218 Open Daily games
1 Open Realtime game
    Pages:   1   (1 in total)
  1. #1 / 7
    Pop. 1, Est. 1981 Alpha
    Rank
    Brigadier General
    Rank Posn
    #60
    Join Date
    Dec 09
    Location
    Posts
    991

    So, for a while I have been trying to figure out what should be done to fix the problem with the Swiss System Tournament. If there is no interest in fixing the problem then please ignore this.

    Here is the current issue:

    With more that 2 players per game, the tournament does not always produce a clear winner.  Several players have the same top record, but only one of them wins.
    The two underlying issues are that Swiss-Style tournament were developed with 1on1 games in mind and secondly that the current scoring system cannot differentiate players after 2 rounds.

    One way around this is to change the Swiss-Style tournament into an Elimination-Style tournament by making sure the number of players in the tournament is an exponent power of the number of players per game.  This has many disadvantages (for one, once you lose a game you cannot win the tournament, but you have to play out more games).  This was okay before elimination tournaments were around, but now this is unnecessary.

    Before, I argued that allowing the number of rounds to be undecided and let the tournament continue until a clear winner can be determined (one player with a better record than anyone else).
    Issues:  Might not ever end (not likely, but could last a while), players with bad records might have to play even more games than before and still not be able to win.

    Here is my new suggestion for a fix.
    Let the first 3 rounds of the Swiss-System tournament commence as usual.
    If there is a clear winner, then we are done, tournament ends
    Otherwise, keep the players with a 2-1 record and eliminate the players with an 0-3 record.  For the players which are 1-2, keep the player with the highest score, where the score of a player is determined by:
      player's score (PS) = (# of opponents per game)*(# of wins) + # of wins of opponents beaten.

    Using this PS, the players will usually be differentiated enough to make good decisions.  Here is an example:
    15 players, 3 players per game
    After three rounds (A-E have 2 wins, F-J have 1 win)
    Player  PS
    A          11
    B           7
    C          8
    D          6
    E          6
    F          6
    G         3
    H         2
    I          2
    J          2
    (These came from a randomly generated tournament played out).

    Keep A-F and let ABC and DEF play a game at this point.
    If no winner, keep only the top three and let them play until one has the best record.
    In general, after three games, keep only players who have won and only enough to fill in games for the 2-1 players.  Then have one less game per round dropping off the lowest players.

    This scoring system will take into account the "strength of schedule" as was suggested by others previously.

    Round 1 Winners Round 2 Winner Round 3 Winner
    1 3 3 13 13 1
    2
    4
    8
    3
    13
    1






    4 4 8 8 7 2
    5
    11
    3
    6
    12
    2






    7 8 10 7 4 11
    8
    7
    11
    9
    14
    15






    10 11 9 2 6 5
    11
    15
    14
    12
    2
    5






    13 13 6 1 9 9
    14
    5
    10
    15
    1
    12

    Never Start Vast Projects With Half Vast Ideas.

  2. #2 / 7
    Pop. 1, Est. 1981 Alpha
    Rank
    Brigadier General
    Rank Posn
    #60
    Join Date
    Dec 09
    Location
    Posts
    991

    cool, I broke the forum already.

    Never Start Vast Projects With Half Vast Ideas.

  3. #3 / 7
    Standard Member RiskyBack
    Rank
    Colonel
    Rank Posn
    #105
    Join Date
    Nov 09
    Location
    Posts
    1190

    I am not happy about you beating me to that............

    ALPHA!!!!!

    People Always Leave

  4. #4 / 7
    Brigadier General M57 M57 is offline now
    Standard Member M57
    Rank
    Brigadier General
    Rank Posn
    #73
    Join Date
    Apr 10
    Location
    Posts
    5083

     

    I think your chart broke the forum..

    Alpha wrote:

    Round 1

    Winners

    Round 2

    Winner

    Round 3

    Winner

    1

    3

    3

    13

    13

    1

    2

     

    4

     

    8

     

    3

     

    13

     

    1

     

     

     

     

     

     

     

    4

    4

    8

    8

    7

    2

    5

     

    11

     

    3

     

    6

     

    12

     

    2

     

     

     

     

     

     

     

    7

    8

    10

    7

    4

    11

    8

     

    7

     

    11

     

    9

     

    14

     

    15

     

     

     

     

     

     

     

    10

    11

    9

    2

    6

    5

    11

     

    15

     

    14

     

    12

     

    2

     

    5

     

     

     

     

     

     

     

    13

    13

    6

    1

    9

    9

    14

     

    5

     

    10

     

    15

     

    1

     

    12

     

    If no winner, keep only the top three and let them play until one has the best record. 

    So if there are three players left in a 2-player per game tournament, you would necessarily include a fourth player taken from the top of rest, and that player would have no chance of winning the tournament? (This is ok, I just want to make sure I understand it.)

    However, in the case of a 3-players per board tourney, with two remaining unclear winners, they would need to play a game to determine the winner, borrowing the best from the loser pile (you could have them play head to head - but what if the board is "set up" for 3 players?). If the borrowed player wins, then you will ostensibly have three players with identical records and they can play a rubber match.  Is this what you had in mind?

    Essentially, you turn the final determining rounds into an elimination tournament, inconveniencing the occasional odd man out with an extra game where in some cases, that player is back in the running. 

    BAO alternative:
    https://sites.google.com/site/m57sengine/home
    Edited Sun 6th Feb 06:52 [history]

  5. #5 / 7
    Pop. 1, Est. 1981 Alpha
    Rank
    Brigadier General
    Rank Posn
    #60
    Join Date
    Dec 09
    Location
    Posts
    991

    Strange, I didn't try to insert a table, but I did copy and paste from word; I wonder where that table came from.

    First question: Yes this would be correct for two players per game.  This would just provide a better way to select number 4, good observation (I hadn't thought of this).

    Second question: Yes.

    Last comment: Yes, those at the bottom which are included still have a chance, but it is a long shot.  At least this way there is some incentive to continue trying.

    Never Start Vast Projects With Half Vast Ideas.

  6. #6 / 7
    Standard Member RiskyBack
    Rank
    Colonel
    Rank Posn
    #105
    Join Date
    Nov 09
    Location
    Posts
    1190

    I'm using this thread to answer a question and to show, once again, that there are issues with the Swiss System.

    In this tournament there is 1 more game left with 4 players.  Each have 1w 1L and 11000 points.  I am leading the tournament with 2W 1L and am wondering if the winner of the last game is going to win the tournament or if I can make room in my trophy cabinet.  I really don't understand the scoring.

    http://www.wargear.net/tournaments/view/167

    People Always Leave

  7. #7 / 7
    Pop. 1, Est. 1981 Alpha
    Rank
    Brigadier General
    Rank Posn
    #60
    Join Date
    Dec 09
    Location
    Posts
    991

    Start making room, the winner of that game will have 12500 score.

    Never Start Vast Projects With Half Vast Ideas.

You need to log in to reply to this thread   Login | Join
 
Pages:   1   (1 in total)