What do “Injured Man-Games Lost”, “Did Not Dress”, “Did Not Play”, “Suspended” and “In Minors” represent? How is it calculated?
This NHL injury data is manually compiled and tracked by me (as of 2010-2011 season).
- Injured Man-Games Lost (MGL) = player missed game due to injury, injury recovery, or illness
- Did Not Play (DNP) = player was a healthy scratch for the game, or was the backup goalie, and did not play
- Did Not Dress (DND) = player was unavailable for personal reasons (family event, immigration issue, personal event)
- Suspended (SUS) = player missed game since he was suspended by the team and/or the league, with or without pay
- Minors (MIN) = player had been assigned to the team’s minor league affiliate and did not play
What are Points Shares (PS)? What is the Lost-ps metric?
Point Shares are an estimate of the number of team points in the standings contributed by each player. I use the Point Shares values posted at hockey-reference.com.
The Lost-ps metric attempts to determine the point shares lost due to a player’s injury. It is the difference between the expected point shares generated by the player and the actual point shares they have generated.
Lost-ps = Point Shares Expected – Point Shares Actual
Lost-ps = (PlayerPointShares/PlayerGamesPlayed) x (PlayerGamesPlayed + PlayerGamesInjured) – PlayerPointShares
What is IIT? How is it calculated for the NHL?
Injury Impact To Team (IIT) is a single-value metric that attempts to quantify the impact of a player not playing for their team due to injury. It weights injured NHL players and their time missed based on how much playing time they would have had if they were healthy (or on relative Corsi (CFrel%) for the IIT-cfrel metric). Coaches (good ones) play ‘better’ players more, and not vice versa.
IIT-skater and IIT-goalie utilize a skater’s average time on ice (ATOI) or minutes played for goalies, the number of games missed due to injury, the number of games played by their team, and the number of games that the player has played in (necessary in calculations due to players who aren’t on the team’s roster or IR at all times i.e. AHL call-ups, free agent signings).
A higher IIT number equals a greater impact of injured players to the team. A high IIT value for a player reflects a high average of playing time, a high number of games missed due to injury, or both combined. The exact number is not as important as the basic range, relative to other players. This is the case with most metrics in the analytics of sport performance.
For players who haven’t yet played in a current season, I use their previous season ATOI to calculate IIT-skater. ATOI for players is assumed to be fairly stable across seasons, unless they’ve changed teams.
For IIT-cfrel, I do not use previous season CFrel% for players who haven’t yet played in the current season. It is a more volatile metric. So, no IIT-CFrel calculation for those players.
Using the number of games a player has played in for IIT purposely devalues players who miss more than half the season, since their teams are better able to adapt to their loss as the season progresses beyond a 41 game absence.
IIT-skater = (PlayerGamesPlayed x ATOI x PlayerGamesInjured) / TeamGamesPlayed
IIT-goalie = (PlayerGamesPlayed x MinutesPlayed x PlayerGamesInjured) / TeamGamesPlayed
IIT-cfrel = (PlayerGamesPlayed x CFrel% x PlayerGamesInjured) / TeamGamesPlayed
At the team level, the IIT metrics represents the sum of all the IIT-skater, -goalie, or -cfrel values for injured players on the team.
What is CMIP? How is it calculated?
CMIP (Cumulative Minutes of Injured Players) sums the amount of time a player
would have played if they weren’t injured. It represents the amount of playing time, in minutes, that a team has lost due to a player being injured. At the team level, CMIP represents the sum of all the time missed by injured players on the team.
To do this, their average time on ice in the games they’ve played so far in the season is simply multiplied by the number of games they’ve missed due to injury. This metric was pioneered by Rob Vollman of
HockeyAbstract.com.
CMIP = PlayerATOI x PlayerGamesInjured