Taking a look at game scores after 41 games

We’re past the halfway point of the 2017-18 Calgary Flames season. The Flames have two more points than they did at this time a year ago – hooray – but it’s probably prudent to take a more individual look at the club’s performance. We’ve been tracking player game scores for each and every game, and we’ve aggregated them to provide a snapshot of half-season performances.

Game score explained

Game score was conceptualized by Dom Luszczyszyn, currently of The Athletic, in an attempt to capture player performance in one comparable number. Here’s how he summed it up for Hockey Graphs:

It’s not a perfect stat – it’s meant to be a rough measure after-all – but I do think it works well for its intended purpose and is effective at what it does. It’s meant to answer “who had the best game” by adding proper perspective to a combination of a player’s total contributions and into an easily understood all-in-one stat.

For skaters, game score uses goals, primary assists, secondary assists, shots on goal, blocked shots, penalty differential, faceoffs, 5-on-5 corsi differential and 5-on-5 goal differential. For goaltenders, it uses saves and goals against. I’ve made one minor modification, reflecting goaltender assists and penalty differential, to reflect Mike Smith’s contributions in those areas.

The rundown

Here are the Flames’ average game scores through 41 games. If a player has played less than that, it’s been noted. Last season’s game scores are also included in parentheses, as well as changes from last season’s score.

Great (0.95 and up):

  • F Johnny Gaudreau: 1.153 (0.857, +0.296)
  • G Mike Smith: 1.057 [35 GP]
  • F Sean Monahan: 0.958 (0.662, +0.296)

Here’s the good news: take away the slow-ass starts they had to open 2016-17, and Gaudreau and Monahan are very, very good. Smith has also been precisely what Brad Treliving hoped he would be: a very strong starting goaltender.

Good (0.45 to 0.95):

  • F Matthew Tkachuk: 0.901 [39 GP] (0.786, +0.115)
  • D Mark Giordano: 0.879 (0.691, +0.188)
  • D Dougie Hamilton: 0.812 (0.810, +0.002)
  • G David Rittich: 0.800 [5 GP]
  • F Micheal Ferland: 0.729 [40 GP] (0.312, +0.417)
  • F Mikael Backlund: 0.681 (0.777, -0.096)
  • F Michael Frolik: 0.665 [37 GP] (0.688, -0.023)
  • F Garnet Hathaway: 0.544 [18 GP] (0.139, +0.405)
  • D Rasmus Andersson: 0.500 [1 GP]
  • F Mark Jankowski: 0.490 [33 GP]

Nobody on the Flames has improved, on average, more than Ferland. He’s made the leap from being a complementary player to perhaps being considered part of the team’s core. Tkachuk has improved a bit, while Backlund and Frolik have dipped by roughly the same amount – this seems to be some of the “noise” in the numbers, the trio overall is about as successful as they were a year ago. Giordano has made a big leap, Rittich has established himself as a steady hand in the backup role, while Hathaway’s big improvement and Jankowski’s arrival have really given the Flames a lot more depth in this group.

Fine (0.15 to 0.45):

  • F Sam Bennett: 0.415 (0.280, +0.135)
  • F Jaromir Jagr: 0.407 [22 GP]
  • D T.J. Brodie: 0.399 (0.400, -0.001)
  • F Kris Versteeg: 0.353 [22 GP] (0.488, -0.135)
  • D Brett Kulak: 0.249 [31 GP]
  • D Travis Hamonic: 0.239 [37 GP]
  • F Andrew Mangiapane: 0.228 [3 GP]
  • D Michael Stone: 0.216 (0.297, -0.081)
  • F Troy Brouwer: 0.202 (0.166, +0.036)

Let me pose a question: if the Flames spent three high draft picks to bring in a more suitable partner for Brodie, and his performance is essentially unchanged from a year ago, who should be considered the disappointment? In other news, Versteeg’s bad numbers are potentially from playing hurt (and on the fourth line), while Bennett’s really improved after being rescued from his third line center purgatory. (I fully expect him to nudge into the next grouping within the next few weeks.)

Bad (-0.15 to 0.15):

  • F Curtis Lazar: 0.126 [27 GP] (0.445, -0.319)
  • F Matt Stajan: 0.123 [31 GP] (0.242, -0.119)
  • D Matt Bartkowski: -0.023 [13 GP] (-0.053, -0.030)
  • F Freddie Hamilton: -0.034 [8 GP] (.109, -0.143)
  • F Tanner Glass: -0.078 [7 GP]

Lazar’s played more than four games this season and his numbers are bad. Stajan is what he is. After them, we get into the muck with fringe bodies (and below that).

Awful (below -0.15):

  • G Eddie Lack: -0.950 [4 GP]
  • G Jon Gillies: -1.350 [1 GP]

Gillies’ number is based on a partial game that was way out of hand. Lack’s? Not so much.

  • RealMcHockeyReturns

    Get a Franson for free on waivers then adjust our D and RW with a trade. How about Brodie and Brouwer to Montreal great for Gallagher and a 2nd

  • Flint

    I worked on this a bit last year, but I wanted to compare game score (with its flaws) to salary to find out who were the best value skaters for the team. This is only the halfway point but there are some consistencies from last year.

    The top 5 best value skaters on the Flames by game score and cap hit are:
    1. Garnett Hathaway 21.4
    2. Matthew Tkachuk 19.8 (prev. #6)
    3. Michael Ferland 17.1 (prev. #5)
    4. Mark Jankowski 12.1
    5. Brett Kulak 11.9 (prev. #1)

    AVG or expected value is ~6.5 … so it’s good that Gaudreau is 7.0 and Monahan is a 6.2.

    The worst of the regular players (min. 10GP) excluding Rittich who is quite good value @ 7.6.
    1. Matt Stajan 1.2
    2. Troy Brouwer 1.8
    3. Travis Hamonic 2.3
    4. Michael Stone 2.5
    5. TJ Brodie 3.5

    • Flint

      I should add that scores are affected by number of games played (if the guy doesn’t play it’s worse value right?) not that it’s a perfect system. But if Jankowski and Kulak played all 41 games they would switch places and be closer to Ferland. If Hamonic played all 41 he would jump Stone and be closer to Brodie. Stajan would still be last.

    • Flint

      Ok, one last thing cause I’m geeking out a bit too much here. But Tkachuk is currently providing the expected value by this method of a 5.4AAV player. Hathaway 1.51AAV. Stajan on the otherhand a 576k player…. not even league min. I think it’s especially interesting when thinking of Tkachuk and his next contract.

      • Chucky

        Waive Bartkowski and bring up Wotherspoon. Anderson will benefit from the time in Stockton, Wotherspoon is at a now or never point. Either they want him or get his traded value up because after a great year in the AHL he is gone next year with no return.

      • The Beej


        That wouldnt be very good for Andersons development. Being 6/7 here instead of 1 playing 20+min a night in the minors is not a good development strategy for a 2nd year pro prospect. Maybe for the playoffs.

        But when Anderson is brought up for good they will make sure he is in a position to succeed and he will have an actual spot.

    • ThisBigMouthIsRight

      You can’t… nor can you correlate line mates into the equation. Look what Bennett has been able to accomplish away from his usage last year and unfortunately that is the Lazar situation all over, except Lazar is on the fourth line instead of third line. These states don’t really include opportunity(or lack of opportunity) and the “roll” the player is being used for or asked to do with there given TOI.

  • The Fall

    Player Game Score = (0.75 * G) + (0.7 * A1) + (0.55 * A2) + (0.075 * SOG) + (0.05 * BLK) + (0.15 * PD) – (0.15 * PT) + (0.01 * FOW) – (0.01 * FOL) + (0.05 * CF) – (0.05 * CA) + (0.15 * GF) – (0.15* GA)

    God, I hate this stat.

    • Chucky

      It seems to agree with the eye test so it tells us exactly what we want to know. How can anybody dislike a stat that keeps the calculators working away for hours and contributes nothing to the conversation. The one I hate is Corsi because it give misleading information and some of the confused use it like it means something.

      • freethe flames

        The problem for me with Corsi and some of the other stats is that some people use them as if they can predict something; none of them can. After a number of good corsi games and events a team can still have a stinker. Witness the Montreal game.

  • dontcryWOLF

    Game score fails to account for the health of each players prostate. It even fails to catelouge the number of times each player touched the puck with their leg or foot.

    Clearly worthless information.

    Please write an article about how Glen gulutzun uses lots of gel. Or about how Gaudreau now has a Mohawk as a declaration of war against Calgary and it’s fans.

  • Jumping Jack Flash

    Small sample size for Mangi who could easily get a couple of points and move up to the Good category. Mangi is smaller than I thought but has the Crosby stockiness which will serve him well. He does not seem to be easy to knock off the puck and is crafty with the puck. Maybe they can try him in Versteeg’s role on the PP to compliment Johnny.

  • Sven

    Those top numbers are astounding-
    noting the obvious difficulty in achieving consistency – if you extrapolate those numbers over a full season
    some things stand out

    A) Gaudreau is an elite forward

    if he could maintain these numbers over a full season, his score of 1.153 would be behind only the 2011-2016 five season averages of Crosby (1.31) and Malkin (1.24) …
    putting him ahead of the average seasons of Seguin (1.15) , Kane (1.12), Kopitar (1.10) and Ovechkin (1.08)
    B) Despite their warts – Giordano and Hamilton are rock stars

    Giordanos score of 0.88 would be the fourth best score in the league – trailing only Karlsson (1.07), Letang (1.00) and Burns (0.92)
    and Hamiltons score of 0.812 would be seventh in the league – ahead of Muzzin (0.79) , Keith (0.78) and Lidstrom (0.77).

  • Flint

    It’s getting slammed, but personally I kinda like game score because it’s like the Ironic ‘advanced stat.’

    It is so full of bais, that it effectively simulates the mental arithmetic we all do in our heads when assessing players or teams. It just isn’t SO biased. but you could do really fun ones like a Brian Burke game score and add 0.10xheight in cm + 0.20xweight inkg + 0.50Xpunches thrown/60.

    The thing is, it aligns with ‘eye test’ because the eye test is biased and so is game score. Does that make it horrible? No. Does is make it tremendously useful? Probably not, but that also depends on the question you’re asking.