? I'm so over it.

The problem is advanced stats aren't used properly. You ask 99% of posters how the stat they are using actually explains what is happening on the ice and they don't have a clue but they found a stat that fits what ever arguement they want to win. They just shrug and say the result is the result and have zero context of what one of the 6 players on the ice is responsible for.

 

[rmthomson21]

I am not an advanced hockey theorist/statistician.

But isn't PDO based on actual hard data? Such as actual shot percentage and actual save percentage?

Please correct me if I am wrong.

But if I am correct, then PDO is much better than all of the "expected blah blah / blah blah" that gets mentioned around here.

To me, ACTUAL data is better than PROJECTED/EXPECTED data.

Things that have actually happened are better proof of something than something that some model predicts to supposed to happen.

I'm gonna try to make this very simple for you - you need to establish an expectation of what a player's performance will be in order to determine whether what actually happens during the season is within the range of normal expectations, or above or below it significantly.

..ok, we'll enjoy watching the game for entertainment, you can geek out over how many times McDavid blinks his eyes between scoring goals

..ok, we'll be taking your money in sports betting.

 

[a mangy Meowth]

Somehow I'm always left astonished at how much people just fundamentally misunderstand statistical analysis.

Happens on both sides of this. You'll have people who claim to "hate statistics" but actually still seem to understand the underlying reality of probability and causation/correlation, and then you'll have people who "love statistics" but bastardize and completely misuse it just to prove a point, while having no actual understanding of it.

 

[ricky0034]

I find it funny how "hey maybe this entire team scoring twice as many goals per shot this year as they usually do isn't sustainable" is being portrayed as this super advanced stat in this thread

 

[njdevils1982]

You should do stand up

like stand up and walk away from the keyboard?

 

[canuckster19]

Do you mean is anyone else sick of information?
You're free to watch the games and not care about any information you don't understand.

Yeah but advanced stats bros are like vegetarians, you're going to hear about their lifestyle whether you want to or not.

 

[mandiblesofdoom]

Do you mean is anyone else sick of information?
You're free to watch the games and not care about any information you don't understand.

"show me where on the doll the information hurt you"

 

[Korpse]

I don't believe it's about luck, it's about sustainability. Being far too high or far too low is very likely unsustainable over a long period of time. That's not to say there aren't outliers, but they are exactly that, an outlier and shouldn't be expected.

This. A PDO of 102 has only been exceeded 15 times in the last 10 seasons,15/308. 90% of teams fall in between 98-102 for a single season. When you take it further and combine all ten seasons all teams fall between the range of 98.99-101.08.

 

[Filthy Dangles]

Doncha know, good goaltending and capitalizing on chances efficiently is bad

You need to outchuck the other team and bring those %'s down a bit to earn some rispek

 

[Bedards Dad]

Yeah but advanced stats bros are like vegetarians, you're going to hear about their lifestyle whether you want to or not.

Oh for sure. Some believe advanced stats are more important than the games on the ice and that's just stupid. The reality is games are played and advanced stats help understand what happened in those games and what is likely to happen in the future based on the games played.
There is nothing wrong with ignoring advanced stats, but complaint about their existence is telling people you wear velcro shoes and your mom cuts up you chicken before you eat.

 

[zar]

I’m actually exhausted from the reliance and focus on advanced analytics. They can be an evaluation factor at times but some of these people who quote these numbers, not always objectively, don’t even understand what they mean and how flawed it is to look at specific statistics without broadening the scope of the context.

Anyone who devalues the Canucks season based solely on PDO need to grow a pair of eyes they can see out of and actually stay up late enough to watch a game. … and I hate the Canucks.

Oh well, carry on.

 

[Martin Skoula]

I find it funny how "hey maybe this entire team scoring twice as many goals per shot this year as they usually do isn't sustainable" is being portrayed as this super advanced stat in this thread

No, you see Tocchet after a decade of no-name mediocrity was actually spending all that time devising such a radically new and improved system that absolutely dwarfs all other hockey minds and doubles the offensive and defensive efficiency of the team beyond what any borderline all-star stacked dynasty teams have dreamed of achieving in the past.

Rick Tocchet is simply the truth and Petterson is his prophet, advanced stats nerds will have to cope with this fact.

 

[Mattilaus]

Let's just stop there.

Is there really "like [sp] to be some regression?" Maybe. Maybe not. I don't know. Do you know? Does anyone know? No, no one does. All we know is there's an expectation of what "the mean" is [which again, is probably unknowable in the context to which the regression comment is lobbed] and an implied belief that obviously, a stat has to tend toward that value over the long-term which can not happen for all kinds of reasons.

When might this regression occur? Tonight? Tomorrow night? 3 weeks from now? Right before the playoffs start? 2 years from now? No one knows. Well, eventually it will. OK, thanks Baba Vanga, that's really helpful for me trying to analyze what's going on with a team and understand what to expect. Going to be great when I go into a team's front office and things are going great and someone offers up hey I know you're all in a great mood, the team is doing fantastic in the standings, but this stat says it's unsustainable, you guys are going to regress to the mean. No, I have no idea what the means. No, I have no clue when that's going to happen. No, I have no clue what you can do to prevent it. What do you mean you're calling security? Look, I'm trying to help you folks out!

While the past is obviously not a perfect predictor of the future, it is a predictor and we're arguing over how much weight to put on it. Or, in my world, how much credibility to assign to it. When a stat outperforms expectations [either positively or negatively], you have to ask is this real, or is this a fluke? And the longer it outperforms, the more you have to ask that question - and simply being lazy and saying ah, well, it's going to regress to the mean is a worthless comment because it lacks both context and knowledge.

Thank you for being rude right off the hop. You have a misunderstanding of what PDO is and what people talk about when they talk about regression. Not going to get into it because it's clear you do not care to have a discussion on it with your rude tone right off the hop. It's clear this is too personal an issue for you.

Yeah but advanced stats bros are like vegetarians, you're going to hear about their lifestyle whether you want to or not.

You are in a thread specifically created to bash PDO and people who like stats. It's literally filled with people making fun of anyone who likes stats. Who exactly is the one telling everyone about their opinion again?

 

[Bank Shot]

Yeah but advanced stats bros are like vegetarians, you're going to hear about their lifestyle whether you want to or not.

What is so advanced about shooting percentage?

I mean its simple to look at someone's shooting percentage when its above 20% and say its not sustainable.

Last year when Kuzmenko shot 27% it was a no brainer to say that he wouldn't repeat it, and this season he gets traded away. Last year so many Oilers fans were pushing to keep Kostin. This year he has fallen completely off the map.

This scenario plays out every year with multiple players and its crazy that people still get surprised when someone that shoots 20%+ one year can't maintain it the next.

 

[SnowblindNYR]

They’re not disqualified from winning the cup, teams that rely on PDO are simply rarely good enough to actually win the cup, and they get exposed in the playoffs by teams that rely less on luck.

The issue with PDO is it’s not controllable, sustainable or predictive. So a team racking up wins on the basis of PDO is s ticking time bomb, and their success isnt predictive in a way that you could reasonably say ‘this team has won a lot of games, therefore it is likely they will continue to win games’.

Could the heater last all the way to cup? Sure. Can you count on it? No. And without the heater, PDO-reliant teams fall fast.

My question is this and this is what I don't get why is a team that's 8-2 with a high PDO that won their games on average by 3 goals any less likely to win than a team that's 8-2 with an average PDO that won their games on average by 1 goal? So if the former team won in closer games they're more sustainable. All the PDO tells you is that the margin of victory is unsustainable. There's enough PDO surplus to cut and still continue to win.

 

[SK13]

I’m actually exhausted from the reliance and focus on advanced analytics. They can be an evaluation factor at times but some of these people who quote these numbers, not always objectively, don’t even understand what they mean and how flawed it is to look at specific statistics without broadening the scope of the context.

Anyone who devalue the Canucks season based solely on PDO need to grow a pair of eyes they can see out of and actually stay up late enough to watch a game. … and I hate the Canucks.

Oh well, carry on.

Generally speaking, a person can have eyes and watch those games and not have anything of substance or value to say about anything and, in fact, be completely out to lunch about every aspect of what they're saying.

A misused stat can be countered with why it's misused. Random posters say stupid shit about other teams and players all day long and it can only really be countered with "nah uh", at least without duelling subjective statements or some interjection of objectivity.

Nobody should "discount" the Vancouver Canucks season by pointing out their SH% or PDO. They still won the games. They've played the hockey required to have the record they need to have. It's also not unfair to point out that teams with half of these flags always, always, always hit the brick wall. They didn't hack the matrix and fundamentally change the game of hockey forever, which is what it would take to find 13.5% repeatable in any way, shape or form.

 

[ijuka]

I mean, PDO is fine. It's effectively a goalie stat. Your team's goalie > opposing team's goalie means your PDO will be high, etc.

But they're historical trends. PDO has relatively little predictive power. A team with a high xGF% vs a team with a high PDO, you'd probably pick the team with a high xGF%.

 

[Mattilaus]

My question is this and this is what I don't get why is a team that's 8-2 with a high PDO that won their games on average by 3 goals any less likely to win than a team that's 8-2 with an average PDO that won their games on average by 1 goal? So if the former team won in closer games they're more sustainable. All the PDO tells you is that the margin of victory is unsustainable. There's enough PDO surplus to cut and still continue to win.

You are thinking about it wrong. Winning by multiple goals is good. If a team wins every game by 6 goals, that doesn't mean they are more likely to lose later on. We are trying to look at whether the current play is sustainable. If a team only takes 8 shots a game, lets their opponent put up 50 shots, but wins 6-2, did that team perform better than a team that put up 40 shots, only let 25 shots from their opponent, and won 4-2?

Which would you rather see from your team? Extremely low shot totals while giving up a lot of chances but getting by on high shooting percentage and high sv%? Or a team that dominates both offensively and defensively when it comes to shot totals and chances against? This is all people talking about PDO are saying. You can for sure win with high goal totals and high goal totals are not a bad thing.

You can have a team with a high PDO that is likely to regress a bit, but is also performing at a high level and is likely to keep winning. Both can be true.

 

[SnowblindNYR]

You are thinking about it wrong. Winning by multiple goals is good. If a team wins every game by 6 goals, that doesn't mean they are more likely to lose later on. We are trying to look at whether the current play is sustainable. If a team only takes 8 shots a game, lets their opponent put up 50 shots, but wins 6-2, did that team perform better than a team that put up 40 shots, only let 25 shots from their opponent, and won 4-2?

Which would you rather see from your team? Extremely low shot totals while giving up a lot of chances but getting by on high shooting percentage and high sv%? Or a team that dominates both offensively and defensively when it comes to shot totals and chances against? This is all people talking about PDO are saying. You can for sure win with high goal totals and high goal totals are not a bad thing.

You can have a team with a high PDO that is likely to regress a bit, but is also performing at a high level and is likely to keep winning. Both can be true.

What you're saying is fair but at least in the Canucks case they were winning with large margins of victory. If a team wins by a large margin of victory AND has a high PDO then it could probably normalize their PDO by winning close games. Sure ideally you win by a lot with a PDO around 100 but how common is that? Now if you win close games and have a high PDO that's a problem. But the Canucks were blowing teams out and had a high PDO. Their PDO regressed but their winning didn't. They probably just had fewer blowouts.

 

[andora]

You are thinking about it wrong. Winning by multiple goals is good. If a team wins every game by 6 goals, that doesn't mean they are more likely to lose later on. We are trying to look at whether the current play is sustainable. If a team only takes 8 shots a game, lets their opponent put up 50 shots, but wins 6-2, did that team perform better than a team that put up 40 shots, only let 25 shots from their opponent, and won 4-2?

Which would you rather see from your team? Extremely low shot totals while giving up a lot of chances but getting by on high shooting percentage and high sv%? Or a team that dominates both offensively and defensively when it comes to shot totals and chances against? This is all people talking about PDO are saying. You can for sure win with high goal totals and high goal totals are not a bad thing.

You can have a team with a high PDO that is likely to regress a bit, but is also performing at a high level and is likely to keep winning. Both can be true.

I think a little creedance needs to put into where are all those shots coming from.. are they all from the outside and low %

I think a bit of the catalyst here is that vancouver gets represented as a team like your example when they are actually middle of the pack in all these shots and chances metrics. So instead of a team getting lucky and gettimg great goaltending it can present as a team with a good goaltender and they are capitalizing on their chances

Score effects would be interesting to see as well

 

[Bank Shot]

I think a little creedance needs to put into where are all those shots coming from.. are they all from the outside and low %

I think a bit of the catalyst here is that vancouver gets represented as a team like your example when they are actually middle of the pack in all these shots and chances metrics. So instead of a team getting lucky and gettimg great goaltending it can present as a team with a good goaltender and they are capitalizing on their chances

Score effects would be interesting to see as well

NHL EDGE tracks all players and puck location with micro chips.

NHL EDGE Puck and Player Tracking Statistics - Home

Canucks are 9th in total number of high danger shots. They are tied for most games played so if you convert it into high danger shots per game they probably fall a few spots.

So I think its fair to say they are a team that is middle of the pack in terms of creating high quality chances.

 

[andora]

NHL EDGE tracks all players and puck location with micro chips.

NHL EDGE Puck and Player Tracking Statistics - Home

Canucks are 9th in total number of high danger shots. They are tied for most games played so if you convert it into high danger shots per game they probably fall a few spots.

So I think its fair to say they are a team that is middle of the pack in terms of creating high quality chances.

Where do they rank when leading / tied / trailing

And if i think of it later i am curious where they rank relative to these game states and periods as well