SuperCoach top 50

Discussion in 'AFL' started by SuperCoachSuperStar, Oct 16, 2011.

  1. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    Criteria should include:
    Weightings for forwards, midfielders and rucks.
    Scoring when it really matters eg. Rounds 21, 22, 23, 24
    Time rookies started playing
    Bonuses for DPP
    Time taken to increase in price
    Premiums should cop penalties for missing games.
     
  2. pbabs

    pbabs New Member

    Joined:
    Jan 3, 2016
    Messages:
    13
    Likes Received:
    0
    Nomination: Gary Ablett
    Overall Points:2676, ave 127.4
    Games: 21/22 (one early injury score of 19)
    Last 4 Rounds: played all games ave 143.5
    Optimal Strategy: Lowered to $526k after rd 7, about the time to cash in the first cow. Averaged 135 from this point.

    MVP Ranking: Top 5
     
  3. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    Fyfe has got to be right up there was priced at $353,100 at the start of the year, averaged 108, and DPP. The only thing going against him is he missed the Grand Final (Round 24).
     
  4. Fez

    Fez Moderator

    Joined:
    Jan 3, 2016
    Messages:
    650
    Likes Received:
    8
    Deduct Abletts 19 when he was injured , is that enough for number 1?
     
  5. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    It would help him a bit that he scored his 19 during round 16 (No league games).

    @Fez

    It probably would be enough to get him the number 1 ranking, although he would've missed the three games (assuming he didn't play the game he scored 19).

    He averaged 135 from round 8 (including the 19)
    Was priced about $526,000 prior to round 10
    Averaged 143.5 in finals (Rounds 21-24)
    He still scored the 6th most points despite missing 2 games
     
  6. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    This is a rough draft of the MVP formula (without all the criteria):

    MVP = Total points+70*(22 - games played)-((number of scores under 65)*70-(sum of all scores under 65))
    *(fwd=1.03,mid=1,ruc=1.01,def=1.04)

    This formula is missing things like:
    Price
    Rookie rules
    Point weighting
    DPP bonus
    and other criterea

    Any thoughts
     
  7. Lucas

    Lucas Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    1,974
    Likes Received:
    121
    Happy with my top 50 haha

    Fair enough SCSS if you want to go for the league wins. A bit of a different criteria for the overall. I think my top 50 pretty much stands up, and took me about 30 mins to do in my lunch break.

    I reckon my top 50 was better than Mike Sheahan's!

    Personally I rate PPG but I know others rate overall. I guess it should work itself out in the wash but maybe that's why I have rated players like Rioli when I should have preferred consistent scorers like Zaharakis.

    PPG works during the season because of match differentials caused by byes, but overall works at the end of a season, as it's then injuries and general inconsistencies that stuff people up in the end.
     
  8. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    I wasn't sure whether to do it by overall or league games. I think overall is probably more important so I will get rid of the weighting system.
     
  9. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    I think overall player points is much more valuable than PPG because if you have Sandilands a lot of the time he was sitting on the bench, so you would only be getting the ~70 points from the rookie replacement.
     
  10. stkildathunda

    stkildathunda Moderator

    Joined:
    Jan 3, 2016
    Messages:
    1,555
    Likes Received:
    0
    @ Lucas

    Wouldnt Ben McEvoy get in over few other ruckman in your list? Points increase of 37 this year alone? Think he was in top 5 in the league for that.
     
  11. JPK

    JPK Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    4,785
    Likes Received:
    3,313
    If there is anyone who's really good at permutations and combinations, and algorithms and computer code, I'd like to see a top 57 (33 starters plus 24 trades) that, SELECTED AT THE RIGHT TIME, would have gotten you the maximum possible points haul!

    I know that this is a HUGE amount of work, but think about it - for the overall, its all about selecting the best possible starting team, trading players in and out at the right time, avoiding donuts due to injuries and byes, and generally scoring the maximum possible points each week.

    So, if anyone can put an algorithm together that could select the best possible starting team, and then make all the right trades at all the right times, the 57 players that this spits out would be a definitive top players list, as (assuming its done properly) you couldn't possibly score any more than this!!!

    Anyone game?
     
  12. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    That is a terrific idea, earlier in this season i wan't thinking of doing that but i decided it would take too much work.
     
  13. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    It would very hard to do as there are millions of possible combinations.
     
  14. shanaz

    shanaz New Member

    Joined:
    Jan 3, 2016
    Messages:
    166
    Likes Received:
    0
    @JPK, thats near impossible mate. I dont think it can be put down to statistics. Theres a whole lot of luck going around too, so i dont think its possible to make an algorithm for supercoach
     
  15. JPK

    JPK Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    4,785
    Likes Received:
    3,313
    @shanaz - I wouldn't have thought its impossible, just very hard. All the stats exist for points scored per round, price changes, rounds missed, etc, so the only thing required is the algorithm that can work out the optimal starting team and the optimal trades to maximise points each week. There are probably a few million possible combinations, so some serious computing power would be required, and some serious time and committment.

    Now that I re-read your comment, were you thinking I was talking about 2012? I'm looking for the best possible team for 2011 - a "perfect season" if you will.
     
  16. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    There would be a lot more possible combinations than a million assuming there was 792 players in 2011. Just say there were 230 possible defender, mid, and foward selection and 100 for rucks you would get.

    230^3*229^3*228^3*227^3*226^3*225^3*224^3*223^3*222^3*100*99*98*97 possible intial team selections. I suppose you could exclude certain players or condition it but there is still a lot just for initial team selection alone.
     
  17. SuperCoachSuperStar

    SuperCoachSuperStar New Member

    Joined:
    Jan 3, 2016
    Messages:
    418
    Likes Received:
    0
    I think we should have a too serious competition for who can make the best team for the season.
     

Share This Page