Random pairings in Swiss tournaments

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

  • Random pairings in Swiss tournaments

    I am starting a dedicated thread so that arbiters, organizers and players can weigh in together on this "hottest new thing."

    Originally posted by Wayne Komer View Post
    Isle of Man International 2017
    The Man Behind the Random Pairing Idea is Unmasked

    Moving on from random Norwegians to random pairings: yesterday evening the top eight players lined up on stage at the tournament hall to draw the names of their opponents from a tombola machine. But before describing the fun of the draw, let's first consider what was behind the move to a random pairing for round one. I can reveal here, exclusively, that it started some months ago with a conversation between tournament director Alan Ormsby and... let's just call him Mr Y for now.

    During this telephone discussion, Mr Y expressed his dissatisfaction with certain aspects of Swiss tournaments. He found round one of traditional Swiss tournaments particularly tedious, with a long litany of mismatches and only the very occasional newsworthy David success against Goliath. Why, he asked rhetorically, was it axiomatic that the top players should be kept apart until the latter stages of a tournament? He empathised with chess scribblers and promoters who had to try and make reports of early rounds of Swisses interesting with so little useful material to work from. Why were pairings so slanted in favour of higher rated players anyway? It was no longer necessary to add sweeteners to attract strong players these days since the prize list was sufficient inducement in itself. Why were pairing regulations so complicated? He expressed the view that the regulations should be sufficiently simple so that they could be explained to the man in the street in five minutes. Alternation of colours - yes, of course - pairing with someone on the same or similar score - yes - but why all the other palaver? If random pairings make it a bit harder for norm seekers - tough. The world has enough titled players anyway, and this would compensate for the rating inflation that has made the gaining of titles too easy.

    On and on Mr Y pontificated in his usual long-winded way. Only he usually does this in writing... and some of you might be starting to wonder whether Mr Y reminds you of the current writer. Well, I suppose it is time to come clean. I shall now out myself as Mr Y. Yes, I was responsible for the initial idea. Alan Ormsby thought it sounded interesting and took it away to discuss and work through with arbiters and officials. I rather thought that would be the last I heard of it - in reality I was only sounding off, like the grumpy old man I am turning into - but a few weeks later Alan phoned me back to say that they had checked it over and decided it was a runner. The implementation has been shorn of some of my wilder ideas, which is probably just as well. But if anyone wants a scapegoat for the random pairing idea I guess that will have to be me. I shall ascend to the verbal guillotine of Twitter with as much dignity as I can muster. "It's a far, far better thing I do...," etc, etc.

    Round 1 Draw

    First man to the tombola machine was number one seed Magnus Carlsen. His first job was to decide the colour of the top board pairing and he duly drew a white queen from the bag. (Thereafter the boards alternated colour for the leading players so no further colour draw was required.) Then he drew out the name of an Icelandic player, Bardur orn Birkisson, rated 2167. Master of ceremonies Mike Klein referred to him as "a fellow Scandinavian" but the world champion swiftly corrected him: "Iceland is not Scandinavia." Nice pairing for Magnus - I think he now owes me a tenner for my random pairings suggestion, just as he does the arbiter for that illicit coach journey (oops! I think I just blurted out the second Mr X's identity).

    Next up was Vladimir Kramnik, who already knew he was going to be Black in the first round.

    Vlad's face registers surprise - "Caruana..."

    "... Fabiano!" Vlad Kramnik's surprise turns to amusement as he finds he is playing the number three seed.

    Vlad Kramnik read out the name as it was given (surname first) on the slip of paper: "Caruana Fabiano!" and suddenly his faced was wreathed in smiles. Fabiano Caruana also grinned at the news. At least one person in the auditorium - namely me - was greatly relieved at their good-humoured reaction at being paired with the highest rated player remaining in the draw. Generally the players seemed to have reacted quite well to the random pairing idea. We shall have to see how they feel after the game. I might need a safe house. Any offers?

    Vishy is paired with an IM from the USA. Arbiter Matthew Carr writes the name down, while Mike Klein asks the questions.

    Vishy Anand was the next to step up and found he was paired with Marc Esserman, an IM from the USA. Vishy smiled as he was reminded that Esserman had held him to a draw last year in Gibraltar.

    http://iominternationalchess.com
    The key sentence:
    "If random pairings make it a bit harder for norm seekers - tough. The world has enough titled players anyway, and this would compensate for the rating inflation that has made the gaining of titles too easy."

  • #2
    Re: Random pairings in Swiss tournaments

    GM Jon Ludvig Hammer advocates for an even more radical approach.

    "Random pairings - is this going to be a thing?"
    "We got some excitement in the world of chess on Friday! I was one of those who hadn't read the 2017 Chess.com Isle of Man regulations, and was taken aback by the announcement of Caruana-Kramnik as a 1st round pairing. Like most people, when faced with something completely contrary to systems we're used to, I thought this seemed plain wrong. Random pairings, and then to have the biggest outlier of them all - the number two seed drawn against the number three seed - on the very first try! As added spice, it's THE integral pairing of the rating race for two rating spots in the Candidates.

    It has taken me some hours to digest, but I'm liking it more and more. Here's why:

    I think this is the first basic point: What could be more fair than everyone having the same risk - presumably a disadvantage - of facing the World Champion? If we say this system is giving the best players an disadvantage, I think that logically also means that traditional swisses actually GIVE the best players an advantage - as they don't run the risk of a very tough first round pairing.

    Greg elaborates on his blog. I want to highlight these two paragraphs:

    The Swiss System is inherently slightly unfair. Over a short sample of tournaments it will likely be MORE fair than the results of randomized pairings, but over a long sample of tournaments the randomized pairings will be fairer to all players in the event (not just the top ones).

    The issue is that chess players cannot see past the one individual tournament, and therefore they are happy to accept some small degree of inherent unfairness in order to assure that any one tournament isn’t too affected by “lucky pairings”.

    In fact, I think where the Isle of Man organizers went wrong is that they aren't principled about this - they are just using it as a novelty. Like Greg notes, this system will create lucky (and unlucky) pairings in the short run. That's part of what makes it exciting. Could a 2200-player be on full score after five rounds - and face off with the other guy on five points - a certain 2800-rated World Champion? But precisely for this reason, you need to stick with it for the duration of the event.

    We've actually had a cousin of this in Norway the last four decades: Our way (the monrad system) is similar to the Isle of Man system, as starting numbers are drawn by random - but after that it resembles a swiss again. Start number 1 has priority for floating, and as such is considered an advantage. By doing a purely random draw, no one has floating priority (and depending on which tiebreak you use, it may not even matter). It was scrapped by the Norwegian Chess Federation this summer, as the swiss system was considered more likely to find a fair winner. But do keep in mind: Most Norwegian weekend tournaments are five rounds. That's too short to avoid outlier outcomes to affect many tournaments. The monrad system wasn't great for short tournaments - and neither is a system of completely random pairings.

    That's why I think the Isle of Man pairing was unfair - they didn't follow through. It's essentially a one round monrad, and that doesn't give the weird "accidents" a chance to correct themselves in the long run. After losing to Caruana in the first round, Kramnik is back to the swiss system - one point behind, trying to catch up - while Caruana doesn't run the same risk of suddenly getting paired with Carlsen in round 2. Caruana's high rated privileges of the swiss system has been given back to him. Not to mention Carlsen, who got through the first round unscathed, with the full point never being in doubt. The World Champ enjoys a one-point head start on Kramnik.

    I would love to see a tournament who takes the concept as a principle: Random pairings (within the point group, of course) all the way. I think it would make for one heck of a show. It's a swiss in which everyone is unrated. And after all, isn't that what we tell our kids? Disregard rating - just play the game!"
    Last edited by Vadim Tsypin; Monday, 25th September, 2017, 02:50 PM.

    Comment


    • #3
      Re: Random pairings in Swiss tournaments

      At the club we decided to use 1 vs 2, 4 vs 3 etc pairings for the first round. We'll see how it works tonight :) We'll switch to normal Swiss for later rounds.

      Comment


      • #4
        Re: Random pairings in Swiss tournaments

        Pairing Caruana against Kramnik in the first round of an open tournament is so dumb it defies logic.

        The problem with random pairings is simple: the larger the tournament, the more likely you're gonna see a guy with a low rating raise to the top boards just by pure luck, being paired down round after round. Then it messes up the results on the top boards at some point. Which means a relatively poor show for the viewers.

        Just tighten the sections or use accelerated pairings. These concepts are not really hard to explain to your average Joe.

        I'm not against random pairing per se, but keep it for a tournament where the difference in ratings is not that high.

        Comment


        • #5
          Re: Random pairings in Swiss tournaments

          Originally posted by Vadim Tsypin View Post
          I am starting a dedicated thread so that arbiters, organizers and players can weigh in together on this "hottest new thing."



          The key sentence:
          "If random pairings make it a bit harder for norm seekers - tough. The world has enough titled players anyway, and this would compensate for the rating inflation that has made the gaining of titles too easy."
          That, I have to agree. Some players got their titles just because of inflation.

          However, I'm not sure random pairing are the solution. Ratings will keep going up nonetheless.

          Comment


          • #6
            Re: Random pairings in Swiss tournaments

            Originally posted by Egidijus Zeromskis View Post
            At the club we decided to use 1 vs 2, 4 vs 3 etc pairings for the first round. We'll see how it works tonight :) We'll switch to normal Swiss for later rounds.
            And why would you want to do that??

            Comment


            • #7
              Re: Random pairings in Swiss tournaments

              The Isle of Man tournament has been using normal Swiss pairings starting in the second round.

              Comment


              • #8
                Re: Random pairings in Swiss tournaments

                And I just don't understand the reasoning behind the equation "random = fair".

                Randomness is not fair. Far from it, actually. By the very definition of randomness it means someone will be given a better 'hand' (here, easier matchups). How is that fair?

                Being fair means acknowledging past results. In chess, this is conveyed in the very simple form of a chess rating, which we can use in order to decide on future actions, like pairings in a swiss system.

                We can advocate for random pairings, I'm not even against the idea. But let's not pretend it's 'fair'. Pairing Kramnik in the same way as we pair a 2100 guy is not fair. It's just a loss of time for everyone involved.

                Comment


                • #9
                  Re: Random pairings in Swiss tournaments

                  Originally posted by Mathieu Cloutier View Post
                  The problem with random pairings is simple: the larger the tournament, the more likely you're gonna see a guy with a low rating raise to the top boards just by pure luck, being paired down round after round. Then it messes up the results on the top boards at some point. Which means a relatively poor show for the viewers.
                  As opposed to the thrill ride that is round 1 for viewers?

                  In any event, your scenario is not very possible. To be paired down round after round, there must be continually players lower than you who are also somehow winning. In the end, it will all work out if the tournament is sufficiently long.

                  To be honest, I would actually find someone running white hot at the pairings draw and somehow ducking all the top people more interesting.

                  Comment


                  • #10
                    Re: Random pairings in Swiss tournaments

                    Originally posted by Mathieu Cloutier View Post
                    And why would you want to do that??
                    no need for warm-up time :) The club games mostly to have a good game as no prizes for regular nights.

                    Comment


                    • #11
                      Re: Random pairings in Swiss tournaments

                      The regular Swiss system is not the best, but nobody knows anything better. Any attempts to "improve" it should be taken very carefully. Let's say organizers decided to make random pairings fo all 9 rounds. I'll take this Isle of Man tournament as an example.

                      They have 160 players. Group A (amateurs) consist of lowest 64 players of this tournament, rated 2400 and below. They play one another in first 6 rounds. Sure, it's very unlikely, but not impossible. They don't make draws.

                      After 6 rounds, one player L (lucky amateur) has 6 points. 6 players from group A have 5 points. What happens next? Now our L will play against strong players and lose? No. Nobody among top-96 (160-64) players has 6 or 5.5 points, which is very possible. Some of them have 5 points, but L got "friendly" pairings again and plays against a player from group A.

                      L wins in round 7. 2 players from group A played one another, one of them won and has 6 points. He plays against L in round 8. L wins.

                      After 8 rounds L has 8 points. Nobody from his group has 7 or even 6.5 (they don't make draws). 2-3 top players have 6.5. Let's say, Carlsen is one of them. He played 8 games against top players, average rating of his opponents is around 2700.

                      6.5/8 is a great score even for Magnus, he gains more than 10 rating points. Finally, in round 9 Carlsen plays against L and wins. L finished 1st with 8, Carlsen 2nd with 7.5.

                      Is it fair?

                      Comment


                      • #12
                        Re: Random pairings in Swiss tournaments

                        I played in a tournament in the 1970's where the organizer new nothing about ratings...so it was all random...I did not think it made sense at the time...and today...I still don't think it makes any sense...but I did win the event LOL

                        Comment


                        • #13
                          Re: Random pairings in Swiss tournaments

                          Originally posted by David Ottosen View Post
                          As opposed to the thrill ride that is round 1 for viewers?

                          In any event, your scenario is not very possible. To be paired down round after round, there must be continually players lower than you who are also somehow winning. In the end, it will all work out if the tournament is sufficiently long.
                          If the tournament has enough players, you are almost sure to see one lower rated guy being paired down for the first 3-4 rounds, then it messes up the pairings for round 5, which has an impact on round 6 and so on...

                          It's a logical fallacy to randomly pair 300+ chess players on the premise that an outlier is unlikely in terms of favorable pairings. Each and every player is unlikely to be the outlier, but you have hundreds of them, so the most likely result is that one of them will be the outlier and will get favorable pairings for the first X rounds.

                          Comment


                          • #14
                            Re: Random pairings in Swiss tournaments

                            And on a general basis, I don't know of any competitive activity where amateurs are routinely paired against professionals, or even against the very best in the world... except for chess.

                            Pairing a 2100 with a 2800 doesn't make an ounce of sense. But we do it and plenty of chess aficionados applaud.

                            Comment


                            • #15
                              Re: Random pairings in Swiss tournaments

                              I would have no problem playing in an event with completely random pairings. However if they are going to ignore the ratings then they shouldn't have any rating category prizes.

                              Think of it like a poker tournament where the luck of the draw puts three top level pros at the same table while the next table over has no pros and a bunch of fish.
                              "Tom is a well known racist, and like most of them he won't admit it, possibly even to himself." - Ed Seedhouse, October 4, 2020.

                              Comment

                              Working...
                              X