PDO and Unsustainable Stats

[toquester]

It seems a lot of Stanley Cup Champions had something in common with the Calgary Flames. Even the Edmonton Oilers of the 80s had PDOs that suggested they were lucky.

This is from an excellent article written by a stats guy, about jumping to conclusions.

http://www.thestar.com/sports/hockey...app=noRedirect

"Assuming half of the last 23 Stanley Cup winners weren’t simply lucky, a PDO above 101 clearly is sustainable. While that may not sound like much, the difference between a PDO of 100 and 101 is about 25 goals over the course of a season. Or if you prefer, it’s the offence of last year’s Penguins vs. last year’s Islanders."

He concludes the article with the following:

"Many advanced stats critics complain that the statisticians are so in love with the math that it leads them astray. They have a point here: data provide a useful screen, but they’re only part of the picture. If you don’t understand what else is happening in the game, you’ll reach some very foolish conclusions."

[driveway]

An important quote from the article, from the perspective of the Flames.

If you go to sportingcharts.com, you’ll find the PDO generated by each team in each season since 1990. Of those 642 teams, only a dozen posted a regular season PDO of 103.0 or better, the best of which was 103.7.

Calgary currently has a PDO of exactly 103.

Assuming half of the last 23 Stanley Cup winners weren’t simply lucky, a PDO above 101 clearly is sustainable.

If Calgary's shooting percentage slipped by 2%, bringing us from 2nd in the league, to 10th, giving us that sustainable PDO of 101, we'd drop from 87 goals scored to 66 goals scored, which would take us from +18 to -3 in goals for.


Edit: Hang on ... NHL.com has two different totals for how many goals we've scored this year. The standings page gives us 87 scored, but the Goals For stats page says 84.

[getbak]

Originally Posted by driveway

Edit: Hang on ... NHL.com has two different totals for how many goals we've scored this year. The standings page gives us 87 scored, but the Goals For stats page says 84.

The standings gives a team one goal for every shootout win (and one goal against for every shootout loss), regardless of how many goals were scored in each shootout.

The stats page only counts goals scored during game play (regulation and OT).

So far, the Flames have won three shootouts. That's why the standings show them with three more goals than they've actually scored.

[EldrickOnIce]

I think your maths are off though, driveway.
Couldn't find Flames shots, but must be around 700, so can't be 20 less goals.
But your point stands. Unless the Flames are now an elite scoring team (and maybe they are), something has to give.

[Yrebmi]

PDO is based on parity.

It stems from the concept that all shots and all goalies are more or less equal and will more or less equal out over the course of a season. The discrepancies in the more and the less are labeled luck.

While there is a lot of parity in the NHL and that has led credence to this line of thinking, it is by no means as certain as so much of the NHL fan base has come to accept. The concept is completely fallible to elite goaltending/defense or high percentage shots. I wonder what the PDOs look like in leagues with less parity.

By the way. PDO was conceived by an Oilers fan, whose user name was PDO.
PDO is no good.

[albertGQ]

Ok. I'll be the first to ask and hopefully I'm not the only one that doesn't know.

What does PDO stand for and what does it mean?

[sureLoss]

Originally Posted by albertGQ

Ok. I'll be the first to ask and hopefully I'm not the only one that doesn't know.

What does PDO stand for and what does it mean?

http://blogs.thescore.com/nhl/2013/01/21/pdo-explained/

If you are too lazy to click the link: The short of it is PDO was the username of the person who first brought the stat to people's attention. It is a stat that tries to determine "luck" by simply adding save% and shooting %. Too high of a PDO and the implication is that the team has too much good luck and too low of a PDO means that a team has bad luck. The theory is that too high or too low PDO will eventually regress to a mean value (i.e. a lucky team's luck will run out and regress or an unlucky team will eventually play better when more pucks start going in and/or a goaltender starts putting up some lucky saves).

That's what some "advanced" stats people believe anyways.

[getbak]

Originally Posted by albertGQ

Ok. I'll be the first to ask and hopefully I'm not the only one that doesn't know.

What does PDO stand for and what does it mean?

PDO is a number calculated by adding a team's shooting percentage to its save percentage. PDO apparently was the username of the person who first suggested this number would have some value in measuring a team's "luck".

So, rather than giving it a logical name like Combined Shot Percentage, someone decided it made sense to call it PDO after that person's username. I believe this is one of the biggest reason people dismiss a lot of these "advanced" stats ideas, the names make no sense and don't lead to helping people understand what they mean.

Plus-Minus is a flawed stat, but at least it's easy to understand just based on its name, as are all the other "basic" stats. Imagine if Time-on-Ice was called the "LPC" for "Lester Patrick Clock" number because Lester Patrick was the first person in the NHL to track that, it would be ridiculous.

[EM11]

I understand how a PDO is generated, but the idea that it can "measure" something as subjective as luck seems ridiculous. What is luck? How can any statistic account for it? Clearly some teams are lucky and some are unlucky at any given time. Entire playoff series (and even championships) can turn on this. But no statistic could tell you how to quantify it.

If you listen to any Oilers' executive talk it becomes clear that they feel their luck is bad, that things haven't gone their way.

[Bend it like Bourgeois]

Originally Posted by getbak

PDO is a number calculated by adding a team's shooting percentage to its save percentage. PDO apparently was the username of the person who first suggested this number would have some value in measuring a team's "luck".

So, rather than giving it a logical name like Combined Shot Percentage, someone decided it made sense to call it PDO after that person's username. I believe this is one of the biggest reason people dismiss a lot of these "advanced" stats ideas, the names make no sense and don't lead to helping people understand what they mean.

Plus-Minus is a flawed stat, but at least it's easy to understand just based on its name, as are all the other "basic" stats. Imagine if Time-on-Ice was called the "LPC" for "Lester Patrick Clock" number because Lester Patrick was the first person in the NHL to track that, it would be ridiculous.

Don't worry, they'll rename it skill, or coaching, or some incredibly broad term to go with 'possession' and make it sound way more meaningful.

The article was really good.

[Lanny_McDonald]

What I continue to see is that these many measures were founded by people with no math knowledge or understanding of advanced statistics (no #### Sherlock). Many of these goofy numbers were put together by, not surprisingly, Oiler fans trying to explain why their team was better than it produced against the others in the NHL. That's using statistics to try and explain away an outcome rather than gain greater insight into the events themselves (poor method). Also, picks two loosely related statistics out of the air, slamming them together, and saying it is a measure of something else, completely unrelated and not tracked. I mean, luck is now just the difference between having a positive or negative result versus the mean? I thought luck was an event where a completely unexplainable event took place, like a puck bouncing off a stanchion and going into the net, or a defense man inadvertently shooting or directing the puck into his own net. Those are actual events that can be tracked and quantified. That is an actual statistic. Claiming luck is having a high comparative shooting percentage is, I don't know what, but it isn't a statistic, and it doesn't makes much sense unless you fall into the logic trap dreamed up in the original scenario.

[ricardodw]

Looking for a team that has an unsustainable RGI St.L is in deep trouble.

They have Steen, Lindstrom, Taresenko, Lehtera and Stastny that are soft players by the RGI (5). Last year they had Roy and Tarasenko. Steen did enough to get him over the 1.0 cutoff as a soft players.

The Blues are a a different team than John Davidson left behind.

[Lanny_McDonald]

Originally Posted by ricardodw

Looking for a team that has an unsustainable RGI St.L is in deep trouble.

Try as you might, the Ricardo Grit Index is not going to catch on.

[edslunch]

My gripe with PDO is the notion that everything must regress to the mean. Meaning if you played a long enough season every team should end up at 100%. And presumably every team then would be tied in the standings, every team would have an equal chance of winning the Cup.

That's patently ridiculous. Some goalies are elite, some aren't. Some players are more skilled scorers than others. Some teams have more talent, some teams have better systems, some teams have more injuries, etc., etc. Let's not assume the league is a bell curve.

Still, though, over time we can build up a pretty good idea of the normal range. If a team is putting up numbers that are unprecedented in NHL history then it's a fair bet they are over their heads. So while the idea that the Flames must drop back to 100 is BS, they are certainly outliers and may not be able to sustain it.

In short, I don't think the PDO crowd is necessarily wrong, but they undermine any credibility by a) being too absolute about it, without considering other factors and b) expressing themselves very poorly.

[Resolute 14]

Originally Posted by New Era

Try as you might, the Ricardo Grit Index is not going to catch on.

And it would be appreciated ricardodw you would stop pushing that crap in legitimate discussions. It adds absolutely nothing to the conversation.

[Street Pharmacist]

Originally Posted by edslunch

My gripe with PDO is the notion that everything must regress to the mean. Meaning if you played a long enough season every team should end up at 100%. And presumably every team then would be tied in the standings, every team would have an equal chance of winning the Cup.

That's patently ridiculous. Some goalies are elite, some aren't. Some players are more skilled scorers than others. Some teams have more talent, some teams have better systems, some teams have more injuries, etc., etc. Let's not assume the league is a bell curve.

Still, though, over time we can build up a pretty good idea of the normal range. If a team is putting up numbers that are unprecedented in NHL history then it's a fair bet they are over their heads. So while the idea that the Flames must drop back to 100 is BS, they are certainly outliers and may not be able to sustain it.

In short, I don't think the PDO crowd is necessarily wrong, but they undermine any credibility by a) being too absolute about it, without considering other factors and b) expressing themselves very poorly.

If PDO regressing to 1 is absolute, the only way to win is to shoot more than the opposition. It simply reinforces the fallacy that better corsi equals better team.

[Resolute 14]

Originally Posted by edslunch

My gripe with PDO is the notion that everything must regress to the mean. Meaning if you played a long enough season every team should end up at 100%. And presumably every team then would be tied in the standings, every team would have an equal chance of winning the Cup.

That's patently ridiculous. Some goalies are elite, some aren't. Some players are more skilled scorers than others. Some teams have more talent, some teams have better systems, some teams have more injuries, etc., etc. Let's not assume the league is a bell curve.

Still, though, over time we can build up a pretty good idea of the normal range. If a team is putting up numbers that are unprecedented in NHL history then it's a fair bet they are over their heads. So while the idea that the Flames must drop back to 100 is BS, they are certainly outliers and may not be able to sustain it.

In short, I don't think the PDO crown is necessarily wrong, but they undermine any credibility by a) being too absolute about it, without considering other factors and b) expressing themselves very poorly.

While advanced stat guys can be dicks a lot of the time, I find it likewise unnecessary for people who hate advanced stats to be unnecessarily literal in a bid to disingenuously dismiss a stat.

The argument that teams will inevitably regress to the mean is not a literal statement. What it says is that people who track the stat believe teams will regress toward the mean. Nowhere has it ever been suggested that every team would hit 100 exactly if the season was long enough. Also, no one has ever suggested there won't be outliers.

Likewise, people on both sides - depending on their belief or desire to dismiss the stat - take the "luck" comment far too literally. Take Josh Jorris. He has a shooting percentage of 27.4%. As much as some people hate the word, that simply is not sustainable. He's had good chances, and he's buried those chances. But at some point, his "luck" will change. A shot that is going in now hits the post instead. The goalie starts cheating to stop more of his shots. His teammates get fewer passes into prime areas through and he instead takes more shots from lower percentage areas. Numerous reasons why it will happen. Some the result of literal luck, some the result in a metaphorical sense.

Right now, we have a lot of rookies scoring a lot of points - the Flames lead the league in that regard by a ridiculous margin. There is no reason why people on the outside would not expect that the Flames will indeed regress toward the mean. And there is no reason why Flames fans should be upset by this. Otherwise, you all might as well get mad at yourselves and each other for predicting we'd finish near the bottom of the standings.

[edslunch]

Originally Posted by Resolute 14

The argument that teams will inevitably regress to the mean is not a literal statement. What it says is that people who track the stat believe teams will regress toward the mean. ......

Likewise, people on both sides - depending on their belief or desire to dismiss the stat - take the "luck" comment far too literally.

Hence my point B. If regress to the mean doesn't mean regress to the mean then use different words. Toward the mean gives a much better picture.

If luck doesn't literally mean luck then stop saying it and find a better way to express it.

Note these comments are not directed specifically at you, but almost every advanced stat discussion involves these exact terms.

[PantherExtreme]

From watching games and the team SH%,Corsi, etc. I would have to say they have been getting a bit of luck. The offense from the D, comeback wins, etc..

But the Flames 5 on 5 close SH% is apparently only 8% which is 10th and 5 on5 close SV% near the bottom of the league. (NaturalStatTrick.com)

[Jay Random]

Originally Posted by Resolute 14

The argument that teams will inevitably regress to the mean is not a literal statement. What it says is that people who track the stat believe teams will regress toward the mean. Nowhere has it ever been suggested that every team would hit 100 exactly if the season was long enough. Also, no one has ever suggested there won't be outliers..

PDO himself, from what I have read, devised the stat in the belief that it was a measure of luck. To be a measure of luck, it cannot be a measure of skill. Therefore, ALL deviations from 100 must be a matter of luck, and teams must regress TO the mean, not merely towards it, over a sufficient number of trials.

Some people do use the stat for other purposes, and I have no quarrel with them. But the interpretation ‘PDO measures luck’ (the original interpretation) has caught on in the media, and it’s a damned lie. It measures nothing of the kind.

Take Josh Jorris. He has a shooting percentage of 27.4%. As much as some people hate the word, that simply is not sustainable.

Fine. And you can see that by looking at nothing but the shooting percentage itself. PDO adds nothing to that discussion. (Does Jooris have a save percentage? Of course not.)

But you know who else has an unsustainable shooting percentage? Joe Colborne, at 0.0%. Oh, in theory he could deliberately shoot every puck square at the goalie and never score – but he’s not trying to do that, nor is anyone else. If you assume that good shooting percentages must come down, without paying any attention to the bad shooting percentages that are likely to come up, you are just cherrypicking – which is the first cardinal sin of statistics.