SO WHO'S GOING TO SHEFFIELD? I AM!

[IM Jack Rudd]

It can, but there are some pretty specific constraints on that number; if {draws where the seed is white}!=1 and {draws where the seed is black}!=1, I don't think it can happen.

Surely your argument is assuming equal distribution of colours? Since colours take precedence, any imbalance between top and bottom half will potentially see the player in question "switch".

It won't - in the ½/1 scoregroup, the number of top-half white-seekers will always match the number of bottom-half black-seekers, and the number of top-half black-seekers will similarly always match the number of top-half white-seekers. With more than one draw on each side, you'll then always be able to pair with the right colours without a median-flip.

[Richard Bates]

It can, but there are some pretty specific constraints on that number; if {draws where the seed is white}!=1 and {draws where the seed is black}!=1, I don't think it can happen.

Surely your argument is assuming equal distribution of colours? Since colours take precedence, any imbalance between top and bottom half will potentially see the player in question "switch".

It won't - in the ½/1 scoregroup, the number of top-half white-seekers will always match the number of bottom-half black-seekers, and the number of top-half black-seekers will similarly always match the number of top-half white-seekers. With more than one draw on each side, you'll then always be able to pair with the right colours without a median-flip.

40 players:

22v2 draw
24v4 draw
26v6 draw
20v40 draw

In round 2 number 20 has to be paired against no 2,4 or 6, no?

[David Sedgwick]

Surely your argument is assuming equal distribution of colours? Since colours take precedence, any imbalance between top and bottom half will potentially see the player in question "switch".

It won't - in the ½/1 scoregroup, the number of top-half white-seekers will always match the number of bottom-half black-seekers, and the number of top-half black-seekers will similarly always match the number of top-half white-seekers. With more than one draw on each side, you'll then always be able to pair with the right colours without a median-flip.

40 players:

22v2 draw
24v4 draw
26v6 draw
20v40 draw

In round 2 number 20 has to be paired against no 2,4 or 6, no?

Yes, but Jack referred to "more than one draw on each side". In your example, there is only one draw on the "White top half side" (and three on the "Black top half side").

I suspect this is a case of arbiter speak confusing the rest of the populace.

After my faux pas in the other thread, I hope I'm on surer ground here.

[Roger de Coverly]

I suspect this is a case of arbiter speak confusing the rest of the populace.

Almost certainly. If you are number 20 in a 40 player tournament and draw round one , what are the relative chances of playing 1 to 19 or 21 to 39?

It depends on other results, but is it that far removed from 50:50?

[Stewart Reuben]

That is what Michael and Nigel intend to do.
Without having done the actual statistical analysis, I suspect there is more variance in the first round of international tournaments than in the others. This will be less pronounced in Sheffield because travelling is much less to national Championships.
In my experience, Nigel being a consummate professional will probably arrive Saturday.

Stewart Reuben

[E Michael White]

Why wasnt 20 Black v 40 in round 1 like 2 4 and 6 ?

[IM Jack Rudd]

I suspect this is a case of arbiter speak confusing the rest of the populace.

Almost certainly. If you are number 20 in a 40 player tournament and draw round one , what are the relative chances of playing 1 to 19 or 21 to 39?

It depends on other results, but is it that far removed from 50:50?

It's almost certainly quite a way away from 50:50 - but how far depends mostly on the probability of your being due the upfloat. (If you're upfloating, you'll very likely play 1 to 19; if not, you'll very likely play 21 to 39.) I can't be more specific about the probabilities without some actual simulation assumptions.

[David Sedgwick]

Why wasnt 20 Black v 40 in round 1 like 2 4 and 6 ?

A very good point.

However, I imagine Richard's original question, and the answers, would be equally valid if he'd referred to 19 v 39, or if his "other side" had been 21 v 1, 23 v 3 and 25 v 5.

[Richard Bates]

Why wasnt 20 Black v 40 in round 1 like 2 4 and 6 ?

A very good point.

However, I imagine Richard's original question, and the answers, would be equally valid if he'd referred to 19 v 39, or if his "other side" had been 21 v 1, 23 v 3 and 25 v 5.

Yep, that's what i intended to write. And i understand Jack's point now.

Anyway up to black against Adams now, I think. Probably less psychologically scared of that one, putting aside the 0-7 record or whatever

[Stewart Reuben]

I played John Nunn 5 times, all 5 with Black and lost them all. I have always thought, 'At least I didn't waste a white on him.'

Stewart Reuben

[Richard Bates]

I played John Nunn 5 times, all 5 with Black and lost them all. I have always thought, 'At least I didn't waste a white on him.'

Stewart Reuben

Yes, well, I've wasted 6 on Mickey. The only time i ever had black was the only game i've ever played needing a result (a win!) for a GM norm...

[Stewart Reuben]

The only time I played Mickey properly was at the British in Plymouth in 1989. It was the rapidplay and he had come to the event in order to score 6/6 for the Grand Prix. He had white, I never fully equalised and I went wrong on about move 65 in rook and pawn endgame. He was castigated for that for putting his whole day's trip at risk. His response, 'Oh well.'
It occurs to me he wins many games by eking out small advantages. The contrast of his style to that of Nigel's is fascinating and they have only once before both played in the British Championship at the same time.
You might expect there to be some measure of bitter rivalty between them. But I have never observed any of this.

Stewart Reuben

[John Upham]

Do anyone know which bookstall(s) (and operators) will be in attendance at this years Championships?

[Alex McFarlane]

Chess & Bridge supplied the best tender and were awarded the contract.

[E Michael White]

I meant to post a reply at the time to this posting by Stewart Reuben concerning accelerated parings, numerous pages back.

http://www.ecforum.org.uk/viewtopic.php ... 04&#p64404

It would be a good idea to ask the players in Sheffield which system(s) they would prefer to be used in the future.

It would be a very bad idea to sample in this way. You need to sample in addition those who might have played with their preferred system in use. But how do you sample players not there ?

Lottery
Seeded by Rating 1v33, 34v2 etc.
Seeded by Rating 1v64, 63v2 etc.
Accelerated Fractional
Accelerated Bonus points
Dubov
Burstein
For first round divide in 6 groups 1v3, 2v4, 5v6.
'Reuben' System
'Lamb' system 1v2 4v3

Only the third and last are totally flawed in concept in my opinion, and people can of course disagree with that.

Have any of the above systems been assessed by professional mathematicians ? To use a chess analogy chess players/arbiters often view pairings with what could be described as tactical arithmetic rather than look at the underlying structure.

Harry Lamb's suggestion is perfectly feasible and has some basis in the science of Information Theory, developed by Claude Shannon and others. Basically this shows that if a question has three possible outcomes in this case W/D/L then maximum information is gained if the probability of responses is roughly equal ie 33.33%. Harry is attempting to get close to that. Harry's system is similar to what I referred to in this post in 2009:-
http://www.ecforum.org.uk/viewtopic.php ... 87&#p13287

I don’t agree fully with Harry Lamb’s view as I think he will generate too many draws and without determining the range exactly, I believe players need to be paired against those roughly 100-200 points higher or lower in the early rounds. This will introduce a 2nd type of float and to balance things out this paring method could be used for 4 rounds. Some accelerated systems currently in use occasionally come close to this structure fortuitously.

Where are new ideas for improving systems going to come from if people just rely on computers?

I suspect these might emerge at some point from Monash University, or similar, where David Dowe’s team have the necessary credentials as long as they don’t get too tied up with MML.