Midpoint player ratings

Discussion in 'MLS: News & Analysis' started by Jough, Jul 23, 2012.

  1. Jough

    Jough Member+

    Jul 30, 2007
    Kansas City
    Club:
    Sporting Kansas City
    Nat'l Team:
    United States
    So, because I'm a giant spreadsheet nerd, I took a bit of time and crtl+v -> crtl+c'ed the player data for each team into a big ole spreadsheet to crunch some numbers. More specifically, I used the player scores from the MLS fantasy game to come up with some sort of arbitrary player ratings, based on player score and minutes played (in actuality time played barely mattered, but I didn't know that until the end). The average player (who has seen the field) has played 837 minutes this year, by the way.

    It was sort of a quick and easy way to "rate" every player in the league but I thought some of the results were interesting, if not perhaps a little obvious. Ratings are more or less average player score across the league = 5 and each individual score = 5 + (player's score - average score)/(standard deviation of all players scores).

    The easiest way to share it is just to throw up the Google Doc, so here it is. I've thought about adding full stats (i.e. goals and such) as well. I might do that in the near future.

    MLS Player Ratings

    For brevity, I'll just put up the Top 25.

    [​IMG]
    The only thing that kind of surprised me was how many keepers made the Top 25. Maybe it shouldn't be a surpise though, since one could argue it's the most important role. If one wanted to.

    I did check out each team's average rating as well, and, for the most part, it mirrored the league table, with the weird exception of Dallas and, kind of now that they're turning around, LA.

    [​IMG]
    So, yeah. That's it really. I was just a little pleasently surprised that it wasn't too hard to manipulate a large chunk of player data into some kind of meaningful numbers that some people might find intersting. Feel free to ask questions or tear it all apart as warrented.
     
  2. tomásbernal

    tomásbernal Member+

    Sep 4, 2007
    Club:
    Portland Timbers
    Nat'l Team:
    United States
    Very cool. Thanks for doing this.
     
  3. footballfreak

    footballfreak Member

    Jun 30, 2003
    Vancouver, BC, Canad
    Club:
    Vancouver Whitecaps
    Nat'l Team:
    Canada
    These stats reinforce what we already know in Vancouver - YP Lee should be at the All-Star game as one of the best defenders in the league. I'm happy he's not there and gets an opportunity to rest, but still a little perplexed at the exclusion.
     
  4. RedBullFootball

    Apr 7, 2008
    Is there really any way in the world these numbers could be actually meaningful?

    A fun little exercise maybe, BUT meaningful? Come on.
     
  5. Jough

    Jough Member+

    Jul 30, 2007
    Kansas City
    Club:
    Sporting Kansas City
    Nat'l Team:
    United States
    I meant more that the individual player ratings don't seem too far off from what my own subjective ratings would probably be, but, no, there's really nothing too deep going on here. That wasn't what I intended to convey. Just fun with numbers at the end of the day.
     
  6. Jough

    Jough Member+

    Jul 30, 2007
    Kansas City
    Club:
    Sporting Kansas City
    Nat'l Team:
    United States
    Annnnd of course I can't edit a post from a bit ago.

    I just wanted to add that, if given nothing but the player data and what team that person plays for, you could concievably guess at what team is doing well and what team isn't. Team rating correlates pretty well with season point total so far.

    correlation (r squared) for team points ~ team average rating = 0.84

    Of course the followup question would be do good players make a team good, or does a good team make players good? Would Wondo be playing this well on Philadelphia? Could Fabian Castillo be a stellar player with New York? etc. etc.

    Maybe I'm overthinking it, I haven't had coffee yet this morning.
     
  7. JazzyJ

    JazzyJ BigSoccer Supporter

    Jun 25, 2003
    So what is the player score from the MLS fantasy game based on? Seems like kind of an important question if we want to evaluate how meaningful the numbers are.
     
  8. Jough

    Jough Member+

    Jul 30, 2007
    Kansas City
    Club:
    Sporting Kansas City
    Nat'l Team:
    United States
    It's a very important question and it's where the arbitrary 'wtf why is that that way' aspect comes in. In no way do I think this is the correct, perfect, absolute way players should be ranked. It was just the easiest way to do it since all the data is readily available in table form.Here's the breakdown.

    +1 point for playing in a game
    +2 points for playing at least 60 minutes in a game (includes playing points mentioned above)
    +6 points for each goal scored by your goalkeeper or defenders
    +5 points for each goal scored by your midfielders or forwards
    +3 points for each assist
    +4 points for your goalkeeper conceding 0 goals (must also play at least 60 minutes)
    +4 points for your defenders conceding 0 goals (must also play at least 60 minutes)
    +1 point for your midfielders conceding 0 goals (must also play at least 60 minutes)
    -2 points or every penalty miss
    +5 points for every penalty save
    -1 points for every 2 goals conceded by your goalkeeper or defenders
    -1 points for every yellow card
    -3 points for every red card (includes any yellow card points)
    +1 point for every 3 saves
    +1 point for every 3 crosses (A cross is defined where the ball is played to their own player in the penalty area)
    +1 point for every 3 key passes (A key pass is one that leads to a shot at goal)
    +1 point for every big chance created (A big chance is one where the analyst determines the player should score)
    -1 point for every big chance missed
    +1 point for every 6 clearances, blocks or interceptions
    +1 point for every 6 recovered balls
    -1 point for every error leading to a goal
     
  9. JazzyJ

    JazzyJ BigSoccer Supporter

    Jun 25, 2003
    OK, cool, so there's a little more to it than just the obvious stuff like goals and assists. If it was just based on the obvious stuff that shows up in the score sheet, than calculating the cumulative numbers and comparing to a team's place in the standings would be kind of a circular exercise. The teams with the most goals scored and the least conceded, etc. would naturally have the highest cumulative scores, etc.
     

Share This Page