SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
MattF
Posts: 925
Joined: March 14th, 2003, 7:15 pm

Re: Is Hou Yifan right??

February 6th, 2017, 6:04 pm

Yes! Tonight I'll split it by round. I've read somewhere that in the "Swiss system" you initiate pairing based on ranks (the off disgonal lines?) and in the next rounds you let winners of the previous round play losers of the previous round. If the observed pairing becomes explainable then we can look for manipulation (or reject it).
Don't bother. Do you really think chess geeks aren't all over this? The entire tournament pairings have already been duplicated independently using the same software. Hou Yifan is simply wrong.
Thanks!
They did?? Has anyone written something about it (link?)?
I didn't give a link as you'll have to pick out the posts from this thread about the whole tournament http://www.ecforum.org.uk/viewtopic.php?f=31&t=8904 
Alex Holowczak has duplicated the pairings. Also there was some discussion here http://en.chessbase.com/post/gibraltar-flash#discuss
 
User avatar
outrun
Topic Author
Posts: 4573
Joined: April 29th, 2016, 1:40 pm

Re: Is Hou Yifan right??

February 6th, 2017, 7:11 pm

Don't bother. Do you really think chess geeks aren't all over this? The entire tournament pairings have already been duplicated independently using the same software. Hou Yifan is simply wrong.
Thanks!
They did?? Has anyone written something about it (link?)?
I didn't give a link as you'll have to pick out the posts from this thread about the whole tournament http://www.ecforum.org.uk/viewtopic.php?f=31&t=8904 
Alex Holowczak has duplicated the pairings. Also there was some discussion here http://en.chessbase.com/post/gibraltar-flash#discuss
Thanks!! Very interesting. Backdoors... Israel.. It's everywhere the same!
I though nobody was looking at this, also didn't know about official reproducible pairing algorithms. All very interesting.
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Is Hou Yifan right??

February 6th, 2017, 7:36 pm

Hmmm.. So it seems the pairing algorithm may be deterministic as a function of the original rank list and then subsequent patterns of win-loss. If so, the only way to manipulate the gender ratio of a given player is to permute the original list so that the chosen player is pushed into a tree of possible matches that has a skewed gender ratio. That would take a lot of manipulation with lots of anomalous female-female pairs so that no matter who won, Yifan would face a female. That seems unlikely!

Curiously, Hou Yifan and Piorun Kacper have the same numerical rating yet Yifan was assigned the higher rank. Why? Had they flipped it, Kacper would have presumably faced a run of female opponents.
 
User avatar
katastrofa
Posts: 8772
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Is Hou Yifan right??

February 6th, 2017, 8:57 pm

She may be right or wrong, but it's certainly stupid to claim that she's either. What's the p-value of drawing at least 7 female competitors in 10 tournaments?

total No of players: n = 1000
female share: r = 1/10

p = 1/ binomial(n, 10) * (binomial(r*n, 7) * binomial((1-r)*n, 3) + binomial(r*n, 8) * binomial((1-r)*n, 2) + binomial(r*n, 9) * binomial((1-r)*n, 1) + binomial(r*n, 10)) << 0.05
 
User avatar
outrun
Topic Author
Posts: 4573
Joined: April 29th, 2016, 1:40 pm

Re: Is Hou Yifan right??

February 6th, 2017, 11:33 pm

this is what I got. n = 255, r =42/254, p ~= 1/5000

However, if it's not random like t4a and mattf say then p=1
 
User avatar
MattF
Posts: 925
Joined: March 14th, 2003, 7:15 pm

Re: Is Hou Yifan right??

February 7th, 2017, 12:16 am

Curiously, Hou Yifan and Piorun Kacper have the same numerical rating yet Yifan was assigned the higher rank.  Why?  Had they flipped it, Kacper would have presumably faced a run of female opponents.
Just alphabetical probably ... Hou is her surname. Kacper would only have faced the same run of females if both Hou and Kacper perfectly duplicated the other's results. Anyway it's now been established that the only tampering the organisers did was not to pair Israeli and Iranian players (which results in the Iranians having to default to avoid dire consequences back home). Of course this is undesirable but semi-official practice. Ironically it only came to light due to verification of the correctness of the pairings (and it didn't affect Hou's pairings in any way).
 
User avatar
Paul
Posts: 10093
Joined: July 20th, 2001, 3:28 pm

Re: Is Hou Yifan right??

February 7th, 2017, 6:57 am

Curiously, Hou Yifan and Piorun Kacper have the same numerical rating yet Yifan was assigned the higher rank.  Why?  Had they flipped it, Kacper would have presumably faced a run of female opponents.
Just alphabetical probably ... Hou is her surname. Kacper would only have faced the same run of females if both Hou and Kacper perfectly duplicated the other's results. Anyway it's now been established that the only tampering the organisers did was not to pair Israeli and Iranian players (which results in the Iranians having to default to avoid dire consequences back home). Of course this is undesirable but semi-official practice. Ironically it only came to light due to verification of the correctness of the pairings (and it didn't affect Hou's pairings in any way).
Which means in a year or two they will start tampering with the gender pairings, and then by race, by sexuality,...Yawn.
 
User avatar
katastrofa
Posts: 8772
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Is Hou Yifan right??

February 7th, 2017, 2:32 pm

Since women to men ration in Iranian team was 3:1 and in Israeli team was 1:8, it's easy to see that tampering with these matches could affect Hou's draws. Anyway, as calculated above by outrun and I, the result is highly unlikely (H0 is that the draws are random).

I had a colleague who was an Olympic swimmer. She told me that everyone knew that eliminations were rigged or bought. It looks like chess tournaments are the same.
 
User avatar
MattF
Posts: 925
Joined: March 14th, 2003, 7:15 pm

Re: Is Hou Yifan right??

February 7th, 2017, 3:01 pm

Since women to men ration in Iranian team was 3:1 and in Israeli team was 1:8, it's easy to see that tampering with these matches could affect Hou's draws. Anyway, as calculated above by outrun and I, the result is highly unlikely (H0 is that the draws are random).

I had a colleague who was an Olympic swimmer. She told me that everyone knew that eliminations were rigged or bought. It looks like chess tournaments are the same.
I've tried to explain in this thread several times now why Hou was wrong and her pairings weren't rigged. I'm fairly knowledgeable about the subject matter and know where to look and who to believe for info about this specific case. If you continue to assert "it's rigged" you just look like Amin and his significant breakthroughs. The Israel/Iran pairings were too low down the field to affect Hou who is a 2600 GM. Some rigging does happen in chess, particularly where titles or norms are concerned, but not in Hou's case.
 
User avatar
katastrofa
Posts: 8772
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Is Hou Yifan right??

February 7th, 2017, 3:53 pm

I realise I sound crazy to people who have no clue about statistics. Can you point me to your post actually explaining something (I've read the thread and couldn't find any)?
 
User avatar
outrun
Topic Author
Posts: 4573
Joined: April 29th, 2016, 1:40 pm

Re: Is Hou Yifan right??

February 7th, 2017, 3:55 pm

Kata, the pairing is done with a deterministic algorithm (based on elo rank and surname?), not a random one. The algorithms has been re-rerun by a 3rd party giving the exact same pairing.
 
User avatar
katastrofa
Posts: 8772
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Is Hou Yifan right??

February 7th, 2017, 4:01 pm

Thanks, outrun. Some posts in the linked chess forums misled me.
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On