Frozen Tools Forensics: PDO levels early in the season
I will preface this article by saying that I know there are severe limitations around PDO as a stat and this is not meant to provide an analytical critique or an in-depth scouting report of NHL players. It is intended to give fantasy owners a clue as to whether a player might continue to perform at their current level.
The definition of PDO is:
“PDO, called SPSV% by the NHL, is the sum of a team's shooting percentage and its save percentage. PDO is usually measured at even strength, and based on the theory that most teams will ultimately regress toward a sum of 100, is often viewed as a proxy for how lucky a team is.” (Wikipedia)
Essentially, if the sum is well above 100, we can assume the team or player is abnormally lucky; conversely, significantly below 100 means unlucky. We will start with the teams in the NHL that are having the worst puck luck and should improve. Admittedly, PDO is influenced by exceptional goaltending by your team or the opposition, but over the course of the season it will balance out to within certain ranges.
We are focusing on PDO numbers outside of those ranges.
Los Angeles is the worst in the league at 94.3 as of October 24th, followed by Vegas at 96.9. Other teams that range from 97.2 to 98.2 and should improve are Columbus, San Jose, Calgary, Florida, Winnipeg, and Dallas.
Next are the teams with the best puck luck at the moment, likely to revert as more games are played. The luckiest team at the time of writing is Colorado at 105.9, followed by Buffalo at 105.4. Other teams that range from 103.9 to 102.2 are Nashville, Pittsburgh, Arizona, and Montreal.
Put into perspective, the teams with the highest PDO for all of 2018-2019 were the NY Islanders (102.3), Washington (102.2), Tampa Bay (102.0), and Toronto (101.9). The unluckiest teams in 2018-2019 were Minnesota (98.5), Florida (98.6), and New Jersey (98.7).
St. Louis won the Stanley Cup with a PDO of 100.2 for the whole season. On January 1st, 2019, after 37 games, their PDO was 98.7 and they were fifth lowest in the league. They reverted back to the middle for the rest of the year and had a PDO of 101.6 in the second half.
The highest PDO in 2017-2018 was Tampa Bay with 102.3, and the lowest was Buffalo with 97.8, so I think it is safe to assume that teams outside of 102.5 and 97.8 will revert to within that range over the course of this season.
Individually, the ranges are a bit wider over the past two seasons for players (50 games or more), with a high of 106.9 (Auston Matthews in 2017-2018) and a low of 92.5 (Brandon Dubinsky in 2018-2019). Both those PDOs are a little out of the typical range, which is 94.0 to 105.0. We will focus on players starting this season outside the typical range, starting with the unluckiest players (after a minimum of six GP) taken from Dobber’s Frozen Tools Report Generator:
Adrian Kempe (81.6) and most of the Kings, including Ilya Kovalchuk (83.5), Jeff Carter (90.3), and Drew Doughty (90.8), should get back up above 94.0 in PDO by the end of the season. Kovalchuk and Doughty are off to decent starts and on pace for 64 and 55 points each, so it is their plus/minus that should improve the most, and perhaps their points marginally. Kempe and Carter have only two and three points apiece; Kempe has no goals on 22 shots and Carter has one goal on 33 shots, which will improve.
For Toronto, it is surprising to see Mitch Marner (87.2), John Tavares (90.4), and Tyson Barrie (92.9) on this list as Marner is on pace for 89 points and Tavares had seven points in his first eight games before being injured. Barrie has been a little disappointing with only four assists in 11 games, but that is hopefully temporary.
Looking at all the players on this list, their 5 on 5 shot percentages are much lower than normal, along with their team’s 5 on 5 save percentage.
Next are players in the range of PDO from 108.7 to 114.2, much higher than the usual top end of 105 or 106. All of these players will fall into the normal range with more games played.
Joel Armia (113.3) has had a very good start to the season with five goals in eight games and is on pace for 62 points. His career high is 13 goals and 29 points. With his early start, there are a few people who figure he is a cinch for 20 goals. PDO is perhaps telling us that his start might be too good to be true. It’s not that his shooting percentage of 13.3 is particularly high in the NHL for a forward, but it is high compared to his historical percentages.
Brock Boeser (112.7) is next and Canucks fans should be a little worried with Elias Pettersson (109.5) and J.T. Miller (110.6) also being on this list. Miller (who has probably been the best Canuck forward to date) should level out closer to his previous career high of 56 points rather than his current 91-point pace. Boeser and Pettersson have actually had a slow start to this season (by their standards) and with their PDO being so high, it’s a little troubling, Of course their PDO numbers are inflated because of the great goaltending the Canucks have had so far, but it’s 100% certain that all three players will have less success on net or the team will allow more goals while they are on the ice. Neither scenario is good for them moving forward.
Other notable players that will be affected are Casey Mittelstadt (112.4) on a 57-point pace, Jeff Skinner (109.3), Brad Marchand (109.5) on a 128-point pace and Ryan Ellis (110.5) currently on a 98-point pace. If you’re a fantasy owner and have a player on this bottom list, perhaps look to sell high if it’s not a keeper pool and your pool has plus/minus.
I’ve heard people say that PDO is the most overrated and useless stat in hockey, and that five on five shooting percentage should be looked at first. I agree with that for the most part, but the reason I like PDO in addition to shooting percentage is that it is a certainty for teams or players outside the usual range to revert to that range 100 percent of the time. Players can and do have high or low shooting percentages in back to back seasons, so it’s more difficult to tell within a season how they will do.
If you have any comments or suggestions, please let me know below and/or follow me on Twitter @gampbler15.
No data at this moment.