Swiss Pairing Systems

Historical knowledge and information regarding our great game.
Roger de Coverly
Posts: 18057
Joined: Tue Apr 15, 2008 2:51 pm

Swiss Pairing Systems

Post by Roger de Coverly » Mon Jun 24, 2019 11:42 am

I was wondering when these were first introduced. The British Championship of 1949 was the first "British" to use such pairings.

According to a wiki article
https://en.wikipedia.org/wiki/Swiss-system_tournament
the first example was
The system was first employed at a chess tournament in Zurich in 1895, hence the name "Swiss system".
It also claims
The first national event in the United States to use the Swiss system was in Corpus Christi, Texas in 1945;
Any known events between 1895 and 1945?

Angus French
Posts: 1589
Joined: Thu May 15, 2008 1:37 am
Contact:

Re: Swiss Pairing Systems

Post by Angus French » Mon Jun 24, 2019 3:31 pm

Interesting Wikipedia page.

It often strikes me that the rules for Swiss pairings - or at least the ones in use in ECF-graded events - are significantly complex and I wonder if there couldn't be simpler, more elegant, designs which are easier to understand and code/operate? Has anyone written a thesis on pairing systems?

Roger de Coverly
Posts: 18057
Joined: Tue Apr 15, 2008 2:51 pm

Re: Swiss Pairing Systems

Post by Roger de Coverly » Mon Jun 24, 2019 4:08 pm

Angus French wrote:
Mon Jun 24, 2019 3:31 pm
I wonder if there couldn't be simpler, more elegant, designs which are easier to understand and code/operate?
They have to be deterministic to avoid accusations of arbiter bias, hence the complicated rules. I'd suspect that "random" pairings driven by a pseudo random seed would be simpler and more elegant. The historic development in the UK was the arbiters adopted semi deterministic methods some years before anyone attempted to use a computer engine for pairings. That was long after computer engines could play at a reasonable or strong level.

Mick Norris
Posts: 7429
Joined: Tue Apr 17, 2007 10:12 am
Location: Bolton, Greater Manchester
Contact:

Re: Swiss Pairing Systems

Post by Mick Norris » Mon Jun 24, 2019 4:33 pm

Angus French wrote:
Mon Jun 24, 2019 3:31 pm
Interesting Wikipedia page.

It often strikes me that the rules for Swiss pairings - or at least the ones in use in ECF-graded events - are significantly complex and I wonder if there couldn't be simpler, more elegant, designs which are easier to understand and code/operate? Has anyone written a thesis on pairing systems?
Pennsylvania not quite what you may be looking for? It references Weighted matchings in chess tournaments which has
In many chess tournaments, e.g. when the Swiss system is used, the number of players is much larger than the number of rounds to be played. In such tournaments the pairing for a round depends on the results in earlier rounds, and the pairing process can be very complicated. In these pairing systems the main goals are to let players with equal scores play together, and that each player should alternately play white and black, with the restriction that no player may face the same opponent more than once. The paper describes how a weighted matching algorithm is used to find ‘the best pairing’ by converting the pairing rules into penalty points.
Last edited by Mick Norris on Mon Jun 24, 2019 4:39 pm, edited 2 times in total.
Any postings on here represent my personal views and should not be taken as representative of the Manchester Chess Federation www.manchesterchess.co.uk

Kevin Thurlow
Posts: 2932
Joined: Wed Apr 30, 2008 12:28 pm

Re: Swiss Pairing Systems

Post by Kevin Thurlow » Mon Jun 24, 2019 4:36 pm

Ask Stewart Reuben!

Other games have used Swiss systems as well (Bridge and Go, I think.) Definitely, independent "Countdown" tournaments and even a colleague at work ran a snooker tournament on a Swiss system, which confused everybody except the two chess players in it. The same guy Elo-rated the players as well!

John McKenna
Posts: 3712
Joined: Tue May 17, 2011 2:02 pm

Re: Swiss Pairing Systems

Post by John McKenna » Tue Jun 25, 2019 8:27 am

Swiss pairings have been used in the Olympiad from 1976 on.
To find a for(u)m that accommodates the mess, that is the task of the artist now. (Samuel Beckett)

Angus French
Posts: 1589
Joined: Thu May 15, 2008 1:37 am
Contact:

Re: Swiss Pairing Systems

Post by Angus French » Tue Jun 25, 2019 8:55 am

Mick Norris wrote:
Mon Jun 24, 2019 4:33 pm
Angus French wrote:
Mon Jun 24, 2019 3:31 pm
Interesting Wikipedia page.

It often strikes me that the rules for Swiss pairings - or at least the ones in use in ECF-graded events - are significantly complex and I wonder if there couldn't be simpler, more elegant, designs which are easier to understand and code/operate? Has anyone written a thesis on pairing systems?
Pennsylvania not quite what you may be looking for? It references Weighted matchings in chess tournaments which has
In many chess tournaments, e.g. when the Swiss system is used, the number of players is much larger than the number of rounds to be played. In such tournaments the pairing for a round depends on the results in earlier rounds, and the pairing process can be very complicated. In these pairing systems the main goals are to let players with equal scores play together, and that each player should alternately play white and black, with the restriction that no player may face the same opponent more than once. The paper describes how a weighted matching algorithm is used to find ‘the best pairing’ by converting the pairing rules into penalty points.
Thanks Mick, I'll take a look at Christopher Hua's (University of Pennsylvania) paper. I've also found a shorter paper.

Roger de Coverly
Posts: 18057
Joined: Tue Apr 15, 2008 2:51 pm

Re: Swiss Pairing Systems

Post by Roger de Coverly » Tue Jun 25, 2019 9:47 am

Angus French wrote:
Tue Jun 25, 2019 8:55 am
I've also found a shorter paper.
That quotes 1942 as the first usage in the USA.

The BCF seems unlikely to have leapt totally into the unknown in the British Championship, so are there UK examples earlier than 1949?

Mick Norris
Posts: 7429
Joined: Tue Apr 17, 2007 10:12 am
Location: Bolton, Greater Manchester
Contact:

Re: Swiss Pairing Systems

Post by Mick Norris » Tue Jun 25, 2019 9:52 am

Angus French wrote:
Tue Jun 25, 2019 8:55 am
I've also found a shorter paper
Thanks Angus; the notes reference the 2015 paper by Csato, which itself has a large number of references (not least half a dozen former works by Csato), so there's obviously quite a bit of research been done
Any postings on here represent my personal views and should not be taken as representative of the Manchester Chess Federation www.manchesterchess.co.uk

Nick Grey
Posts: 1148
Joined: Thu Dec 01, 2011 12:16 am

Re: Swiss Pairing Systems

Post by Nick Grey » Tue Jun 25, 2019 12:03 pm

As the Swiss system is a 19th century invention it may well have been used earlier. I suggest looking in Europe. I was surprised of a mention of a dutch swiss system.

Roger de Coverly
Posts: 18057
Joined: Tue Apr 15, 2008 2:51 pm

Re: Swiss Pairing Systems

Post by Roger de Coverly » Tue Jun 25, 2019 12:13 pm

Nick Grey wrote:
Tue Jun 25, 2019 12:03 pm
I was surprised of a mention of a dutch swiss system.
That's a dutch pairing system which is now the most popular official FIDE version.

E Michael White
Posts: 1333
Joined: Fri Jun 01, 2007 6:31 pm

Re: Swiss Pairing Systems

Post by E Michael White » Tue Jun 25, 2019 12:39 pm

Is it not the case that, excluding random pairings, most if not all, ECF and FIDE approved pairing methods will produce the same first round pairings for the same group of players whether the tournament is 5 or 6 rounds? Information and probability theory would say otherwise if the objectives are the same.

User avatar
IM Jack Rudd
Posts: 3914
Joined: Tue Apr 17, 2007 1:13 am
Location: Bideford

Re: Swiss Pairing Systems

Post by IM Jack Rudd » Tue Jun 25, 2019 4:01 pm

E Michael White wrote:
Tue Jun 25, 2019 12:39 pm
Is it not the case that, excluding random pairings, most if not all, ECF and FIDE approved pairing methods will produce the same first round pairings for the same group of players whether the tournament is 5 or 6 rounds? Information and probability theory would say otherwise if the objectives are the same.
It's only the case for an even number of players. For an odd number of players, they'll produce different pairings, because CAA pairings will give the bye to the median and FIDE pairings will give the bye to the bottom seed.

E Michael White
Posts: 1333
Joined: Fri Jun 01, 2007 6:31 pm

Re: Swiss Pairing Systems

Post by E Michael White » Tue Jun 25, 2019 4:28 pm

ah what I meant was that 5v6 rounders would produce the same pairings for either ECF or FIDE, not necessarily ECF=FIDE.
Last edited by E Michael White on Wed Jun 26, 2019 12:12 pm, edited 1 time in total.

Roger de Coverly
Posts: 18057
Joined: Tue Apr 15, 2008 2:51 pm

Re: Swiss Pairing Systems

Post by Roger de Coverly » Tue Jun 25, 2019 5:03 pm

IM Jack Rudd wrote:
Tue Jun 25, 2019 4:01 pm
FIDE pairings will give the bye to the bottom seed.
In practice that's not a very satisfactory way of doing it, as the same player will get a bye in consecutive tournaments by virtue of being the bottom seed either by rating or even worse by alphabet.

Post Reply