Human factor in making the draw

Discussion in 'Japan Open / China Open 2006' started by 2cents, Oct 7, 2006.

  1. 2cents

    2cents Regular Member

    Joined:
    Aug 9, 2003
    Messages:
    2,263
    Likes Received:
    3
    Occupation:
    Landlording, providing vacation homes at Disney wo
    Location:
    Vacation Home @ Disney World
    There will be 12 China singles in China Open. It is strange that 9 out of 12 will be in the first half, the other 3 in the bottom half. It sounds away from random selection. There are 6 more players in the upper half than the bottom half. This is unexplainable from statistics normal distribution.

    Since there are some rules to make seeds distributed evenly. So upper to bottom half ratio for seeds is 3:2, it is okay. But

    For non-seeded CHN players, 6 out of 7 in the upper half, while only 1 in the bottom half. The chance for this happening is about 4%.

    More, those unseeded CHN players concentrated in Bao CL and Lin Dan's 1/16.

    It is more like a plot to murder those unknown CHN players as early as possible, to avoid them destroying the pecking order of world badminton. :mad:
     
  2. maa2003

    maa2003 Regular Member

    Joined:
    May 7, 2006
    Messages:
    1,449
    Likes Received:
    1
    Location:
    Jakarta, Indonesia
    we do not know how IBF did a draw .........
    but to my understanding, IBF put the seeded players first in the position, which is almost fixed position, then randomly pick other players from the rest ..........
    so, the draw shall be fair enough.

    if I see correctly, in the MS of China Open draw, 6 in the top-half and 3 in the bottom half.

    anyway, just enjoy the games .......

     
  3. 2cents

    2cents Regular Member

    Joined:
    Aug 9, 2003
    Messages:
    2,263
    Likes Received:
    3
    Occupation:
    Landlording, providing vacation homes at Disney wo
    Location:
    Vacation Home @ Disney World
    There are 9 in the top-half and 3 in the bottom half.

    You see 6 in the main draw, there are also 3 other chinese in the qualification round, which are also in the upper half.
     
  4. smashko

    smashko Regular Member

    Joined:
    May 6, 2006
    Messages:
    22
    Likes Received:
    0
    Location:
    calgary
    2cents, your plot theory does not make any sense to me. Some of these unknown Chinese players are more dangerous to top seeds like LD and BCL than many of the known non-Chinese players. They are unknown mainly because they haven't got opportunities to play in international competitions and if they do they often produce major upsets, just like LD and BCL before becoming known. If it were the otherway round in the draw, namely 9 new Chinese players in the other half, I would think it as a plot, because they might be used to kick the other main title contenders out or tire them out. I would not be surprised to see either LD or BCL be beaten by one of these unknown players.
     
  5. 2cents

    2cents Regular Member

    Joined:
    Aug 9, 2003
    Messages:
    2,263
    Likes Received:
    3
    Occupation:
    Landlording, providing vacation homes at Disney wo
    Location:
    Vacation Home @ Disney World
    What you said was the same as I said. Because those unknown Chinese are pretty good, if they played more often, a lot of players, like Eric Pang, would be out of top 20. That's the exact reason IBF arranged the draw, to localize or contain Chinese players. Let them play each other first, so that other players still have chances to guaranteed positions in quarterfinals. That's my thinking. :)
     
  6. event

    event Regular Member

    Joined:
    Jan 27, 2005
    Messages:
    2,901
    Likes Received:
    278
    Location:
    Korea
    Read you statistics textbook again. The only time you can expect a normal distribution is with a large sample size. You're talking about a sample size of exactly 1(one)! One tournament looks suspect. Do some research and plot the distribution of top/bottom half ratios of Chinese entries and then start drawing conclusions about this so-called "human factor" (aka bias). On the other hand, if you wish to look at the individual selection, supposedly random, of Chinese players, you're looking at a sample size of 7, maximum. Still not enough to expect a normal distribution.
     
    #6 event, Oct 7, 2006
    Last edited: Oct 7, 2006
  7. event

    event Regular Member

    Joined:
    Jan 27, 2005
    Messages:
    2,901
    Likes Received:
    278
    Location:
    Korea
    In the Japan Open, you have a ratio of 4:3 bottom half to top half, including 2:1 among unseeded players. That sample size of 3 is not statistically significant but neither is 7, sadly. If you're suspecting bias, though, keep going until you've got a sufficiently large sample.
     
  8. 2cents

    2cents Regular Member

    Joined:
    Aug 9, 2003
    Messages:
    2,263
    Likes Received:
    3
    Occupation:
    Landlording, providing vacation homes at Disney wo
    Location:
    Vacation Home @ Disney World
    I have to point out that, event, you misunderstood normal distribution. This forum is not about statistics, I'm sorry to talk about it a little bit. however I would like to just give you some hints to understand what I'm talking about.

    Whether or not should be in normal distribution, depends on the intrinsic of the physics process. Not the sample size. Sample size will determine the parameters, which shapes the bell curve. The true normal distribution intrinsic property is the equal chance, or absolutely random. Like in the theory of error, the random error is normal distribution, no matter you measure just once, or ten thousand times. While the system error is not a normal distribution, no matter how many times you measured.

    6 of 7 on one side, the other 1 of 7 on the other side, which is good enough. For example, common sense, if you pick a ball blindly from a bag which has only black and white balls. Say you pick 7 times, and 6 times you got a black ball, only 1 time you got white ball. Is that enough to disapprove that there are equal number of black and white balls in the bag? Well, we still cannot 100% disapprove, but we can disapprove at p value of 5%, which means 95% confidence(of course this is not accurate, just use it for common sense). There are a lot tests for the normality tests, such as normal probability plot, Jarque-Bera test,...
     
    #8 2cents, Oct 7, 2006
    Last edited: Oct 7, 2006
  9. event

    event Regular Member

    Joined:
    Jan 27, 2005
    Messages:
    2,901
    Likes Received:
    278
    Location:
    Korea
    I didn't misunderstand normal distribution. You are correct that if you draw one ball out of a balanced bag blindly seven times you have 128 colour permutations and only 7 of those are going to be exactly six of one colour and one of the the other. So only a 5% chance of getting that outcome. Do the same experiment 100 times, and you'd expect to get an outcome like that only 5 times. Do the experiment once, and you get only one of 128 possible outcomes. If you want us to believe there is bias, show us a pattern of hometown players getting stacked in one half of their home tournament draws.
     
  10. previah

    previah New Member

    Joined:
    Sep 12, 2005
    Messages:
    4
    Likes Received:
    0
    Location:
    delft
    Yay....statistics....!!!!!!:crying: :crying: (always hate this course:p )
     
  11. hcyong

    hcyong Regular Member

    Joined:
    Mar 7, 2004
    Messages:
    5,558
    Likes Received:
    104
    Occupation:
    Software Engineer
    Location:
    KL & Sg
    Agree with event here. 2cents, your understanding of statistics is somewhat inaccurate.
     
  12. Baderz_Jas

    Baderz_Jas Regular Member

    Joined:
    Jan 18, 2006
    Messages:
    2,360
    Likes Received:
    0
    Occupation:
    Student, business man wannabe =P
    Location:
    Cambridge, UK
    What? :eek: :D 6 in top-half and 3 in bottom-half :eek: :crying: why? :crying: :D They just don't want China to dominate the world of badminton :D :D :D
     
  13. jug8man

    jug8man Regular Member

    Joined:
    Nov 7, 2003
    Messages:
    2,123
    Likes Received:
    6
    Occupation:
    MultiTasked guy, Stress Addict, Leisure Bum, mad c
    Location:
    Malaysia
    What ever you guys say............. I don't think Andrew Smith will fall on the first round to the Chinese Player 'You Hao'.

    Any takers????

    TBBMBB(N)
     
  14. jug8man

    jug8man Regular Member

    Joined:
    Nov 7, 2003
    Messages:
    2,123
    Likes Received:
    6
    Occupation:
    MultiTasked guy, Stress Addict, Leisure Bum, mad c
    Location:
    Malaysia
    The Funny thing about stats.... assuming this is an actual honest draw.

    1) You get a lopsided natural & unmanipulated random draw result like this and statiscians call it BIASED.

    2) You manipulate a draw so that it goes fairly down the middle and call it UNBIASED.

    Its a funny world we live in.

    TBBMBB(N)
     
  15. Baderz_Jas

    Baderz_Jas Regular Member

    Joined:
    Jan 18, 2006
    Messages:
    2,360
    Likes Received:
    0
    Occupation:
    Student, business man wannabe =P
    Location:
    Cambridge, UK
    I hope he doesn't fallonthe first round too! :D But you never know how good the Chinese players are, they never played international! :D :p
    BTW, what does TBBMBB(N) :confused: :)
     
  16. jug8man

    jug8man Regular Member

    Joined:
    Nov 7, 2003
    Messages:
    2,123
    Likes Received:
    6
    Occupation:
    MultiTasked guy, Stress Addict, Leisure Bum, mad c
    Location:
    Malaysia

    Hello :)

    this is short for :
    The Big Bad Mean Borneon BaddyNut.

    This short form version was given by hcyong, another BadmintonCentral member.

    Well, Andrew Smith did beat Sato Shoji earlier this year... So my $$$ is on him.


    TBBMBB(N)
    har har
     
  17. Baderz_Jas

    Baderz_Jas Regular Member

    Joined:
    Jan 18, 2006
    Messages:
    2,360
    Likes Received:
    0
    Occupation:
    Student, business man wannabe =P
    Location:
    Cambridge, UK
    Hahaha!!! :)

    oh i see! hahahah :D :D :D Go Andrew Smith!

    thanks
    Baderz_Jas :)
     
  18. Cheung

    Cheung Moderator

    Joined:
    Apr 25, 2002
    Messages:
    23,793
    Likes Received:
    4,776
    Occupation:
    wannabe badminton phototaker
    Location:
    Outside the box
    Before you can take one tournament as an example of a conspiracy theory, have you got other data to back up your claim? A draw like this could be due to random error rather than systematic error. Just because it happens with a p<5% doesn't mean the result cannot happen (as you pointed out yourself later).
     
  19. blessing

    blessing Regular Member

    Joined:
    Jun 30, 2006
    Messages:
    686
    Likes Received:
    1
    Location:
    Malaysia
    dun worry.. dey r goin to redraw da whole thing.. relax la..
     

Share This Page