Every Friday, I make a PDO breakdown graph. This season, tweeting out that graph generally leads to two different discussions:
- How unlucky the Anaheim Ducks have been.
- How very lucky the New York Rangers have been.
Here’s an example – the PDO breakdown graph from Friday, March 11th.
PDO is described really well here. For the uninitiated, PDO is the sum of a team’s save percentage and shooting percentage. In general, we expect these to add to 100.0 over a season.
A team with a value much greater than 100 (see NYR above) is getting lucky and should be expected to regress back towards 100 over the course of a season. A team with a value less than 100 (see ANA) should regress to the mean too.
As of March 11th, most teams had settled in within striking distance of 100. The Leafs, Flames, and NYR are examples of teams the furthest from 100 at this point in the season.
The reasons are varied:
- TOR’s sh% has lagged behind league average all year
- CGY’s sv% is awful
- NYR’s been very lucky
But a question I get into after posting the weekly PDO breakdown is just how lucky/unlucky have teams been? Is NYR’s PDO really far out of line with what we see in other seasons?
To answer this, I’ve graphed each team’s full-season PDO score for every season since 2007 (9 seasons total). Have a look.
*check the larger and interactive view here.
The shaded gray area represents two standard deviations below/above zero. 95% of team seasons are within that range. That PDO range is from ~97.75 to ~102.25. This alone is useful to know – over 9 seasons of data, PDO has almost always landed between 98 and 102.
5% of team seasons fall outside the range. Those seasons (and team seasons on the cusp of exceeding 2 standard deviations) are labeled with the season.
What do we find?
The lockout-shortened 2012-13 season (48 games long) was short enough to allow for several teams to fall outside of the expected range. Also, the Bruins have enjoyed several seasons of good PDO, including 3 seasons in the extremely good range.
To swing back to this season’s NY Rangers, they currently own the fifth-highest PDO since 2007. So, that’s pretty darn high. Even if Henrik Lundqvist is the greatest goalie of our generation, it’s still reasonable to expect NYR’s PDO to fall back to league averages.
As for the Ducks, their PDO this season is low but not historically low. The team should continue to regress towards league average as the season wraps up.
As Travis Yost tweeted, the Ducks remain a scary playoff team.
I can’t think of a team I would rather play less in the playoffs right now than ANA. Not one.
— Travis Yost (@travisyost) March 15, 2016
PDO can only be tracked back to 2007. With 9 seasons of data, the sample size still isn’t large. As more seasons worth of information are added, the range may shift – for now, at least we have a starting point to guide our expectations.
And watch out for the Ducks.
But, what about 2012-13?
A good point raised by Ryan Stimson was that many of the PDO outlier seasons happened during the 2012-13 season. How would the graph look if the shortened season was removed from the analysis?
This is a great point and I’ve done so below:
The crazy (and bad) outlier season for the Panthers and the crazy (good) season for the Leafs disappear, along with a handful of others. After removing the 2012-13 season from the analysis, the standard range for PDO pinches together a little – just under 98 to just above 102.
Can a team like the NY Rangers maintain a PDO score out of expected range for an entire season? Sure, we expect this 5% of the time. But, for the most part, teams have fallen between PDO scores of 98 and 102 over a full season. This is a helpful guideline when viewing my PDO breakdown graphs and a useful piece of knowledge to keep in general.
Questions? Comments? Concerns? Hit me up on twitter @SeanTierneyTss or seantierney1 @ hotmail.com