Method
The players were ranked by their total Productivity Rating Scores (see Introduction to Productivity Rating for more information) over the last three seasons. PR-Score was designed so that it could be added across seasons, which allows for the calculation of average PR over several seasons and for the calculation of Value Rating (VR).
Using PR-Score this way would be like identifying the best goal scorers since 2019 by finding the players who scored the most goals in that period. It also suffers from a shared perception problem: players who miss games will have a lower than expected PR-Score, just as players who miss games will score fewer goals than expected.
I’ll show the top fifty players of the last three seasons in groups of 10, starting from the top.
Best Players Since 2019 – Places 1 Through 10
Here’s a quick guided tour of the table:
- Rank is the player’s rank in all players who have played since 2019/2020, and is followed by his name, his team and his position.
- Tot-PR is the total PR-Score of the player from 2019/2020 through 2021/2022.
- The Big 5 Stats are:
- GP/82 – games played per 82 scheduled games.
- TOI – average time on ice per game played.
- G/82, A/82, Pts/82 – goals, assists and points per 82 games played.
- The last three columns show the PR-Categories a player has been in each of the last three seasons.
While Tot-PR is an abstract number, seeing a bunch of them will give you context. Tot-PR is like the marks for diving in the Olympics: the underlying calculations that lead to the number is complicated; a higher score is better; and the bigger the spread between two scores the more likely it is that the higher score is truly the better player (or dive). I’d say a difference of more than 0.25 in PR-Score is convincing, while a difference of less than that leaves room for debate.
If you believe that Mika Zibanejad was better than Adam Fox since 2019, I will grant you that the “evidence” for Fox being better leaves room for reasonable doubt (their Tot-PR difference is 0.004). Your opinion would be reasonable. If you believe that Victor Hedman was better than Auston Matthews, the “evidence” that Auston Matthews was better is convincing (Tot-PR difference is 2.741) and there is no reasonable doubt. Your opinion would be unreasonable.
The top two players from the last three seasons are Connor McDavid and Leon Draisaitl, both from Edmonton. Edmonton hasn’t had much playoff success these last three seasons, but PR is a measurement of regular season play only. They do play on the ice at the same time a lot, and it absolutely affects their statistics, as the table below shows.
The data in the table is based on five-versus-five (5v5) play from the last three seasons. When Connor and Leon are on the ice, it is the return of the high-scoring game that was played in the 1980s. The pace of scoring is 133% greater when they are on the ice together than when they are on the ice individually, 175% greater than when they are both off the ice. As most hockey statistics measure offense, anything that increases offense will inevitably increase PR-Score.
Number 3 on the list is Auston Matthews. While he doesn’t have the absolute star teammate that McDavid and Draisaitl have with each other, he does play Mitchell Marner (13th on the list), Tavares (55th), Nylander (76th) and Rielly (78th).
The top three are clearly the top three players in the last three seasons: fourth place is 2.741 PR-Score points behind. Speaking of fourth place, Victor Hedman is the only guy in the top ten with bona fide playoff success, having won two Stanley Cups and appeared in three finals in the three seasons.
Roman Josi is one heck of a defenseman. Brad Marchand is the oldest guy in the top ten, and John Carlson the second oldest (tied with Hedman and Josi). Adam Fox is the youngest, and his teammate Mika Zibanejad has the most hair of anybody in the top ten. J.T. Miller rounds out the top ten. His team is on the left coast so we in the East don’t get to see many of his games, but the data says he is exceptional.
Best Players Since 2019 – Places 11 Through 20
Now is as good a time as any to pay a little attention to the last three columns in the table. The top ten players had 19 PR-Elite seasons, 8 PR-Star seasons and 3 PR-First5 seasons.
The second ten have 7 PR-Elite seasons and 23 PR-Star seasons. That’s a clear drop in the quality of seasons this group of players produced. No player outside the top ten has had multiple PR-Elite seasons since 2019.
The spread in total PR-Score between 11th and 20th (1.277) is smaller than the spread between 2nd and 3rd (1.364) and much smaller than the spread between 3rd and 4th (2.741).
From about 8th place in the list, the differences between slots are slight, meaning that the guy in 20th place could arguably be 18th, or could arguably be 22nd.
Maybe it’s because of where he plays compared to where I live (Ottawa), but Ryan O’Reilly is either underrated or under-covered (in terms of the press).
Patrice Bergeron and Brent Burns were the oldest players in this group, at 36. The youngest players, at 24, were Mitchell Marner and Sebastian Aho. Anze Kopitar, who it feels like has been playing since the 1940s, is 34 years young.
Barkov would have ranked higher, possibly in the top ten, had he not missed so many games. MacKinnon, who has the only Stanley Cup (in the last three seasons) in this group would also have ranked higher with fewer games missed. Darnell Nurse may well get a boost in his PR-Score because he has teammates named McDavid and Draisaitl.
Best Players Since 2019 – Places 21 Through 30
The third ten players have 3 PR-Elite seasons, 19 PR-Star seasons and 8 PR-First5 seasons. As expected, a drop in the number of outstanding seasons.
This group delivers the first players who are both in the top fifty and who played for multiple teams, so they have a “2” in the team column. Pietrangelo played for St. Louis and Vegas; Jones played for Columbus and Chicago; Devon Toews played for the Islanders and Colorado. Jones got to Chicago just in time for them to ditch almost every good player they have, while Toews got to Colorado just in time to win the Stanley Cup.
Cale Makar is youngest in this group, at 23. Kris Letang and Patrick Kane are the seniors, at 34 and 33 respectively.
Over the last three seasons, Kyle Connor has been scoring 40+ goals per 82 games, Mark Scheifele is scoring 87 points per 82 games, Patrick Kane is at 98 points per 82 games, and Artemi Panarin is at 109 points per 82 games. It’s not top-ten offense, but it’s certainly nothing to sneeze at.
What can one say about Charlie McAvoy? He’s a very good player who also delivers hits: he’s 7th in the top 50 in hits delivered per 60 minutes. His offensive numbers are similar to Pietrangelo, Seth Jones and Devon Toews, which is a good level of offense for a defenseman. And he’s young (25).
Best Players Since 2019 – Places 31 Through 40
The 31-to-40 group had 1 PR-Elite season, 21 PR-Star seasons and 6 PR-First5 seasons.
Mikko Rantanen and Joe Pavelski are the first two players in the top 50 to have had seasons rated below PR-First5. Rantanen played just more than half-a-season in 2019 for his PR-Regular rating. Pavelski played nearly a full season in his 2019 Dallas debut and played poorly (for him). In his last two seasons, he has rebounded nicely. They were his 3rd and 5th best seasons since 2007/2008 – not bad for a guy who was 37 last year.
Quinn Hughes is the youngest of the top 50 players, at 22. I’ve mentioned it in passing before, and so I’ll mention it here as well: it is a very good omen for a player when he shows great play early in his career.
It’s strange, but I would have bet money that Pastrnak was older then Huberdeau. It must be the crowd that Pastrnak hangs around with in Boston: age by association.
Ivan Provorov is another good young talent, but it would be surprising if his career keeps pace with Quinn Hughes. While they are almost identical in PR over the last three seasons, Provorov is two years older and scores less frequently. The same is true for Jacob Slavin, but he’s five years older than Hughes.
It took a free-agent signing, but Ottawa has a top-50 player in the league. Giroux actually became their second top-50 player, as a trade brought #50 (DeBrincat) to Ottawa just before he signed. They go together with Thomas Chabot (64th) and Brady Tkachuk (77th). They would have had a fifth top-100 player if they had held on to Connor Brown (84th). As he moves from one nation’s capital to another nation’s capital, I wish him well.
Ryan Suter, the second oldest player in this group, makes a headline when he misses a game. In the last ten seasons, he has missed fewer than a dozen games. He’s been off his early-30’s form these last two years, but he’s damn good for a 37-year-old.
Best Players Since 2019 – Places 41 Through 50
The final group had only one PR-Elite season, 17 PR-Star seasons and 8 PR-First5 seasons since 2019.
Matthew Tkachuk and Alex DeBrincat (brother of a Senator and a newly acquired Senator) are the youngest, at 24 years-old. Mark Giordano is the oldest, at 38. His signing with Toronto at the league minimum should be very good for the Leafs. He is old, but he’s not league-minimum old.
Jeff Petry is as old as Sidney Crosby, but not quite as accomplished over the course of their careers. The reason Petry is above Crosby is called “games played”. Petry played 30 more games in the last three seasons.
MacKenzie Weegar got a late start to his career. His first good season (PR-Regular or better) came at age 25. He really hit his stride at age 27, with two PR-Star seasons.
Steven Stamkos has had an up and down career. His best stretch was 2009 through 2012, with three PR-Elite and one PR-Star season. 2013: injury. 2016: injury. 2019, 2020: lots of missed games. But last season was another PR-Elite.
Jared Spurgeon’s career value rating (VR) is exactly what you’d think career value should look like: it starts out low, it hits a peak, and it starts to decline.
To be fair, many career VR charts look like this, but the regularity of Spurgeon’s increases and decreases from season to season is simply amazing.
Summary
The question was: who were the best players these last three seasons? I chose to answer the question by ranking players by their total PR-Score. There isn’t any controversy at the top of the list. If you find that some players are too lowly ranked in (or missing from) the top fifty, there are three probably reasons.
The Time Frame
This article ranks the best players from 2019 through 2021, period. In 2012, Zdeno Chara would have easily been in the top twenty. After the 2022 season, Kirill Kaprizov would almost certainly make this list.
Click on this link to see the complete list of 122 players.
Games Played
Connor Brown was a bit of a surprise on the list, ranking 84th. At 83rd place is William Karlsson of the Vegas Golden Knights while Andrew Copp (WPG, NYR) is 91st.
All three are very close in their Big 5 Statistics, so it makes sense that their total PR-Scores are close to each other. Given that other players with quite similar statistics are ranked similarly, it is likely that their rankings are “correct.”
Connor Brown played 191 games these last three seasons, and ranks higher than Sean Couturier (143 games played), Jakob Chychrun (166), Gabriel Landeskog (159), Evander Kane (163), Filip Forsberg (171) and Erik Karlsson (158). All those players, and some others, would have ranked higher than Connor Brown if they had played the same number of games as he did.
Fan-centric Point of View
There is an outside chance that the player you think should be in the top fifty isn’t really that good. As noted elsewhere, the PR formula doesn’t look at press clippings. It doesn’t give a player more points because his name is Bobby Clobber or because he plays for the Kraken.