Chess.com Isle Of Man Tournament 23 Sep to 1 Oct 2017

[Kevin Thurlow]

There was a weekender somewhere years ago where random pairings were used throughout, although nobody warned the players in advance. Top seed John Nunn found he was playing the second seed in round 1, so enquired why.

Arbiter - "It's random."
JN - "I can see that - why have you done it?"

Random pairings for a 5-round event are probably silly, having one random round for a much longer event could work. Someone could think, oh well, lost the first game, should be easier tomorrow. Oh, black against Kramnik...

[JustinHorton]

I managed to get Black against Tony Miles the a second round once despite having lost in the first. And that wasn't even random.

[Chris Rice]

Chess24 report, Chess.com report and ChessBase report on the first round.

The reports naturally focus on the new Rd 1 pairing system innovation and the resultant Caruana-Kramnik pairing. Every time something new is introduced there is of course a legion of stick in the muds who don't want to try anything new and John Saunders and the rest of the IoM organisers are to be applauded for giving it a go. For the record the suggestion that it ruined the Candidates race is of course rubbish. The same pairing could have happened in a later round and both players had the option not to enter, if they didn't want to play each other in Round 1 which I'm sure both Kramnik and Caruana would deny vigorously. The random pairing system for Round 1 was known well in advance and cleared with FIDE as I understand it so there really is nothing to whine about though of course that won't stop the whiners doing just that.

Round 2 pairings

[Jonathan Bryant]

Every time something new is introduced there is of course a legion of stick in the muds who don't want to try anything new ....

Well I disagree. The criticism that I’ve seen - and the reservations that I have - are not based around 'newness" but more about lack of logic/outright silliness/novelty for novelty’s sake.

I’m not sure I agree with every word, but this is a good read, I think >> https://www.chess.com/blog/SultanOfKing ... be-a-thing

[Chris Rice]

Every time something new is introduced there is of course a legion of stick in the muds who don't want to try anything new ....

Well I disagree. The criticism that I’ve seen - and the reservations that I have - are not based around 'newness" but more about lack of logic/outright silliness/novelty for novelty’s sake.

I’m not sure I agree with every word, but this is a good read, I think >> https://www.chess.com/blog/SultanOfKing ... be-a-thing

That article seems to argue that the IoM organisers didn't go far enough and make every round random? I suspect the IoM organisers didn't want to go "all in" because it might have been a disaster and just did it for one round which FIDE cleared, seems logical enough to me. Still haven't seen a coherent argument as to why its not a good idea.

[IM Jack Rudd]

Well, there is one major issue with random pairings that you may or may not think is a big one - it's that they are, by their very nature, not reproducible. If you ever get queries about whether a player has been treated unfairly by the pairings, it's a lot easier to answer that if there is a deterministic procedure you can point to. (This was an important point with respect to the Hou Yifan issue in Gibraltar this year.)

[Chris Rice]

Well, there is one major issue with random pairings that you may or may not think is a big one - it's that they are, by their very nature, not reproducible. If you ever get queries about whether a player has been treated unfairly by the pairings, it's a lot easier to answer that if there is a deterministic procedure you can point to. (This was an important point with respect to the Hou Yifan issue in Gibraltar this year.)

OK now we're getting to it, I was expecting more of the "fairness" argument to be played up and you're right, I don't think its a big issue. If Arsenal are drawn away five times in the FA Cup does anyone say its "unfair"? No, it's just the luck of the draw. This fairness argument is a myth. I have played a number of top half v bottom half tournaments over the years when I often had opponents who outgraded me by 400-500 points or more in the odd number rounds and who I similarly outgraded in the even number rounds. Is that a fair pairing sysem simply because its reproducible? Didn't matter to me that much, I still entered the tournaments. Now suddenly Caruana is randomly drawn to play Kramnik in round 1 of an Open tournament and everyone is crying about unfairness? They need to get over themselves.

Have to admit though Hou Yifan continuing to be drawn against female opponents in completely different pairing systems after giving up playing competitively in women's events defies all explanations.

[LawrenceCooper]

If Arsenal are drawn away five times in the FA Cup does anyone say its "unfair"? No, it's just the luck of the draw.

They'll have to get relegated twice before that can happen under the current rules

[Roger de Coverly]

Well, there is one major issue with random pairings that you may or may not think is a big one - it's that they are, by their very nature, not reproducible.

Pseudo-random numbers. It's a system where you always get the same sequence of equal probability 0-9 but the sequence itself depends only on a starting key.

https://en.wikipedia.org/wiki/Pseudoran ... _generator

In practice, deterministic pairing methods took hold before experiments in attempting to generate pairings by computer methods. Thus all the computer pairing systems were deterministic and needed some complex rules to produce a unique solution.

If you did use pseudo-random, you would still need rules about colour sequences and floats and the extent if at all you seeded one side of the pairings.

[NickFaulks]

Have to admit though Hou Yifan continuing to be drawn against female opponents in completely different pairing systems after giving up playing competitively in women's events defies all explanations.

Other than the obvious one, which is that these things do happen by chance.

[Richard Bates]

If you have no objective of trying to ensure the best performing players end up at the top of the tournament then there is obviously no problem with random pairings. Everybody has an equal chance of benefitting or suffering from "luck of the draw". But personally I don't see why this is desirable - and for professional players and with lots of money at stake I think one should obviously aspire to the initial objective. The argument that a player can sometimes benefit from a favourable draw within Swiss systems is an argument for seeking improvements to Swiss systems, not for making it worse and introducing random pairings and significantly increasing the chances of variance from an "average" set of pairings, and the potential size of that variance.

The FA Cup is imo not a great example because it is the exception rather than the rule in how sporting competitions are organised, which almost all retain some element of seeding to the benefit of the best teams/players and/or to seek that the concluding rounds are contested between the best. And even the FA Cup favours the better teams by entering them later in the competition.

As for the specific case of just randomising the first round to make the first round more interesting - well if this is an objective then frankly I think this is far better achieved through well designed accelerated pairing systems.

[Chris Rice]

Random Pairings Make Chess Players go Crazy by Greg Shahade

[Richard Bates]

Have to admit though Hou Yifan continuing to be drawn against female opponents in completely different pairing systems after giving up playing competitively in women's events defies all explanations.

Other than the obvious one, which is that these things do happen by chance.

Actually I read some decent arguments at the time of Gibraltar about how the circumstances of the tournament (unusually large female entry, combined with the rating strength band that a significant proportion of that entry were clustered in) actually made the outcome a bit more likely than pure "chance" would indicate.

[Roger de Coverly]

Actually I read some decent arguments at the time of Gibraltar about how the circumstances of the tournament (unusually large female entry, combined with the rating strength band that a significant proportion of that entry were clustered in) actually made the outcome a bit more likely than pure "chance" would indicate.

They haven't flagged the female players in the ranking list, so difficult to see directly how many female players there are. The field in the Isle of Man is comparable to the one in Gibraltar, so to the extent that she is paired downwards, it's going to be to a similar range of ratings, which as in Gibraltar may be heavy in terms of female players.

[Richard Bates]

The other point I'd make about standard Swiss systems is that (assuming there aren't some particularly irregular features in the rating distribution, and excluding "Swiss gambit" gaming of the system")if they favour anybody it tends to favour the stronger players. Taking two players in isolation the stronger player is more likely to be paired against a player weaker than themselves. Although this is compensated by the fact that when the two players are on the same score group in the same half of the draw the stronger player will usually have the tougher pairing.

And this is fine. If you want to have the best of the draw you just have to get stronger. Where pairing systems really fall apart is when there is a perception (real or imagined) that they actually favour weaker players.