S12 Stat-based NHL team and standing Projections

Discussion in 'Leaguegaming Hockey League (LGHL PSN)' started by memphisonfire, Oct 9, 2019.

  1. erinssilence

    erinssilence Well-Known Member

    Joined:
    Oct 3, 2016
    Messages:
    311
    Trophy Points:
    1,491
    Location:
    Dallas, Texas
    Ratings:
    +709 / 28 / -83
    Alright everyone so I'm bored and when I get bored I play around with stats and spreadsheets and all that fun nerd stuff. I'm sure some of you know that I pretty much have no life and spend a lot of time running projections on players in LG. Guys that were in the AHL a few seasons back will remember the kind of stuff I did on weekly basis for my Scrub Hunting thread and the AHL show with Sibs, Markz, Philly and Don later in the season. Others may remember the pre-season, post-bidding projections I ran for the NHL before S9 as well mainly because you were probably shitting on me as a nobody who had no business posting anything like that in the NHL forums. Anyway, I decided to run a projection based off everyones rosters currently and tried to project which DPs are making it up as best as possible as well.

    Before we get into the fun stuff that will light the comment section on fire as everyone tries to burn it down, I want to preface a few things. First, DPs are tricky to project mainly because most of them don't have enough data for me to run an accurate projection. In other words, I can't project off of one season. Second, any players that switched positions I went and tried to find as much data about those guys in those spots. For some they switched from F to D but had played D in previous seasons so I was able to at least have a better idea of that players value for that spot. As for guys that have never played the spot, or had less then 2 seasons of data for that position, I just decided to play it safe and give essentially the average rating to those players.

    The main stat I used to run these projections was my own PER (Player Efficiency Rating) stat that I've translated over from an NBA stat of the same. I'll spare all the details of the stat because it would bore most of you and also because its my secret formula too but just know that it incorporates both offensive and defensive stats as well as the advanced stats on the LG site to essentially boil everything down that what a players overall value is. I ran a full projection on every rostered skater (480 Fs and Ds) as well as any DPs with enough data as well. I didn't include Gs in this list mainly because over the past few seasons I've realized Goalies are the hardest to run accurate projections for because so much of their stat line is based on the team in front of them. Anyway, getting sidetracked again, I then took those projections and calculated out the PER for each player and added them all up to give a team Forward Rating, Defense Rating and a combined Overall Rating which is simply the Average of the Forward and Defense Ratings. There's also an average rating in there too that simply takes the average of all 15 skaters rather then separating by F or D.

    Now I know this thread is getting a bit long winded but when I previously ran this similar type projection back in S9 I was able to accurately project 75% of the top 4 teams in each division, 62.5% of playoff teams and had 34.4% of teams within 3 spots of their final points standing within the league. Keep in mind that these are simply projections based off players previous 2-3 seasons worth of data. So here are this seasons team strength projections, divisional strength, conference strength, and playoff projection. Hope you guys have fun with this.

    [​IMG]

     
    • Like x 26
    • Dislike x 1
    • Winner x 1
    • Informative x 1
    • Old x 1
    Offline
    memphisonfirex
  2. Hedmqn

    Hedmqn Well-Known Member

    Joined:
    Mar 26, 2016
    Messages:
    566
    Discord:
    xHedman-#8861
    Trophy Points:
    6,391
    Location:
    Canada
    Ratings:
    +477 / 21 / -87
    shut up
     
    • Winner Winner x 8
    • Dumb Dumb x 6
    Offline
    hedmxnn
  3. JClaus33

    JClaus33 Josh

    Joined:
    Feb 12, 2016
    Messages:
    1,316
    Discord:
    JClaus33#0960
    Trophy Points:
    48,921
    Location:
    NC
    Ratings:
    +4,032 / 68 / -188
    Good stuff!! Best d core in da league
     
    • Like Like x 1
    Offline
    JClaus33
  4. Vish x 91

    Vish x 91 Well-Known Member

    Joined:
    Dec 3, 2016
    Messages:
    253
    Trophy Points:
    2,871
    Location:
    punjab
    Ratings:
    +520 / 26 / -32
    @meka DVA meka hahahahahahahahahahahahahahah projected 7th in our division cause u blow
     
    • Funny Funny x 3
    • Like Like x 1
    • Winner Winner x 1
    Offline
    vish_b
  5. meka DVA meka

    meka DVA meka Bring Load Management To LG

    Joined:
    Feb 19, 2016
    Messages:
    1,474
    Trophy Points:
    5,791
    Location:
    USA
    Ratings:
    +3,475 / 92 / -197
    @memphisonfire got all these stats but can’t make the playoffs. Consider me skeptical of these stats. (Disclaimer: good job on doing media tho)
     
    • Funny Funny x 3
    Offline
    Nybarisny
  6. zdball

    zdball Well-Known Member

    Joined:
    Feb 12, 2016
    Messages:
    94
    Trophy Points:
    2,941
    Location:
    Rimouski
    Ratings:
    +101 / 2 / -13
    Could be better ...
     
    No Streaming Account
  7. l Doughty l

    l Doughty l Jeremy

    Joined:
    Mar 4, 2018
    Messages:
    328
    Discord:
    Doughty#9625
    Trophy Points:
    23,851
    Location:
    Toronto ➡️ Los Angeles, California
    Ratings:
    +456 / 27 / -57
    These canes though
     
    • Informative Informative x 1
    Offline
    Doughtyy_NHL
  8. l Doughty l

    l Doughty l Jeremy

    Joined:
    Mar 4, 2018
    Messages:
    328
    Discord:
    Doughty#9625
    Trophy Points:
    23,851
    Location:
    Toronto ➡️ Los Angeles, California
    Ratings:
    +456 / 27 / -57
    These canes though!
     
    • Informative Informative x 1
    Offline
    Doughtyy_NHL
  9. Cahspxr

    Cahspxr cowabunga

    Joined:
    Feb 9, 2017
    Messages:
    276
    Discord:
    Nylander#4206
    Trophy Points:
    3,641
    Location:
    Trois-Rivières
    Ratings:
    +316 / 21 / -26
    calgary way too high
     
    Offline
    vNylander
  10. JHughes86-

    JHughes86- Well-Known Member

    Joined:
    Jun 1, 2017
    Messages:
    161
    Discord:
    frann8686#0
    Trophy Points:
    1,141
    Location:
    frick park
    Ratings:
    +75 / 3 / -19
    this is bad
     
    No Streaming Account
  11. Spooer69

    Spooer69 Well-Known Member

    Joined:
    Nov 30, 2013
    Messages:
    207
    Discord:
    Spooer69#1668
    Trophy Points:
    6,841
    Ratings:
    +379 / 3 / -7
    This is really cool! The best part of statistical models like these are that placements can be explained by the numbers themselves.

    People can argue that some stats are weighed too heavily or historical stats may not be representative of future results, but the numbers are what they are.
     
    • Like Like x 2
    Offline
    spooer69
  12. MarcoGoes5H0LE

    MarcoGoes5H0LE Well-Known Member

    Joined:
    Sep 15, 2014
    Messages:
    130
    Discord:
    MarcoGoes5Hole#9335
    Trophy Points:
    11,181
    Location:
    Houston
    Ratings:
    +222 / 8 / -10
    UCLA Tings
     
    • Like Like x 3
    Offline
    MarcoGoes5Hole
  13. Zamne

    Zamne Well-Known Member

    Joined:
    Feb 20, 2016
    Messages:
    294
    Trophy Points:
    2,201
    Location:
    NS
    Ratings:
    +1,169 / 43 / -56
    This fuckin sucks.., Cheers
     
    No Streaming Account
  14. PatJoyce_

    PatJoyce_ Well-Known Member

    Joined:
    Oct 5, 2016
    Messages:
    107
    Trophy Points:
    5,751
    Location:
    Boston
    Ratings:
    +55 / 0 / -9
    Weighing down your teams D rating feelsbadman .
     
    Offline
    PatJoyce