SO WHO'S GOING TO SHEFFIELD? I AM!

[Roger de Coverly]

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

An old system of doing the British Championship was to arrange the top half players in grade order and then randomly draw their opposition. So you avoided clashes like Penrose-Keene in the first round, but if you qualified even as the lowest graded qualifier you might get a shot at someone famous early on.

Some sort of hybrid between seeded and random might yet have its points if you used computerised pairings to avoid suggestions of arbiter bias. If you use pseudo-random numbers to generate the pairing lottery, you will always be able to reproduce the pairings on another computer. So perhaps the pairing principle is that you sort the players by rating, but you assign opponents by lot. Thus if you are midway in the a score group, but in the top half, you might get to play anyone from about your own rating downwards. This avoids one of the drawbacks of both accelerated and seeded systems, that the players diagonal to you on the next boards in the first round are those whose games you will likely see a lot of, but are among your least likely opponents in the whole event.

[Adam Raoof]

http://www.britishchess2011.com/entrants.htm

Anyone have any opinions about why the Week One events have fewer entrants than the other events on offer?

[Sean Hewitt]

An old system of doing the British Championship was to arrange the top half players in grade order and then randomly draw their opposition. So you avoided clashes like Penrose-Keene in the first round, but if you qualified even as the lowest graded qualifier you might get a shot at someone famous early on.

I like that!

[Roger de Coverly]

http://www.britishchess2011.com/entrants.htm

Anyone have any opinions about why the Week One events have fewer entrants than the other events on offer?

It's by no means obvious. Perhaps you need to survey the players staying for one week only, why they chose the second week in preference to the first. It's not unknown for players in the grading restricted events to play those in the morning and the pm open in the afternoon. The numbers there are a bit second week weighted, but not dramatically so. Thus you have U180 and U140 in the first week with 24 and 45, whilst the U160 and U120 in the second week have 54 and 34.

It might be that the British is expected to be more dramatic in the second week, but with one of the biggest entries being the second week pm open, players in that aren't going to get that many spectator chances.

[Stewart Reuben]

E Michael White >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 ?<

The latter point has been made many times. I believe a questionnaire should go out to people on the ECF website, not just to the contestants in the congress. But the pairing system for the championship is rather esoteric and would not be of interest to the general chess public. The players qualified to play would be a more useful group to consult for tht particular matter. No doubt many have differing opinions on the qualifying regulations as well.

People have mainly done their own valuation of pairing systems. That suffers from the usual problem that the architect of the system wants to prove his system was best. Probably Almog Burstein has done most work on his system. But as I have said before, first you must determine what your objectives are. Naturally different people will have different objectives. Thus Almog wants the system to be as little affected by pre-established ratings as possible. A quite different position from Dubov who places great reliance on rating in his system.

The recent Kaggle Competition for the 'best rating system' could also be done for Swiss Pairings. But there 'best' was defined.

People like to be able to predict their next round opponent. Using a random number generator and having the computer do the pairings does not address that particular non-scientific preference.

Nobody has raised the issue that the Swiss Pairing System is flawed by definition. Players who have more whites than blacks will tend to have a higher score. It is about 51.5-48.5%. Thus players in the top half are more likely to be black-seekers and those in the bottom half white-seekers. Normally more people in the top half will get 6 whites rather than 5. If you have an even number of rounds, as in Gibraltar, inevitably a few people will get a 6:4 distribution.

As far as I know nobody has investigated a pairing system where players can play more than once. Thus Sheffield 2011. Adams and Short are lone leaders with 4.5/5. Why shouldn't they again play, preferably with reversed colours?

Adam Raoof asks why there are usually more entrants for the second week than the first? I suspect this is because people want to see the conclusion of the Championship. After all more people are interested in the later stages of Wimbledon than the earlier. Of course you could ask the contestants.

Stewart Reuben

[Richard Cowan]

As far as I know nobody has investigated a pairing system where players can play more than once.

Yes they have. See: -

http://www.jbfsoftware.com/joomla15/ind ... &Itemid=55

[Stewart Reuben]

Richard I've never come across that. Thank you. But is it a Swiss? A bit more like a ladder system.

The London Schools League when I played scored on Sonneborn-Berger. The top 4 teams then met in a knockout. You could choose different parameters for your preliminary opponents. We always sought the strongest, others chose opponents close by, Still others wanted to play weak opposition. Of course you could play a team in the preliminaries and the knockout.
My school always won. So it was a good system, but not a Swiss.

Another disadvantage for all tournaments is that Michael Adam has a built-in advantage this year. He doesn't have to play himself and thus may meet a weaker field.

Stewart Reuben

[Adam Raoof]

As far as I know nobody has investigated a pairing system where players can play more than once.

Yes they have. See: -

http://www.jbfsoftware.com/joomla15/ind ... &Itemid=55

Thanks Richard - this looks like the ideal pairing system for our club tournament in September! However it doesn't take account of those players who always insist on arriving at the club late every week....

[Dan O'Dowd]

That Keizer system looks horrid! It totally overpunishes lower ranked players (especially in tourns with more players), and seems to be giving a large in-built advantage to the higher players!

Scrabble uses something called Chew pairing for the World Championships and other tournaments. This may prove useful where your number of rounds is close to the number of participants, or it can allow it to become larger. http://www.poslarchive.com/math/softwar ... .html#chew

Some tenets of the system are that repeat pairings are, in some situations, wholly valid and encouraged, and that clear tournament winners are paired lower to avoid effecting ranking of runners up.

For any of our major tourns (British etc) I'm sure this wouldn't be practicable unless we suddenly decided to become the 1940/1950s and have shintysix rounds in every tournament, between very few players )

Has it ever been considered to change the format of the Championship proper in any way, to either reduce the field, or perhaps to hold a separate elite level super tournament of a different format, avoiding said pairing problems? Perhaps a Champion of Champions every four to six years would be a nice addition.

[Stewart Reuben]

You may not be aware that the British Championship was a round robin of 12 players until some time in the 1950s.
Then it became a Swiss for not more than 32 players.
Then Leonard Barden and Richard Beville said to me that the specific number was wrong, it should be however many were good enough. That resulted in expansion when the top of British chess was very strong.
We enacted positive discrimination in favour of females in about 1983.
When Neil Graham took over he lowered the levels.
And here we are. Everybody has qualified according to the regulations. There are 85 players and, if that remains true, another will be added on Monday.
I think there is something to be said for letting the Commonwealth players in for only 2013, the 100th British Chess Championships. Securing the agreement of the Commonwealth Chess Association that this be the Commonwealth Chess Championship. That the winner in each category receive that title. That the highest placed British Isles player receive the British title and, if he wins, also the Commonwealth one. This to apply to all the titles, even Under 8.
But it would require the agreement of the ECF, the players, the CCA and the British Isles Coordinating Committee.
Anybody can make their own suggestions. I remember when the Southern Irish were allowed in. I think 1988.
Rules aren't fixed forever.
Stewart Reuben

[Paul Dargan]

@ Adam - Kaizer system was used at Leiden for he 120-player club championship. Worked very well. They had a slight tweak that they checked who was leading at week 20 (half-way) and at the end. If you were top of the table in both then fairenough - you win. If not then the top 2 at the mid-point and end-of-year played a 4-player All Play All to determine the winner.

One of many nice touches in what was a very well-run club (and one of 7!! in a not overly large University town)

Paul

[Ian Jamieson]

Adam Raoof asked "Anyone have any opinions about why the Week One events have fewer entrants than the other events on offer?"

I was inspired to enter a congress over the summer rather than wait until September before playing again when a clubmate qualified for the Championship but I decided to play in the Weekend Open at the Scottish championships rather than play at the British Championships.

The main reason was location. Edinburgh was more appealing to me than Sheffield although I may be biased as I used to live in Edinburgh.

As Stewart Reuben suggested another reason was I was able to see the conclusion of the Championship while I was there - the weekend open was held on the final weekend of the Scottish Championship while if I had played at the British I would probably have played in the U180 which is the first week of the British.

Other reasons were cost and transport links. It was cheaper to play in the weekend open than the U180 partly because of VAT and partly because I did not have to be a CS member. I discovered there is an overnight National Express service from Stoke, where I live, to and from Edinburgh and National Express were running a summer promotion so it was both cheaper and easier to get from Stoke to Edinburgh than to Sheffield even though Sheffield is considerably closer.

Availabilty of cheap accommodation was also a factor. Even though I only decided to play a month before the event I had no problem finding somewhere to stay. My clubmate who qualified for the British has deferred his entry for a year because he couldn't find somewhere to stay for the whole fortnight in Sheffield.

[Ian Jamieson]

Oh I forgot.

I was also able to see Bobby Fischer against the World while waiting on the Sunday night for my coach back down to Stoke.

[Stewart Reuben]

I am now in Sheffield. The venue seems superb (you can't know until playing for some time). My hotel is full, not just because of chess, but also swimmers. That led to a fascinating conversation this evening comparing children representing England in the two very different sports. The weather became good. The town in the immediate vicinity of Ponds Forge is bustling and lively.
David Welch and I are very pleased.
The numbers of entries will increase until Monday 1 Auust. Do come.
Stewart Reuben

[David Gilbert]

In all the excitement it almost went un-noticed that the number of entries reached 898 today, overtaking Canterbury. It's now a forgone conclusion that Sheffield will exceed 900 enteries by the weekend. A fabulous job done by everyone involved - it is already the sixth most well attended Bristish Championship of the 21st century and maybe, just maybe, a late surge could take that figure beyond that 2009 Torquay:

Edinburgh (2003) 1009
Torquay (2002) 973
Scarborough (2001) 960
Scarborough (2004) 940
Torquay (2009) 935
SHEFFIELD (2011) 898
Canterbury (2010) 891
Great Yarmouth (2007) 824
Swansea (2006) 768
Liverpool (2008) 742
Millfield (2000) 711
Isle of Man (2005) 503