|
Displaying pretext
Team PDO
Vancouver 102.6
Boston 102.5
Minnesota 102.2
Philadelphia 101.8
Dallas 101.4
...
Toronto 99.2
St. Louis 99.2
Tampa Bay 99.0
Ottawa 97.8
New Jersey 97.8
As of March 3, 2011
Three of the top four teams in the league in points are also three of the top four teams in PDO, and four of the five teams with the worst PDO can all currently be found in the league's basement. There's obviously something that can be learned from PDO, so what is it?
PDO was introduced by Vic Ferrari a few years ago and while nobody remembers what it stands for, everyone knows how it's calculated. Simply add save percentage to shooting percentage, and you should wind up with a number relatively close to 100.0 (or 1000, as some people prefer to present it). And technically you're supposed to use even-strength numbers only, something we'll skip for simplicity.
While further description of PDO is outside the scope of this article (but can be gained from either Tyler Dellow or Corey Pronman),it's clear to see why PDO is so closely associated with the standings. When we're talking about upwards of 5000 shots for and against per season, even a fluctuation of 1.0 means 50 goals, which is easily enough to knock a strong team out of the playoffs or to push a weak team in.
While it's been previously established that team PDO has a strong tendency to 100.0 due largely to the heavy dependency that shooting percentage has on luck (about 75%), Gabriel Desjardins discovered that the presence/absence of skilled players can sustain a team at a significantly different level for several consecutive seasons. Therefore, it is probably best to compare each team to their established PDO levels instead of 100.0 before predicting jumps or slides.
| 
Howe and Why (02/19)
2 comments have been left for this article. 
I have long been against PDO for reasons explained in the post Rob linked to.
I think it has merit on the team level, but at the individual level I just don't see the sense in using it unless you were looking at goal-based metrics, but the stats community nowadays doesn't use goal-based metrics as much so its usefulness kind of dies with that.
If you want to see if a player's Assist rate is being influenced by luck, you would look at his on-ice shooting percentage. However for what reason would we ever look at a player's on ice- save percentage other than to see how much his GA/60 is being influenced by luck? Especially since we have much better shot-based metrics which wouldn't be influenced by on-ice save percentage.
Agreed. Player-based PDO basically just tells you how lucky a player is. That can still be a useful thing to know.
And yes, when dealing with players over a time span of a single season or less, many analysts prefer shot-based metrics over goal-based, since the creation and prevention of shots is a more persistent skill than goals, and the larger sample size reduces the influence of luck.
Nevertheless, goal-based metrics can still have their uses (cue Timo).