Login

grok · 08-03-2020, 09:03 PM

[DO NOT GRADE. This media piece is sponsored by SHL Head Office]

PDO: SHN Advanced Stats Guide Chapter One

With the introduction of FHM to the league, we have a ton of fancy new statistics available to measure the performance of teams and players. Of all these new metrics, PDO is one of the most confusing. "It's a measure of team luck," you might have heard. But should we use it that way? And if not, how should we?

What is PDO, literally?

PDO, named for its innovator's online handle, adds together shooting percentage and save percentage. For a team, it's their total SH% plus their total SV%. For a player, it's their team's SH% plus their goalie's SV% while that player is on the ice. Basically, for every minute you're on the ice, add up the whole team's SH% and the goalie's SV% and you get your PDO (usually measured to the nearest tenth of a percent).

Why do we measure PDO?

There are two key reasons we consider PDO. First, if you take all shooting percentages and save percentages across the entire league, it will always equal 100%. If the league save percentage is 92.0%, that means 8% of shots on goal across the league must have gone in. This cross-league balance is immovable--it is impossible to have one shot on goal count for two goals or for two saves, and vice versa.

Second, the distribution of shooting percentage and save percentage values are, effectively, random. Now, they aren't totally random—some players are better than others and some goalies have very good seasons—but across multiple seasons it's incredibly hard to find a stable SH% or SV%. It continually varies as the league evolves, even within a single team. Even the best teams have SH% and SV% variance based on how many shots their team is taking, how they drive play, and how well they suppress shots from other teams.

How can we actually use PDO?

Well, let's re-evaluate the phrase "It's a measure of luck" with this new understanding of what PDO measures. If you're converting more shots into goals and getting more saves than the league average, you're probably overperforming at least a little bit. This is especially true if a player's PDO is dramatically higher than their team's PDO. You can usually expect a player to regress back toward 100 PDO over time. The same works in reverse--if a player's PDO is dramatically lower than their team's PDO, they're probably underperforming and will improve towards 100 over time.

If this seems too good to be true, a one-stop shop for "player luck"...that's because it is. PDO doesn't exist in a vacuum. We can't assume that a person with PDO over 100 is just lucky. What we can do is use PDO combined with other stats for context. A player with a PDO of 105 who is also a top 10% player by Corsi probably deserves that PDO rating. By correlating these two stats, we can determine that the player drives play, outshoots their opposition, and gets their line to score a ton as a result. On the other hand, a player who also has a PDO of 105 but has a bottom 10% Corsi probably got some lucky shots or lucky saves and will regress down to the 100 PDO mean over time.

Ultimately, PDO is most valuable when you use it compared to the league, the player's teammates, and the player's own down-line advanced statistics. It is a measure of luck, but it's one we use best in the context rather than in a bubble.

[DO NOT GRADE. This media piece is sponsored by SHL Head Office]