What Lightning fans should know about PDO
What in the world is PDO? And why should Lightning fans care? Well, given the disappointments fans have suffered this season, this statistic actually might be a sign that things could possibly get better. So, I'm all for it.
Numbers are taking over hockey, if you haven't noticed. I've recently done a lot of posts using numbers and that's quite remarkable. I hate math. I hate trying to add up grades at the end of a semester. The most advanced function my calculator performs is the square root. Which I've never used in any practical application.
But these adventures in statistics are a good way to check your perceptions against reality. Because observation is flawed. Our brains trick us into thinking we're seeing things we're not, and as human beings, we're very, very good at discovering patterns--even when those patterns aren't there. Over generalization is a great example of that phenomenon. A player turns the puck over and the fans start to say, "That guy is always turning the puck over." The next time it happens, you notice it, because your brain is looking for a pattern. Statistical analysis aims to remove some of that and allow for a less biased understanding of what's going on out there.
But the statz nerdz have long insisted that they don't understand everything, that they can't predict outcomes in individual instances or in the short-term, and that there's a great deal of apparent randomness involved in hockey. The rest of us may be thinking, "yeah, I've noticed." The numbers guys and gals are thinking, "I wonder how we can account for that."
And that's what PDO is intended to try to do--account for "luck." Here's the premise: Whether a specific shot on net results in a goal is completely unpredictable. It's almost all luck, or to be more accurate, randomness. There are two measures of whether shots on goal result in goals, shooting percentage and save percentage. For any game, one team's shooting percentage plus the other team's save percentage always adds up to 1 (or 100%.) But sometimes a particular shot gets deflected or the puck is wobbling or the goalie gets a stick on it or something else keeps it out or sends it in. That sort of thing is, statistically speaking, random--it's unpredictable and unrepeatable.
Long story short, for reasons my puny brain doesn't quite comprehend, over the course of a season any team's shooting percentage plus that team's save percentage regress quite strongly toward (that is, tend to move closer and closer to) 100%. (PDO is sometimes expressed as percentage, sometimes as a decimal and sometimes as something around 1000. They all seem to mean the same thing, though.) That regression toward 100% makes this a nifty little measure of whether current performance is sustainable or the result of randomness. You know. Luck.
A team with a low PDO will have a pretty strong tendency toward getting better. A team with a high PDO will have a pretty strong tendency toward falling off. The farther away from 100% your team is, the more likely that change in luck will occur.
The Lightning, as of Sunday afternoon, have a 5-on-5 save percentage of .902 (90.2%) and a team 5-on-5 shooting percentage of .084 (8.4%). Add those together and you get .986. So how low is that, really? Well, it's the 6th lowest in the league right now. According to my figures, the league PDOs range from .971 to 1.037 (the Colorado Avalanche with the lowest and the New York Rangers with the highest.) So, while PDO doesn't tell us anything about what particular areas are most likely hurting team performance or propose any solutions, it does tell us that randomness--those ubiquitous "bad bounces" everyone talks about--have had something to do with the outcome of the team's games. It also reminds us that those bad bounces are unlikely to be sustained over the course of the season.
So, while there's a lot I don't quite understand about PDO, what I do know tells me that there's hope that things will get better. If only I could say when it would happen and how much better it might get....
Thanks to Dirk Hoag at at On the Forecheck for helping me find the data I needed to be able to do this article. You can get it at www.behindthenet.ca/2011/team_data3.php . For more info on PDO, I'd start with the article "PDO: If you were going to understand just one NHL statistic," at Arctic Ice Hockey.Or try "PDO And What It Means" at PensBurgh.com. Also good are "Advanced Team Stats - What the Heck is PDO?" from Pension Plan Puppets and "PDO Numbers" from Tyler Dellow. A lot of the information out there on PDO relates to individual players rather than entire teams, but a Google search for "PDO hockey" brings up a lot of articles from good sources like Arctic Ice Hockey (which used to be the original Behind the Net) and mc79hockey. Look around for it if you like, and maybe you'll see stuff I've missed.
16 comments
|
0 recs |
Do you like this story?
Comments
Really nice lead-in
and good stuff.
I just noticed yesterday that we’re 13th in the East in goals scored and 13th in the East in goals allowed and 13th in the East in scoring margin, and I was just about to throw up my hands and say “well maybe we’re just the 13th best team in the league and it’s not going to get any better.”
Now I have hope. Please that it not be a fool’s hope.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Dec 12, 2011 9:32 AM EST reply actions
Why, thank you.
And it’s certainly nice to have hope. There’s no knowing, of course, whether this correction will come in time to get a playoff spot, or how strong a correction it’s likely to be, but I do take heart from the thought that there’s a fairly strong likelihood of me being able to see actual “Lightning hockey” this season.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
Agreed. Good read.
The fine people over at Winging it in Motown (Red Wings SB site) are very, very stat concious. You can find many articles (not just RW related) on the subject. Some of the writers have even developed some type of program that tallies stats, that we normally wouldn’t think about, and how they play into the future of the league and it’s teams. I’ll try and find a link. Very interesting.
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 12:37 PM EST up reply actions
Ahh. Here we go.
It’s called the CSSI or Common Sense Scoring Index. Here is their explanation of how it works. And then here is the actual site… HockeyCssi.Com
Maybe a bit more Red Wing related than I thought but the same theory could easily be applied to any team.
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 12:50 PM EST up reply actions
OT: Brett Connolly Leaves WJC Scrimmage Game with Injury
Was going to post this in the fan-shots… as I’ve yet to see anything about it here. But decided just to post it here, in the true hope that nothing more comes of this “injury”
However, when dealing with a knee injury on a player of that age, it’s never good.
"He got hurt today," said Team Canada head coach Don Hay. "He fell into the boards, lower-body injury, it doesn’t look too serious. We’ll re-evaluate it (Monday)."
Here is the rest of the article. Please be ok Brett, we need you!!
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 12:30 PM EST reply actions
This morning there was this:
@bconnolly8: Everything is okay. Just a little banged up. Not serious at all.
This tweet has since disappeared from Connolly’s timeline, though. Hmmm…
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
I saw it this morning – replied to it, in fact. I’d said that that didn’t exactly inspire confidence. Maybe that’s why it was removed?
"You don't have enough talent to win on talent alone." -Herb Brooks
Raw Charge, an SBN Tampa Bay Lightning community. Follow me on Twitter: @dagmar27.
by Cassie McClellan on Dec 12, 2011 3:30 PM EST up reply actions
Link to my reply: http://twitter.com/dagmar27/status/146237285850030080
"You don't have enough talent to win on talent alone." -Herb Brooks
Raw Charge, an SBN Tampa Bay Lightning community. Follow me on Twitter: @dagmar27.
by Cassie McClellan on Dec 12, 2011 3:35 PM EST up reply actions
Very good article, and a good run-down of what PDO is (I didn’t know until reading your article).
However, I do wonder about the significance of PDO and shooting percentage, when there’s no definitive measure of what constitutes a shot on goal. I’m no “statz head” but I understand the need for a standardized basis of measurement before comparing one stat with another.
"The Revolution will be complete when the language is perfect"
- Smith in Orwell's 1984
That’s certainly a concern, but my understanding of the number-people’s approach is this:
There is a rink-to-rink difference in counting shots but that’s true of a lot of other things that get tracked (hits, giveaways, takeaways, scoring chances, assists, etc.). There are ways to correct for this problem (using only road data, for instance) but that gets very complicated and brings up other biases as well. You can’t stop using shots b/c it is part of so many other measures (save percentage, Corsi, and other things). So you do the best you can with what you have to work with and meanwhile you call publicly for more consistency across the league.
It’s not that “shot on goal” isn’t defined, for those who haven’t read about this. A shot on goal is defined as any shot that has nothing between the shooter and the net but the goaltender. The problem is that in practice that’s hard to determine and the people responsible for actually doing the counting are going to be affected by their environment (esp the crowd) and so some counters err on the high side and some counters err on the low side.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
I’ve kept stats for night league and pick up games before. It’s certainly no easy task, especially with no training, haha. Surprised their isn’t a more automated way of keeping stats, in the works. Would that actually work though? Taking out the human error/biased factor might help. Then we could just blame it on the machines and open up a whole other door of problems/conspiracy theories.
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 2:10 PM EST up reply actions
At the NHL R&D camp I believe Shanahan was saying they’ve tried putting trackers in pucks, but there’s some difficulty. I don’t know what, though.
Red Line Station and @RedArmyLine, featuring coverage of the most frustrating team in the NHL
I believe in next year.
by red army line on Dec 12, 2011 4:38 PM EST up reply actions
So..
The one thing I’ve still yet to find out and correct me if I’m wrong… is what the hell PDO stands for??? Would think it’s an acronym for something, right?
Only thing I could find so far, was this from Pensburgh (which makes me angry I had to be on that site, at all, haha)
What the heck does PDO stand for? I mean, as an acronym? apologies if I missed it somewhere above.
by ahtrap on Jun 24, 2010 7:43 PM EDT reply actions
. I don’t think it stands for anything, but here’s the response I got at Behind the Net:
PDO was the posting handle of some guy. Vic likes to name stats after people who come up with good ideas – hence, Fenwick
by GoPens! on Jun 24, 2010 7:56 PM EDT up reply actions
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 2:28 PM EST reply actions
That’s exactly the case, as I understand it. PDO is named after someone calling themselves “PDO,” Corsi is named after its inventor, Jim Corsi (a goalie coach trying to figure out how much work his goaltenders were being asked to do), and Fenwick is named after Mr. Fenwick. It’s tradition!
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
Hmm..
Weird. Figured it was “Projected D______ Outcome” or something like that. Guess I’m a bit behind, when it comes to all these new statistical methods/processes. Aside from the CSSI, that I mentioned earlier. One thing is clear, the old stat mold is breaking.
"WINGnut--BOLThead"
Let's GO..Wings, Tigers, Lions, Wolverines (oh my!) & Bolts!!!! (RIP Cory Smoot aka Flattus Maximus of GWAR)
by Let's JOE WINGS-BOLTS on Dec 12, 2011 2:57 PM EST up reply actions
Long story short, for reasons my puny brain doesn’t quite comprehend, over the course of a season any team’s shooting percentage plus that team’s save percentage regress quite strongly toward (that is, tend to move closer and closer to) 100%.
That’s because most teams don’t have extraordinarily good or bad shooters or goalies. For Vancouver, for example, their PDO may regress to, say, 1010, because Luongo/Schneider are around .01 better than an average goalie tandem (my estimate).
That being said, there are some teams that are so good with shooting talent and generating/preventing quality shots (the 09-10 Capitals and Avalanche, for example) that they’ll go to higher numbers. But really, it’s hard to sift out the randomness from PDO.
It’s meant to be quick and dirty, but useful nonetheless.
For what it’s worth, looks like that shooting percentage is 16th and the save percentage 28th (per BtN). I think you can safely expect that sv% to go up, meaning GA should come down a fair bit, at least 5v5.
Red Line Station and @RedArmyLine, featuring coverage of the most frustrating team in the NHL
I believe in next year.

by 





















