MASS's Player Ratings (Week 1, AHL)

Discussion in 'Leaguegaming American Hockey League (LGAHL PSN)' started by xMAssStAtiC-, Oct 18, 2018.

  1. Kingboivin31

    Kingboivin31 Frappeur

    Joined:
    Jan 9, 2018
    Messages:
    24
    Trophy Points:
    3,821
    Location:
    Montreal
    Ratings:
    +14 / 4 / -5
    Were am i elite rw here
     
    • Winner Winner x 1
    No Streaming Account
  2. kgrilli

    kgrilli Well-Known Member

    Joined:
    Jan 27, 2017
    Messages:
    58
    Trophy Points:
    1,041
    Location:
    Ottawa
    Ratings:
    +18 / 1 / -7
    I've considered that premise in weighting, but found it flawed on a few levels. First it weighs only on the teams current record against the players current season summarized stats. It also weighs the success (or failure) by virtue of the teams total record against said players stats (instead of personal record). Finally the team record isn't always indicative of the strength of team. Often players are traded, called up, or quit. This should affect the strength of team and/or opponent.

    The root of the issue arises from where you base the meta data source. If the data was more accessible at the game stat level it would allow for situational metric valuations. Those metrics would be better served in measuring within normal distributions and weighted algorithms to provide a clearer definition of a players impact.
     
    Offline
    kgrilli
  3. cjdtheprofessor

    cjdtheprofessor Well-Known Member

    Joined:
    Jan 13, 2016
    Messages:
    48
    Trophy Points:
    1,841
    Location:
    City
    Ratings:
    +91 / 1 / -2
    My opinion is that the weight hasn't really been adjusted too much for 19 being a new game, specifically how viable hitting actually is now. We're going to see a sharp increase in hits this season as opposed to the '18 seasons and I think placing a 4 on the weight for defensemen is fairly low. For some perspective, you have blocked shots (which EA has never been able to positively track at all whatsoever) with a weight of 5. Granted I wouldn't make it higher than 10, I just think it's actually a reasonable tool in a defenseman's arsenal when used properly now unlike in the past.

    Also, what does the rating accuracy percentage column mean?

    (edit1)
    Oh, and I'm inclined to suggest giveaways for a defenseman being worse than -16. Honestly that's the biggest thing I look at with D stats because every giveaway from a defenseman is magnified. Normally they result in a direct scoring chance for the other team and they seem to track in a reasonably accurate manner. Again, my opinion. I definitely appreciate the time you put into the spreadsheet.
     
    Last edited: Oct 18, 2018
    • Like Like x 1
    Offline
    Cjdtheprofessor
  4. ClumsyTurtle13

    ClumsyTurtle13 xMAssStAtiC-

    Joined:
    Jun 28, 2017
    Messages:
    234
    Discord:
    xMAssStAtiC-#1607
    Trophy Points:
    1,451
    Location:
    USA
    Ratings:
    +370 / 2 / -8
    I hadn't quite thought of that fully about adjusting for 19 but that is a great point. I just have to wait until maybe week 4 when I see more numbers on how much of a change everything needs. But I will definitely tweak each week as we go on. And the accuracy is there to state that if they play a full season by the end their rating should be accurate closer to 90%. If they are an ECU player they may play a quarter of the season so their accuracy should be closer to 25% meaning there is a good possibility they could be much better or worse. Think of it as the potential accuracy that is in Franchise mode on the game. I was mimicking that in effort to create a potential scout report for NHL Owners.
     
    • Informative Informative x 1
    Offline
    Xmassstatic
  5. ClumsyTurtle13

    ClumsyTurtle13 xMAssStAtiC-

    Joined:
    Jun 28, 2017
    Messages:
    234
    Discord:
    xMAssStAtiC-#1607
    Trophy Points:
    1,451
    Location:
    USA
    Ratings:
    +370 / 2 / -8
    Yeah, I want to make the giveaways much more of a penalty to for defensemen. I think that's a big stat and I've been wanting to change it.
     
    Offline
    Xmassstatic
  6. l Bossy l x 22

    l Bossy l x 22 Well-Known Member

    Joined:
    Nov 11, 2015
    Messages:
    1,356
    Discord:
    Bossy#9297
    Trophy Points:
    8,431
    Location:
    Long Island
    Ratings:
    +3,325 / 126 / -162
    How is it that wins aren't factored in for any skaters? Face-offs should also be looked at for centers.
     
    • Useful Useful x 1
    Offline
    bossy_22_
  7. ClumsyTurtle13

    ClumsyTurtle13 xMAssStAtiC-

    Joined:
    Jun 28, 2017
    Messages:
    234
    Discord:
    xMAssStAtiC-#1607
    Trophy Points:
    1,451
    Location:
    USA
    Ratings:
    +370 / 2 / -8
    I figured wins were more of a team stat and I dont want good players who are on bad teams to have horrible ratings because they arent winning. And vice versa, if I played on a line with JClause, I'd look like an elite forward because I wouldn't lose but in reality I'd probly not be able to play forward for the CHL. And faceoffs do need to be incorporated. I've been lazy about adding it in haha. Maybe next weeks post I'll put it in.
     
    • Like Like x 1
    Offline
    Xmassstatic
  8. l Bossy l x 22

    l Bossy l x 22 Well-Known Member

    Joined:
    Nov 11, 2015
    Messages:
    1,356
    Discord:
    Bossy#9297
    Trophy Points:
    8,431
    Location:
    Long Island
    Ratings:
    +3,325 / 126 / -162
    Why would wins count for goalies then? There's absolutely nothing they can do if the team in front of them blows.
     
    • Informative Informative x 1
    Offline
    bossy_22_
  9. ClumsyTurtle13

    ClumsyTurtle13 xMAssStAtiC-

    Joined:
    Jun 28, 2017
    Messages:
    234
    Discord:
    xMAssStAtiC-#1607
    Trophy Points:
    1,451
    Location:
    USA
    Ratings:
    +370 / 2 / -8
    True, that is something I was debating myself. I saw it as a goalie can be really good but when they give up a goal is it in the crucial moments that could cost the team or is it when they are up by 3 and it doesn't matter. But you are right, If I'm counting it for them I should probly reconsider it for the skaters as well.
     
    • Like Like x 1
    Offline
    Xmassstatic
  10. kgrilli

    kgrilli Well-Known Member

    Joined:
    Jan 27, 2017
    Messages:
    58
    Trophy Points:
    1,041
    Location:
    Ottawa
    Ratings:
    +18 / 1 / -7
    I alluded earlier to the use of normal distributions. Each stat could be valued against the variable average of the leagues current distribution. This value would be leveraged into the weighted value of the stat category of the algorithm.

    These distributions could be compared against previous years to see trends in statistics within the game (or patches haha).
     
    Offline
    kgrilli
  11. ClumsyTurtle13

    ClumsyTurtle13 xMAssStAtiC-

    Joined:
    Jun 28, 2017
    Messages:
    234
    Discord:
    xMAssStAtiC-#1607
    Trophy Points:
    1,451
    Location:
    USA
    Ratings:
    +370 / 2 / -8
    Yeah, very true haha. I'm looking into a way for me to compile multiple seasons into one data set without me manually doing anything but it's going to take some time to do the coding for that.
     
    Offline
    Xmassstatic
  12. bobbatta

    bobbatta Big Dick Bobbatta

    Joined:
    Dec 23, 2015
    Messages:
    309
    Discord:
    Bobbatta#7682
    Trophy Points:
    2,841
    Location:
    Ontario
    Ratings:
    +387 / 12 / -33
    Goals against average isn't much of a goalie stat either, just save % in my opinion.
     
    • Like Like x 2
    No Streaming Account
  13. Markz51

    Markz51 Don't Hassle Me. I'm Local.

    Joined:
    Nov 17, 2015
    Messages:
    428
    Discord:
    markz_51#0
    Trophy Points:
    3,641
    Location:
    New Jersey
    Ratings:
    +702 / 11 / -21
    [​IMG]
     
    • Like Like x 2
    • Funny Funny x 1
    • Winner Winner x 1
    Offline
    Markz_51
  14. TeeSibs

    TeeSibs Well-Known Member

    Joined:
    Jan 2, 2016
    Messages:
    751
    Discord:
    TSibs88#2180
    Trophy Points:
    3,111
    Location:
    PA
    Ratings:
    +1,495 / 17 / -65
    We can debate if this value system is fair but people are trying to make it interesting for the league and are putting in the time to do it.
    I always want to promote the desire of people like Mass and Memphis to do more for LG. No one has created anything perfect but let's try and help them refine their work with constructive ideas.
     
    • Like Like x 3
    • Winner Winner x 2
    • Useful Useful x 1
    Offline
    tsibs88
  15. OcEaNtOmBs55

    OcEaNtOmBs55 Degravees55

    Joined:
    Sep 29, 2016
    Messages:
    92
    Trophy Points:
    841
    Location:
    Houston tx
    Ratings:
    +68 / 20 / -14
    Good read.
    Thank you for your time and efforts Mass.

    #ShutUpShitters
     
    • Like Like x 2
    Offline
    Oceantombs55
  16. xMelom

    xMelom Well-Known Member

    Joined:
    Jun 5, 2015
    Messages:
    109
    Discord:
    Melom#7026
    Trophy Points:
    3,141
    Ratings:
    +88 / 7 / -6
    Off my comment he did end up refining his way of rating goalies and plus u must remember I'm toxic
     
    • Like Like x 2
    Offline
    xmelom
  17. TeeSibs

    TeeSibs Well-Known Member

    Joined:
    Jan 2, 2016
    Messages:
    751
    Discord:
    TSibs88#2180
    Trophy Points:
    3,111
    Location:
    PA
    Ratings:
    +1,495 / 17 / -65
    That is true. You are toxic ;)
     
    Offline
    tsibs88