The term overachiever is usually seen as a positive descriptor. By definition, an overachiever is someone who achieves success over and above the standard or expected level. But right there is the key — over the expected level. In hockey, more often than not overachievement leads to a letdown just one year later.
Luck plays a pretty sizable role in hockey. Goals win you games, and the ratio of goals for to goals against with any given player on the ice can be quite volatile year to year. The lack of repeatability, especially in extreme cases, is why things like Corsi and Expected Goals are even desired in the first place. If previous goals had a strong connection to future goals, there’d be no need.
That lack of repeatability is exactly what we’ll be exploring today.
Regression. Not skills regression, but statistical regression. When it comes to the game of hockey, on-ice percentages are the go-to to measure for luck. Added together, a player’s on-ice shooting percentage and on-ice save percentage will give you a player’s PDO. Over time, it is generally accepted that a player’s PDO will regress towards the mean; 100. This also true at the team level.
While PDO itself has fallen out of favor over the last few years, the two statistics it encompasses are still extremely useful when it comes to finding anomalies.
With how strange the 2018-2019 season was, it’s hard to make too much out of anything. The team saw a new general manager take helm, dealt with a mid-season coaching change, and played in front of eight different goaltenders. Still, here’s one that stands out in retrospect.
Among Flyers regulars during the 2018-2019 season, only Jori Lehtera and Ryan Hartman finished the season with a lower on-ice save percentage than Sean Couturier at five-on-five. Opponents scored on a higher percentage of shots with the now-Selke winner on the ice than with any other player that played in at least half of the games. That probably sounds weird, and it should. It had nothing to do with Couturier’s play during that season, because — and this has been said plenty of times before — skaters do not drive their own on-ice shooting, or save percentages, it’s almost entirely random.
This past season, Couturier’s percentage moved up three points and was the sixth-best among lineup regulars, up where you’d expect to, and usually do, find his name. In this case, it really didn’t have an effect on the general public’s perception of Couturier, because he both maintained a high on-ice shooting percentage, and because the entire team had a goals against problem. In some cases, both percentages fall the wrong way and you’re looking at an off-season of “what happened?” think pieces.
Take ‘18-19 Shayne Gostisbehere for one. A middle of the pack on-ice save percentage combined with a team-low on-ice shooting percentage equaled what was statistically the worst season of his career. Of course, we know now that a bounce-back season was not in the works, as he’d go on to have an even tougher ‘19-20. A season that actually wasn’t driven by some unlikely numbers, but by some truly poor outings.
Further examples include Travis Sanheim’s 2017-2018, as well as Jakub Voracek’s 2016-2017 campaign. In other cases, the percentages fall the right way, making expectations soar.
The Flyers were a very good hockey team last season. And as it stands, the Flyers look to be a very good hockey team next season. But what they also have are a handful of players who likely won’t replicate what they were able to do in 2019-2020.
Scott Laughton, Philippe Myers, and Robert Hagg. One thing that these three players have in common is that they each finished the ‘19-20 season with an on-ice shooting percentage north of eleven percent at five-on-five. This is important because history tells us that it’s highly unlikely that they will benefit from such a high on-ice shooting percentage next season, as only a handful of players have over the last ten years. Just five players out of 138, to be exact, had back-to-back seasons with an on-ice shooting percentage as such.
We care about this because it’s a large reason why their Goals-For percentages (GF%) were as high as they were this past season, which has a large impact on our eye tests. The following Tweet from @JFreshHockey, who you should definitely follow if you do not already, shows just how much different Laughton’s expected on-ice shooting percentage was from his actual percentage.
For those curious, here's the 49 skaters with a 5v5 on-ice shooting % above expected of 3% or higher in 2019-20.— JFresh (@JFreshHockey) November 6, 2020
Based on past data, the vast majority of these players will see a significant regression in on-ice shooting next season. https://t.co/CyB9MRphHt pic.twitter.com/7xqxnfZLni
In Laughton’s case, 4.6 percent fewer goals scored while he was on the ice would take his GF% down from 61 to 48 — a sizable difference. It’s fair to assume that had his on-ice shooting percentage even been close to league average, he would have also put up fewer points, and the entire perception of his season shifts.
But here’s the thing, we can’t take away what did actually happen. Laughton had his best season to date, and the Flyers found a lot of success with him on the ice. Him having benefited from luck doesn’t take any of that away from him, but it does tell us that the circumstances surrounding his offensive breakout are unlikely to repeat.
Myers and Hagg are similar cases, but with Hagg, not only did he finish the season with a high on-ice shooting percentage, but also a high on-ice save percentage as well. The only other semi-regular that finished the season with a higher on-ice save percentage at five-on-five was Connor Bunnaman, who appeared in 21 regular season games. With Hagg on the ice the Flyers converted on 11.2% of their shots on goal, while the opposition converted just 5.4% of the time. That’s just not something that repeats if you view all shots as equal, and especially not when you take into account the quality of those shots.
Now, you may have noticed there are two more Flyers on that list of 49 players — Joel Farabee, and Tyler Pitlick, the latter of the two now a member of the Arizona Coyotes. They too are expected to have their on-ice percentages regress next season, and their inclusion has the Flyers tie with the Colorado Avalanche with the highest amount players on the list of overachievers.
While that’s not something you necessarily want to see, this doesn’t mean that the Flyers are set to fall backwards next season. Just as it’s likely these players won’t get to enjoy as much puck luck as they did previously, others will see the pendulum swing in their favor.
While we’ve kept our focus on the on-ice percentages, individual shooting percentages are just as telling. In August of 2019, Craig took a look at the team’s luckiest and unluckiest shooters. Below was his concluding paragraph:
Every season is a new season. Whatever line, linemates, role, and ice time a player has may be the same year in and year out, but there are seasons where a player catches the breaks and years they don’t. If JVR played an additional 15 games this season he most likely breaks 30 goals, but he could play all 82 games in 2019-20 and still post 27 goals simply because his individual Corsi shooting percentage is anticipated to regress. Konecny is expected to spend a good amount of time (if not the entirety) of the next campaign on the top line with the same linemates, so it isn’t hard to picture the winger producing the same chances. Depending on which way the pucks bounce off iron, Konecny may be on his way to his first 30-goal season in the NHL.
Both van Riemsdyk and Konecny’s goal outputs this past season were just about nailed before the puck even dropped. van Riemsdyk’s pace fell even further than the proposed 27-in-82, while Konecny had scored at a 30-goal pace at the time of the stoppage.
However, Konecny can now be found on the opposite side, as both he and Laughton scored on over 17% of their shots on goal this past season — a rate well above their career averages, and well above some of the league’s best shooters. Though not impossible, it would be relatively safe to assume that those individual shooting percentages won’t continue in 2021.
And when they don’t, it won’t be because Laughton or Konecny’s shooting ability will have gotten worse, it will be because they are no longer overachieving. And really, that’s the case for every player we looked at today.
Data via Natural Stat Trick