Is Miro Heiskanen being overworked? Examining ice time

VANCOUVER, CANADA - MARCH 14: Miro Heiskanen #4 of the Dallas Stars waits for a face-off during the second period of their NHL game against the Vancouver Canucks at Rogers Arena on March 14, 2023 in Vancouver, British Columbia, Canada. (Photo by Derek Cain/Getty Images)
VANCOUVER, CANADA - MARCH 14: Miro Heiskanen #4 of the Dallas Stars waits for a face-off during the second period of their NHL game against the Vancouver Canucks at Rogers Arena on March 14, 2023 in Vancouver, British Columbia, Canada. (Photo by Derek Cain/Getty Images) /
facebooktwitterreddit

Miro Heiskanen is an elite #1 defenceman in the National Hockey League. After a career offensive year this past season with 73 points in 79 games, I was curious whether he was being overworked because of both the need for his skill on the ice and the lack of elite, trustworthy alternatives for his on-ice presence.

Some players play better when they get more ice time, as they get into the rhythm and tempo of the game and keep the “blood pumping” rather than sitting “cold” on the bench. On the other side, some players benefit from periods of rest to recharge and breathe before ‘giving it all’ on the next shift.

For reference, Miro Heiskanen averaged 25:29 total ice time and 20:38 of even-strength (ES) ice time in the 2022-2023 regular season. Heiskanen ranked 6th in total ice time and 8th in ES ice time. He was among the top defensemen in the league for this and Dallas was evidently better when he was on the ice. If Dallas had secondary options that could provide such an impact and eat some of Heiskanen’s minutes, it is possible that he may perform better with more rest.

Furthermore, Dallas’ roster, in recent years, has boasted a deep forward group that contains gifted puck-possession-oriented players including Roope Hintz and Wyatt Johnston (among others) that can transition consistently and take the skating burden off of someone Heiskanen in combination with some other defensemen stepping up.

Today, I will take a look at Heiskanen’s 2022-2023 regular season game-by-game results and look at ice time and opponent to see what we can find regarding Miro’s outputs when playing more or less minutes.

Would Heiskanen be better utilized in fewer minutes to be able to have more energy that could result in better overall outputs?

Process

In order to tackle this idea, I used Heiskanen’s game-by-game GameScore info from hockeystatcards.com. GameScore is a number that combines offensive and defensive metrics together at varying weights to try to provide a ‘GameScore’ for a player, with a higher score meaning that a player had a ‘better’ game. GameScore does not always tell the whole story, but it provides a good standard for whether someone like Heiskanen had a good, average, or bad game.

In a 5v5 game, one player’s ice time being marginally more or less will likely have a minimal correlation to team success over a long period of time. The argument here is hypothetical based on past results and “eye test” viewings of the player and this evaluation seeks to optimize a player’s game management for the best possible contributions to his team.

In Microsoft Excel, I broke up the GameScore numbers into 3-minute, 2-minute, and 1-minute rounded intervals (e.g. games he played ~24 mins, ~27 mins, ~30 mins). I also made custom standings using numberfire.com’s power rankings to rank opponents. From these custom standings, I broke up opponents into tiers: Top 5 (1-5), Top 10 (6-10), Top Half (11-16), Bottom Half (17-22), Bottom 10 (23-27), and Bottom 5 (28-32).

Dallas ranked 8th in this group, for what it’s worth. I also took game dates and calculated rest days and full off days (e.g. Monday night game and Wednesday night game is 1 off day on Tuesday but 2 rest days). I used rest days in this evaluation because there is a chance that it could impact how poorly/well an athlete performs. In combining all of these ideas together, I found some interesting insights related to the research question.

Findings

Days Rest

For context, between games, Heiskanen had the following rest day counts:

  • 1 day: 9 times
  • 2 days: 49 times
  • 3 days: 15 times
  • 4 days: 3 times
  • 8 days: 1 time
  • 10 days: 1 time

For Heiskanen’s performance on 1 day’s rest, he averaged 1.17 GameScore which is a great score for a star like him. In these games, he also sported a 57.7% Expected Goals For Percentage (xGF%) meaning that the attempts that Dallas made with him on the ice exceeded the value of the ones they gave up on defense while Heiskanen graced the ice surface. Heiskanen also surprisingly had the exact same average GameScore on 2 days of rest (1.17) with a 53.9% xGF%.

On 3 days of rest, he averaged a 0.82 game score but still had 54.1% xGF%. For the other days with smaller samples, there were inconsistent average GameScores and these other options were therefore omitted from consideration.

The general conclusion to be made from only rest day information is that Heiskanen is a star defenseman who averages a high GameScore no matter the number of rest days, and further research could help find out metrics that are more closely related to the hockey games themselves which could answer the research question rather than days between games.

Miro Heiskanen: Opponent Rank

When looking closely at Heiskanen’s GameScore sorted by opponent ranks, the averages and sums of GSc make a lot of sense. Heiskanen averaged 0.48 against the Top 5 teams, 0.63 against the Top 10, 1.02 against Top Half, 1.44 against the Bottom Half, 1.80 against the Bottom 10, and 1.53 against the Bottom 5 teams.

This scale is exactly what you could expect: Heiskanen has better outputs against weaker teams on average. Similarly, in multiple games against most teams, the sum of his GameScore in those opponent ranking groups follows the same trend of being better against worse teams.

Further to this, Heiskanen had more “Good” games (GameScore above 0.4) against worse teams in a similar trend to the previous ideas. Against the Top 5 teams, only 40% of Heiskanen’s GameScores were above 0.4 (“Good”) and only 50% above the “Bad” threshold at -0.2.

With this in mind, if we add the 3-minute intervals for ice time, it becomes clear that more minutes for Heiskanen shows more “Bad” GameScore results (below -0.2) but does provide a few “Good” results.

There is an overall inconsistent GameScore trend when Heiskanen plays closer to 27 or 30 minutes against various team rankings, and he seemingly has better results in lower minutes against each tier of opponents. If we remove the opponent filter, I wonder if we can best answer the research question by looking at 3-minute intervals and GameScores separated into three categories: “Good” (above 0.4) vs. “Average” (-0.2-0.4) vs. “Bad” (below -0.2).

Minute-Dependant Results

Miro Heiskanen has good results when playing closer to 24, 21, or 18 minutes. Specifically, 29/39 games when playing closer to 24 mins or less were “Good” games (74%). There is a vastly different scenario for games where Heiskanen played more minutes. When playing ~27 minutes, there are mixed results, with only 18/31 in the “Good” range (58%). Furthermore, when playing ~30 minutes against Top 5 teams, he had terrible results.

Despite all this, 6/9 games where Heiskanen played ~30 minutes were “Good”. This shows us that the one-season sample may show some outliers and may contradict the research question in certain areas, as Heiskanen can have an amazing game while playing 30 minutes and can also have a terrible game in the same deployment because of so many factors in a hockey game. An all-encompassing picture of Heiskanen’s GameScore likely does not show an opponent filter with a small sample. It likely comes down to GameScore and minute intervals

Miro Heiskanen: Average GameScore with Rounded Minutes and Binary GameScores

In splitting all 79 GameScores into positive (>0.00) and negative (<0.00), we see an exciting trend. Among all the games that were positive, Heiskanen had a higher average GameScore in all lower-minute games, with a gradually decreasing GameScore as the minute tier goes up.

On the negative side, Heiskanen has his worst average of negative games at the 30-minute tier (-2.51), but also had a poor average at the 24-minute tier (-1.09), so this may not be statistically significant to the study. However, one thing that is relevant is that there were no “Bad” games by Heiskanen in the 18-minute tier and his best “Bad” average was in the 21-minute tier at only -0.38. This further proves that Heiskanen is subject to poor performance in more minutes.

Miro Heiskanen: Verdict

Miro Heiskanen is an uber-talented hockey player. He had such a great season in 2022-2023 that there may be little to his game that the coaching staff would want to change going forward. However, because he is generally averaging worse statistical outputs in larger minutes, and is especially poor against top teams, it may be better to cut a few shifts from his games to conserve his energy to deploy him for greater on-ice success.

With so many other variables in play to understand Miro Heiskanen’s on-ice impact, it is difficult to justify a statistical conclusion in a 5v5 game, but athlete fatigue is real. Miro can play 30 minutes per game, and he may need to continue playing 30 minutes per game to help the Dallas Stars defensively as their #1 Defenseman.

However, there seems to be better consistency and worse lows in Miro Heiskanen’s play when he plays fewer minutes compared to the highly-volatile results I found using GameScore when the star plays closer to 30 minutes of ice time in a game. Overall, Miro Heiskanen should not be held back on the bench, but rather be rested and deployed tactically to bring his overall ice time down so he can be successful against all opponents and best help Dallas in their search for a Stanley Cup this year.

If you are interested in seeing the raw data that contributed to this article, here is the Microsoft Excel file I used.