We’ve already taught you about Corsi, Zone Starts, and First Assist, now it’s time for the fourth installment of the ever popular segment. Today we’re taking on a slightly more complicated one… PDO. The NHL refers to this stat as SPSV%.
Tim Barnes ran the now defunct website Irreverent Oilers Fans. While doing so he used the pseudonym Vic Ferrari. Brian King was another writer on the site and used the nickname PDO. Barnes began delving into stats and came up with the terms Corsi and Fenwick, and eventually with the help of King came up a stat they named after King, calling it PDO. Or something like that, the story gets twisted a lot.
Yeah, but who gives a damn about the name, just tell me what the hell it is. Fine, calm the F down. But first, remember that PDO and SPSV% are the same thing. For the remainder of this article, we are calling it PDO, cause it’s cooler and the NHL is lame for changing it.
PDO is the sum of shooting percentage and save percentage while a player is on the ice. Shooting percentage is number of goals scored divided by total number of shots of all players on that players team while on the ice (oiSH%). Save percentage is number of saved shots divided by total number of shots by the opponent while a player is on the ice (oiSV%).
(Similar to Corsi, it doesn’t matter if an individual player actually takes a shot, all that matters is that he’s on the ice when a shot is taken or a save is made.)
Example time! Erik Haula is on the ice for a total of 10 minutes during a game. During those 10 minutes, the Golden Knights have 20 shots and 1 goal. In those same 10 minutes the Kings have 10 shots and also have 1 goal. Therefore, nine saves were made by the Golden Knights while Haula was on the ice.
On Ice Shooting Percentage (oiSH%) – 1/20 = 5%
On Ice Save Percentage (oiSV%) – 9/10 – 90%
PDO – 5% + 90% = 95
PDO is usually measured against the number 100 because a shot is either scored or saved. Think of it like flipping a coin. It’s either heads or tails, so over the course of all flips, the percentage of heads plus tails will be 100%. Under the same mathematic principle, the average of all shots made plus all shots saved will be 100%.
The working theory in hockey is about 8% of shots are made meaning 92% are saved. So, over time, player’s PDO numbers should regress toward 100 (8+92). Of course, that’s ridiculous because some players are better than others and luck is heavily involved in hockey, but a player’s PDO can often be an indicator of the future of a player’s season or career.
Let’s look at some Golden Knights PDO numbers. Forwards first…
| Player | oiSH% | oiSV% | PDO |
|---|---|---|---|
| Oscar Lindberg | 8.8 | 93.7 | 102.5 |
| Erik Haula | 8.8 | 92.8 | 101.6 |
| Reilly Smith | 8.5 | 92.4 | 101.0 |
| Teemu Pulkkinen | 7.1 | 93.7 | 100.7 |
| James Neal | 9.1 | 91.2 | 100.3 |
| William Karlsson | 7.8 | 92.3 | 100.1 |
| David Perron | 8.7 | 91.2 | 99.9 |
| Pierre-Edouard Bellemare | 6.4 | 93.4 | 99.7 |
| William Carrier | 5.8 | 93.9 | 99.7 |
| Cody Eakin | 8.6 | 91.0 | 99.5 |
| Jonathan Marchessault | 7.5 | 89.0 | 96.5 |
Now defensemen…
| Player | oiSH% | oiSV% | PDO |
|---|---|---|---|
| Nate Schmidt | 8.5 | 93.3 | 101.8 |
| Jason Garrison | 8.1 | 92.7 | 100.8 |
| Clayton Stoner | 7.2 | 93.1 | 100.3 |
| Deryk Engelland | 7.8 | 92.4 | 100.2 |
| Luca Sbisa | 8.1 | 92.1 | 100.2 |
| Griffin Reinhart | 7.8 | 92.1 | 100.2 |
| Brayden McNabb | 6.6 | 93.4 | 100.0 |
| Shea Theodore | 7.8 | 92.1 | 99.8 |
| Jon Merrill | 6.8 | 91.4 | 98.2 |
| Colin Miller | 6.7 | 91.1 | 97.9 |
What PDO does best is predict the future. Outside of the rare instances of elite players, especially ones playing with elite goalies, players’ PDO number will regress to the mean of 100.
So, Jonathan Marchessault, Colin Miller, and Oscar Lindberg should jump off the page. Marchessault and Miller’s numbers should likely come back up towards 100, while Lindberg could be in for a year sending him back down towards 100.
PDO is also an excellent indicator of the “reality” of a players stats. In other words, the likelihood that a player can maintain the numbers he put up a year before or if his numbers were a fluke. Let’s take the two Panthers, Smith and Marchessault.
Smith scored 15 goals last year and 25 the year before. It should come as no surprise that his PDO last year was 98.3 (1.7 below 100) and 103.1 (3.1 above 100) the year before. PDO already predicted that his 103.1 would come back down and he would score fewer goals in 2016-17 than he did 16-15. PDO can also predict the future in saying that Smith should fall somewhere in between this coming season.
Meanwhile, Smith recorded 22 and 25 assists in those seasons, even with a ranging PDO. His assists number would be expected to stay similar no matter what his PDO ends up being in Vegas.
Now Marchessault. He scored 30 goals in 2016-17, his first full season in the NHL. His PDO… 97.2! In just 45 games the year prior, he scored seven times, and his PDO was even lower at 94.3. This is sports, and nothing ever for certain, but 30 goals with a PDO 2.8 below 100 is an excellent indicator that he at least maintain, if not improve on his goal scoring numbers (or at least his +/-) as his PDO regresses to 100.
McNabb, Reinhart, Sbisa, Stoner, Theodore, Karlsson, Neal, Perron, Bellemare, and Carrier, are all within 0.3 of 100, meaning when you dive deeper into their goals, assists, +/-, and even Corsi, you can have a bit more belief in those numbers from last season.
Things are going to change drastically as all of these players come together on the same team. No longer does James Neal have Pekka Rinne and Filip Forsberg factoring into his PDO, or Nate Schmidt having Braden Holtby or even Alexander Ovechkin, but PDO is surprisingly accurate at validating or giving reason to disbelieve prior season’s stats.
So now that you know what PDO is and how it works, dig into some of the new Golden Knights numbers and see what you find. Oh, and post them here, we’re always interested in a good advanced stats nugget.
James
@Jason Pothier
I view PDO as a proxy for how lucky a team is rather than an individual …
Jonathan Marchessault shot an impressive 15.5 percent last season. His average is 13.7 percent. That indicates to me that he wasn’t unlucky.
Cody Eakin shot just 3.7 percent last season. His average is 10.6 percent. That indicates to me that he was unlucky.
Phisig150
Goalie play seems like it would throw this stat off. Although maybe if you look at a 3 year average it might be more accurate. Dustin Brown has horrible 3 year PDO and Jagr’s is good so I like this stat since it confirms my biases.
James
@Phisig150
It is a good indicator of whether or not a team is lucky. A recent case in point is the 112 points 2013–14 Colorado Avalanche who didn’t have the ability to control the play of a game, Semyon Varlamov played above his head in goal and they shot the lights out … PDO indicated that the hot streak wasn’t sustainable in the long run
Ken Boehlke
We purposely left the word luck out of there because there’s kind of a bad connotation to the word luck. I’m with you that luck is involved in a lot of it, many factors are in why a players PDO is so low or so high, and like every stat, if you dig deeper in some aspects, you’ll be able to validate or invalidate some of the numbers. But overall, I do believe individual PDO is a pretty good indicator of future performance.
James
@Ken Boehlke
I’m not an absolutist about advanced stats. I believe that corsi isn’t the be all, end all – The Pittsburgh Penguins were outshot throughout the playoffs. I believe that shot quality for and against matters … The advanced stats absolutists tend to dismiss shot quality because it’s totally subjective. People have different interpretations when watching game film – The advanced stats absolutists pride themselves on being totally objective, so they prefer Corsi numbers.
My favourite soccer team does not dominate possession like Barcelona, but is deadly on the counterattack and doesn’t give up many Grade A chances. It’s harder to set up your defensive structure in transition – hence why most goals are scored on the rush in hockey.
Some strikers finish better than others. Steven Stamkos and Patrik Laine fit the bill in the hockey realm.
James
@Phisig150
‘Dustin Brown has horrible 3 year PDO and Jagr’s is good so I like this stat since it confirms my biases.’
Do you like the relentless media coverage of Tim Tebow? Jagr has become a sideshow. People resent him because he overshadows the team …
He also plays at a slow pace – Arguably a detriment to his young linemates in Florida.
Phisig150
Being a Broncos fan I did enjoy the year of Tebow it was fun. A sideshow is exactly what this team needs. West Point Nevada is a little too dry and boring. We need something to shake things up a bit. He may be slow but advanced stats seem to love him.
James
@Phisig150
It depends on your style of play. Jagr is akin to the traditional plodding big butt power forward in basketball who used to work in the low post. He’s great at creating a body shield, putting his body between the defender and the puck.
Not a good fit for an uptempo offense.
He also wants to play a prominent role on offense to pad his career stats and he practices on his own schedule. He’s better suited to be a role player
James
@Phisig150
I’m more and more coming around to the Jagr idea. I think he’s good enough to start in our top-six forward group. He can still put up good possession numbers by grinding it on offense.
The biggest downside is that he takes away the opportunity from a younger, unproven player with potential. I’m not cool with that. I’m more interested in the future than in the present…