2015 CYCC Pairing Manipulation

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • 2015 CYCC Pairing Manipulation

    Pairings in the first round are important and can play a crucial part especially in Swiss Tournaments where instead of play off's mathematical tie breakers are applied. Computers are used, since especially in later rounds logic becomes more complicated and chances of human errors are increased.

    The pairing logic of the very first round of Swiss Tournament is simple:

    a) players are sorted as per their ratings and ranked top rated 1st, next best 2nd and so on

    b) If there are odd number of players the lowest ranked player gets a bye

    c) The remaining even number of players or if one started with even number of players are grouped into two equal half's.

    d) Top player in first half is paired to play top player from second half, second player from top half is paired against second player from second half and so on. example: if there are 16 players parings would be 1-9, 2-10,3-11, 4-12, 5-13 .... 8-16.

    e) Depending on the coin toss it could be either 1B-9W, 2W-10B ....etc OR 1W-9B, 2B-10W etc

    The actual pairing in the U10 Girls section in CYCC 2015 were not altered and were as follows:



    However in quite a few sections the pairings were altered, one of the sections in which pairing was manipulated was the U14 Girls section.

    Without alteration the pairings in the U14 Girls would have been as follows using the standard SwissSys settings (you can try this yourself as well by using Chessresults starting ranks and SwissSys), Please take note of board 3 and board 4:



    This is the actual pairing:



    In order to achieve this pairing, two forced pairings must be used, there is simply no other way to achieve this without the use of forced pairings:

    These are the forced pairing settings:



    As you can see using these forced pairing settings the same results as in the actual pairings have been achieved:



    This manipulation was done by someone with access to the pairing software/computer, is deliberate plus though not highly, quite sophisticated in nature. It takes quite a bit of ingenuity to confine changes to two boards and not cause ripple effect changes to many boards especially to those that follow.

    People usually have high trust for anything that is generated from a computer, hence those that know how to manipulate the output have a better chance of getting away with their manipulation.

    In past CYCC's I have looked at, forced pairings have not been used in the first round. Why were the forced pairings used in certain sections? Who entered these forced pairings?

    It is highly disturbing that the pairings have been manipulated in such a manner. There are also numerous other rounds and sections where it is clear that pairings have been manipulated through the use of forced pairings. This is only one example.

    After this finding the integrity of the 2015 CYCC is now in question until proven otherwise.
    Last edited by Keerti Nyayachavadi; Sunday, 2nd August, 2015, 07:22 PM.

  • #2
    Re: 2015 CYCC Pairing Manipulation

    Conspiracy theories are always fun, but there is an aphorism, something like "do not ascribe to evil that which may be caused by incompetence".

    I was on site most of the time during the event, and two separate parents asked me to check the accuracy of the posted ratings. So I did. The ratings used were not the latest ratings update, remarkably enough. I exchanged an email with the chief arbiter, Aris Marghetis, and we agreed that it was water under the bridge, at that stage of the event.

    I think that the computer guy probably updated the ratings during the event, which is why the pairings on Chess Results appear not to follow the swiss rules, for the first round at least.

    Comment


    • #3
      Re: 2015 CYCC Pairing Manipulation

      Ah, that aphorism is known as Hanlon's Razor:

      "Never attribute to malice that which is adequately explained by stupidity."

      Thank you, Mr. Google.

      Comment


      • #4
        Re: 2015 CYCC Pairing Manipulation

        Also neglected in this analysis is that the software being used was Swiss Manager and not Swiss Sys.

        Comment


        • #5
          Re: 2015 CYCC Pairing Manipulation

          Yes, I observed during the tournament that the CFC ratings used for the pairings were from the June 24th update. Where FIDE or FQE ratings were higher, those were used instead, but based on the latest lists available (so Ellen's last FIDE result was included, as was Maili-Jade's last FQE tournament). Given the size of the tournament, I can appreciate why things would be done this way.

          If I fetch the June 25th numbers, the pairings I'd expect match the actual pairings. Leaving the higher-rated player on the left...

          Ouellet (FQE 1931) - Xu (CFC 1461)
          Wang (CFC 1816) - Mah (CFC 1447)
          Kaneshalingam (CFC 1769) - Tao (FIDE 1403)
          L Zhou (CFC 1766) - Gillan (CFC 1398)
          Li (CFC 1653) - Fenning (CFC ACTIVE 1274)
          A Zhou (CFC 1583) - Peng (CFC 1271)
          Nguyen (FIDE 1555) - MacIvor (CFC 929)
          Paraparan (CFC 1541) - John (CFC 911)
          Jiang (CFC 735)

          Comment


          • #6
            Re: 2015 CYCC Pairing Manipulation

            Originally posted by Vlad Drkulec View Post
            Also neglected in this analysis is that the software being used was Swiss Manager and not Swiss Sys.
            Just for fun, I downloaded the Swiss-Manager tournament file and loaded in my Swiss-Manager. Then I did new pairings for the first round.
            This is what I got:


            Comment


            • #7
              Re: 2015 CYCC Pairing Manipulation

              Originally posted by John Coleman View Post
              I think that the computer guy probably updated the ratings during the event, which is why the pairings on Chess Results appear not to follow the swiss rules, for the first round at least.
              For me it looks like the "computer guy" tried to avoid paring 2 girls from Windsor for the first round.

              Comment


              • #8
                Re: 2015 CYCC Pairing Manipulation

                There's nothing shady going on. As of the June 25th ratings, Gillan's rating is 1398. The first-round pairings are entirely consistent with the CFC ratings being filled in ahead of the tournament, and the FQE and FIDE ratings being looked up at the start of the tournament.

                Comment


                • #9
                  Re: 2015 CYCC Pairing Manipulation

                  Originally posted by Rene Preotu View Post
                  For me it looks like the "computer guy" tried to avoid paring 2 girls from Windsor for the first round.
                  If so, he failed miserably. Lily Zhou and Rahma Gillan are both from Windsor. :)

                  Comment


                  • #10
                    Re: 2015 CYCC Pairing Manipulation

                    Originally posted by Derek England View Post
                    There's nothing shady going on. As of the June 25th ratings, Gillan's rating is 1398. The first-round pairings are entirely consistent with the CFC ratings being filled in ahead of the tournament, and the FQE and FIDE ratings being looked up at the start of the tournament.
                    It looks like you're right. I've changed Gillan's rating to 1398 and got this pairings:

                    Comment


                    • #11
                      Re: 2015 CYCC Pairing Manipulation

                      Yes, I observed during the tournament that the CFC ratings used for the pairings were from the June 24th update. Where FIDE or FQE ratings were higher, those were used instead, but based on the latest lists available (so Ellen's last FIDE result was included, as was Maili-Jade's last FQE tournament). Given the size of the tournament, I can appreciate why things would be done this way.

                      If I fetch the June 25th numbers, the pairings I'd expect match the actual pairings. Leaving the higher-rated player on the left...

                      Ouellet (FQE 1931) - Xu (CFC 1461) FIDE
                      Wang (CFC 1816) - Mah (CFC 1447)
                      Kaneshalingam (CFC 1769) - Tao (FIDE 1403)
                      L Zhou (CFC 1766) - Gillan (CFC 1398)
                      Li (CFC 1653) - Fenning (CFC ACTIVE 1274)
                      A Zhou (CFC 1583) - Peng (CFC 1271)
                      Nguyen (FIDE 1555) - MacIvor (CFC 929)
                      Paraparan (CFC 1541) - John (CFC 911)
                      Jiang (CFC 735)
                      If FIDE ratings and the June 24th ratings were used explain what happened in U16 Open?

                      My son had his CFC rating matched to his USCF rating on July 1st. He played with that rating throughout this years CYCC. If the ratings were pulled as of June 24th, his rating would have been 1190 compared to 1560. So even if the ratings were wrongly posted on ChessResults, if the first round used his June 24th rating instead of playing Razvan Preotu he would have been paired with Diwen Shi, so clearly his 1560 rating (which was matched on July 1st was used). So did the U14 girls use the June 24/25th ratings and U16 Open used the July 1st ratings? How is that fair to the rest of the sections that supposedly used the June 24/25th ratings? This claim of using the June 24th ratings seems to be false.

                      If the higher of the FIDE/CFC ratings were used, If FIDE ratings were higher that would have been selected and posted, as is the case of CFC/FQE ratings, then why did Stefano Lee who had a CFC rating of 1444 and FIDE rating of 1544 (which was updated well before June 24th) play under was the CFC one? Please also note that U16 and U18 are FIDE rated while all other sections were not.

                      Also neglected in this analysis is that the software being used was Swiss Manager and not Swiss Sys.
                      I have tried this same scenario in SwissManager and the same results are achieved. You should know better Vlad that the Swiss pairing logic isn't different between SwissSys and SwissManager.

                      I think that the computer guy probably updated the ratings during the event, which is why the pairings on Chess Results appear not to follow the swiss rules, for the first round at least
                      SwissManager directly uploads the tournament (including the ratings used for pairings) as it is going on to ChessResults. So are you saying initially in SwissManager, the June 24/25th ratings were imputed into the program for pairings, then changed to the July 1st ratings, which was posted on ChessResults then changed back the next day to the June ratings before the Round 2 pairings?

                      Your contention is total nonsense, since ratings are consistent in all 7 rounds and starting rank. If there was a change since the tournament was posted on ChessResults after the first round, the old ratings would have shown in ChessResults for round 1 if they were used for the round 1 pairings.

                      Comment


                      • #12
                        Re: 2015 CYCC Pairing Manipulation

                        Originally posted by Derek England View Post
                        There's nothing shady going on. As of the June 25th ratings, Gillan's rating is 1398. The first-round pairings are entirely consistent with the CFC ratings being filled in ahead of the tournament, and the FQE and FIDE ratings being looked up at the start of the tournament.
                        That's total nonsense, then Gilan's rating would have shown up as 1426 in both:

                        Ranking after Round 1

                        Board Pairing Round 1
                        Last edited by Keerti Nyayachavadi; Monday, 3rd August, 2015, 01:30 AM.

                        Comment


                        • #13
                          Re: 2015 CYCC Pairing Manipulation

                          Originally posted by Keerti Nyayachavadi View Post
                          If FIDE ratings and the June 24th ratings were used explain what happened in U16 Open?
                          Looking at the data, I'd say that the same general procedure was used for U16 as U14G, except that the latest CFC data was used for the pairings. When I look up the July 1st CFC numbers, the July FIDE rating for Lee and FQE ratings posted up to July 2nd, then split the 26 players in half and pair them up, here's what I'd get (again, I'm leaving the higher-rated player on the left so it's easier to follow the rating trend).

                          Preotu (CFC 2636) - Nyayachavadi (CFC 1560)
                          Yu (FQE 2350) - Colvin (CFC 1553)
                          Zhu (FQE 2318) - Hay (CFC 1546)
                          Doknjas (CFC 2316) - Lee (FIDE 1541, current 1566 minus 25)
                          Shi (CFC 2214) - Du (CFC 1452)
                          Y Li (CFC 2211) - Richards (CFC 1395)
                          Zhong (CFC 2114) - Ban (CFC 1326)
                          Graif (CFC 2025) - Kanwal (CFC 1132)
                          Petersen (CFC 1898) - Mainville (CFC 1083)
                          D'Amore (CFC 1784) - Vu (CFC 1010)
                          Zhang (CFC 1763) - Gaudette (CFC 956)
                          Baskaran (CFC 1669) - Houle (CFC 950)
                          J Li (CFC 1653) - MacIvor (CFC 948)

                          I believe these are identical to the first round pairings in U16.

                          Comment


                          • #14
                            Re: 2015 CYCC Pairing Manipulation

                            Pairings are a sensitive question, and it is understandable they generate strong emotions.

                            As an interested observer who didn't have a stake in any of the sections mentioned so far, I diligently read the arguments of both sides and was genuinely intrigued. As my custom is, I decided to do a bit of fact finding myself to try to better figure out an issue where I am not an expert. When I looked at the Girls U-08 section (again, the one where I didn't have any pecuniary interest), there was a peculiar anomaly.

                            Please take a look at the Rd 1 pairings in Girls U-08:
                            • there are 10 girls in the section, with starting ranks 1 through 10,
                            • on Board 1, girl # 6 is paired with girl #8,
                            • on Board 2, girl # 2 is paired with girl # 7,
                            • on Board 3, girl # 9 with girl # 3,
                            • on Board 4, girl # 4 with girl # 10,
                            • and on Board 5, girl # 1 (!) with girl # 5,


                            Being a layperson, I would think that only pairing on Board 2 was correct, and there was no way any of the pairing software could produce the ratings that were actually used, given the starting rank data.

                            I'd be grateful if any of the experts who made this thread so fascinating could load the data file in their software and test the pairings. It'd be great if someone could explain why these pairings were produced and were fair to the participants.

                            Thanks in advance.

                            Comment


                            • #15
                              Re: 2015 CYCC Pairing Manipulation

                              At a tournament it is useful to post a crosstable with the rating used for pairing, perhaps email the ratings to the players.

                              There is no mischief by a TD. In rare cases a fixed pairing had been done for players not yet in town. I hate pairing siblings but it can't be stopped.

                              I used to pair by hand and could explain the logic. Now with computers we look at a printout and try to figure it out. At a previous Canadian Open garbage was barfed out but wasn't noticeable by looking at the printout, whereas by hand pairing cards we would see rating, colour history and paired up history.

                              Comment

                              Working...
                              X