Are Open Tournaments really interesting?

The act of developing or disclosing that which is unknown.
Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Are Open Tournaments really interesting?

Post by Roger de Coverly » Wed May 02, 2012 6:06 pm

John McKenna wrote: Just one caveat though - in another thread it has come to light that pairings in some events, held during school holidays, in the British Isles may have been 'tweeked' to accommodate groups of juniors who, it was decided, would benefit from not being paired together if possible! Fair/unfair?
If an event is billed as an individual tournament I don't think account should be taken of whether any of the participants form a "squad". At best you might offer entrants an option not to be paired against certain players, but this should be offered to all. That does explain why, one year at Hastings, with a newly acquired international rating, I was paired for the first eight rounds against unrated players, even though around half the players, mostly the top half, had ratings. I'd always suspected pairing manipulation so as to get rated v unrated and might have benefited myself on the way to a rating, but it seems there was an adult v Junior Squad bias as well.

By contrast, in these International Junior events, the players are representing their Federations and measures to reduce the incidence of squad v squad would seem desirable, unlikely as it might be that FIDE would ever be minded to adopt them.

Sean Hewitt
Posts: 2193
Joined: Sat Mar 10, 2012 8:18 pm
Contact:

Re: Are Open Tournaments really interesting?

Post by Sean Hewitt » Wed May 02, 2012 8:29 pm

John McKenna wrote:... in another thread it has come to light that pairings in some events, held during school holidays, in the British Isles may have been 'tweeked' to accommodate groups of juniors who, it was decided, would benefit from not being paired together if possible! Fair/unfair?
Unfair. Would not happen in an e2e4 event.

Alex McFarlane
Posts: 1757
Joined: Sat Aug 02, 2008 8:52 pm

Re: Are Open Tournaments really interesting?

Post by Alex McFarlane » Wed May 02, 2012 9:50 pm

Most British arbiters would alter a round 1 draw to avoid players from the same club/country playing.

In later rounds I would look to see if such pairings could be avoid only if the players to be swapped with are of similar grades/rating.
I would be more flexible at the bottom of the draw than near the top.
If a request to avoid a pairing was received and no simple swap was possible (even if it was, the swap would not be automatic) then it might happen only if the players receiving the 'favour' ended with stronger opposition and the other two players therefore got easier opponents in grading terms. I would not consider pairing two top half players together to avoid a pairing.

In junior events I will note on the pairing card where a clash has occurred between schoolmates. I will try to keep such encounters to a maximum of two in a 5 or 6 round event.

In general terms, unless players are of identical grades, I will not change a pairing where the players are able to win a significant prize.

Alex Holowczak
Posts: 9085
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: Are Open Tournaments really interesting?

Post by Alex Holowczak » Wed May 02, 2012 10:59 pm

We paired two brothers against each other in round 2 of the Birmingham Rapidplay this year, but they happily played their game against each other. Come round 3 though, and another pair of brothers were paired against each other, and they agreed a draw very quickly. Two further brothers played in round 4, and the elder brother complained about it, but the precedent had been set by what happened in round 2.

Through the whole event, we made no effort to avoid pairing players from the same club. I had previously noticed that some entrants, when declaring their club, were declaring clubs so that they could either avoid playing certain people, or declaring clubs so that they could still play certain people. For example, one entrant entered under his school's club, so that he could still play people from his adult club. So the whole concept was being abused.

We were pairing according to the FIDE pairing rules, so that a computer could produce the pairings to make life easier. The FIDE pairing rules don't say anything about avoiding players from the same club or family, so the computer didn't do it. Based on that, the only pairing I'd probably avoid now is a pairing between an Israeli and a Yemeni (for example), on the basis that the Yemeni would have to default the game for political reasons.

Alex McFarlane
Posts: 1757
Joined: Sat Aug 02, 2008 8:52 pm

Re: Are Open Tournaments really interesting?

Post by Alex McFarlane » Wed May 02, 2012 11:21 pm

Swiss Master, and I believe Swiss Manager, the commonly used FIDE approved pairing programs allows you to specify forbidden pairings. So FIDE obviously doesn't object to the idea.

I have never used that setting in any tournament however.

Alex Holowczak
Posts: 9085
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: Are Open Tournaments really interesting?

Post by Alex Holowczak » Thu May 03, 2012 12:07 am

Alex McFarlane wrote:Swiss Master, and I believe Swiss Manager, the commonly used FIDE approved pairing programs allows you to specify forbidden pairings. So FIDE obviously doesn't object to the idea.

I have never used that setting in any tournament however.
Swiss Master comes with an example from an Aeroflot tournament that uses the forbidden pairing function involving an Israeli. It wasn't used for any other purpose, although I haven't looked to see if it could have been. So I think the functionality is present in the software for that purpose, rather than for any other reasons.

The FIDE handbook says that "The pairing system used for a FIDE rated tournament shall be either one of the published FIDE Swiss Systems or a detailed written description of the rules shall be handed over to the participants."

There's nothing in the pairing systems that I can find about avoiding players from the same club. If FIDE were OK with this, they'd say so in their rules. I'm not sure the CAA discretionary rule counts as a "detailed written description", because there's too much "make it up as you go along" about it.

All that said, I'm sure that a key reason for the FIDE pairing rules being written as they are is for the ease of programming the algorithm for computer pairing software.

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Are Open Tournaments really interesting?

Post by Matthew Turner » Thu May 03, 2012 7:36 am

Most people were happy with the idea of avoiding players from the same club, country, family playing each other. Then the rules of the FIDE junior Championships changed, it was no longer just one player from each nation, but multiple entries. Often three, four or five of the top players in each section came from Russia. Avoiding players from the same Federation meeting would hand the Russians a massive advantage and was patently unfair. Whether implicit or explicit FIDE had to drop the idea of avoiding players from the same federation playing each other. Of course Russia are no longer the dominant force of decades past, but the same principles would still apply.

MSoszynski
Posts: 262
Joined: Thu Apr 05, 2007 4:43 pm

Re: Are Open Tournaments really interesting?

Post by MSoszynski » Thu May 03, 2012 7:49 am

Alex Holowczak wrote:Through the whole event, we made no effort to avoid pairing players from the same club. I had previously noticed that some entrants, when declaring their club, were declaring clubs so that they could either avoid playing certain people, or declaring clubs so that they could still play certain people. For example, one entrant entered under his school's club, so that he could still play people from his adult club. So the whole concept was being abused.
Why is this an abuse? A player's club(s) is recorded on the ECF database, so if the club name shouldn't be there then the data is wrong, and if the player filled in the entry form deceitfully, then that is wrong too. However, maybe a player chose to enter under one of his genuine clubs rather than another in order to qualify for a team prize.

It is very frustrating to travel hours in a car with one's clubmates to a distant congress only to be paired against one of them in the very first round. Speaking from personal experience, but not only from that, these pairings often result in contrived games. Crazy openings are played and draws agreed even when it is mate on the move. Pairing players of the same club together greatly increases the possibility of collusion and the result being determined away from the chessboard. The suggestion that avoiding such pairings at least in the first round is surely a reasonable compromise. Obviously, any two players might collude to e.g. share prize money, but what I'm mainly talking about here is enhancing amateurs' enjoyment of tournament chess. If you don't have to pair clubmates together, why do it? If a player declares one club rather than another then he has done it for a reason which should be respected. As long as he is a genuine member of the declared club it is not an abuse.

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Are Open Tournaments really interesting?

Post by Matthew Turner » Thu May 03, 2012 8:30 am

"but what I'm mainly talking about here is enhancing amateurs' enjoyment of tournament chess"

I like your thinking.

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Are Open Tournaments really interesting?

Post by Roger de Coverly » Thu May 03, 2012 8:50 am

MSoszynski wrote: The suggestion that avoiding such pairings at least in the first round is surely a reasonable compromise.
Isn't this something the Manager of Congress Chess should have an opinion on, or at least try to find out current practice and whether it's supported by those who play in Congresses?

More generally

(a) Should it be the expectation that computer pairings are used ?
(b) Should there be "manipulations" to avoid club v club or squad v squad games? It can depend on the context. If an event is styled as the "xxx championship", it would be reasonable to expect that players from clubs in xxx can meet, even in the first round.

Most (all?) computer pairings schemes don't use grades or ratings directly, rather than as a means to an end of generating seeding orders. So by manipulating the seeding order you can avoid first round clashes and give all members of the same family the same colour.

Alex Holowczak
Posts: 9085
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: Are Open Tournaments really interesting?

Post by Alex Holowczak » Thu May 03, 2012 9:35 am

MSoszynski wrote:
Alex Holowczak wrote:Through the whole event, we made no effort to avoid pairing players from the same club. I had previously noticed that some entrants, when declaring their club, were declaring clubs so that they could either avoid playing certain people, or declaring clubs so that they could still play certain people. For example, one entrant entered under his school's club, so that he could still play people from his adult club. So the whole concept was being abused.
Why is this an abuse? A player's club(s) is recorded on the ECF database, so if the club name shouldn't be there then the data is wrong, and if the player filled in the entry form deceitfully, then that is wrong too. However, maybe a player chose to enter under one of his genuine clubs rather than another in order to qualify for a team prize.
For example, if I played in a tournament, I'd be kept apart on occasion from players who played for West Bromwich, Warley Quinborne and Birmingham Checkmates, because I play for all three. In some sections of local events, this can be a significant chunk of the entries in a section.
MSoszynski wrote:If you don't have to pair clubmates together, why do it?
But you do, because the FIDE pairing rules say so, and one of our sections is FIDE rated, so we have to use their pairing rules. It would be odd to pair different sections of the same event to different pairing rules.

We say on the entry form that Swiss Master is being used to do the pairings, so to then undo the pairings it generates would be breaking our own rule. And we have to use Swiss Master, because the organiser isn't prepared to organise the event if he has to use pairing cards, because he's the sod who'll have to spend a whole day writing them up, and a whole day after the event doing the grading and rating. That's time he's not prepared to commit, particularly when a much more simple solution exists that removes the need for both jobs.

Sean Hewitt
Posts: 2193
Joined: Sat Mar 10, 2012 8:18 pm
Contact:

Re: Are Open Tournaments really interesting?

Post by Sean Hewitt » Thu May 03, 2012 10:00 am

Just to be clear. The FIDE rules do not provide for keeping anyone apart but they do allow wide discretion as long as detailed written description of the rules are provided.

So by all means, keep family members, club members etc apart in round 1, early rounds, players on less than 50% - whatever you like.

But you need to document your rules in a detailed way and publish.

This ensures a level playing field, makes sure everyone knows the rules, and should help to avoid arguments about the legitimacy of the resulting pairings.

Alex Holowczak
Posts: 9085
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: Are Open Tournaments really interesting?

Post by Alex Holowczak » Thu May 03, 2012 10:06 am

Sean Hewitt wrote:Just to be clear. The FIDE rules do not provide for keeping anyone apart but they do allow wide discretion as long as detailed written description of the rules are provided.

So by all means, keep family members, club members etc apart in round 1, early rounds, players on less than 50% - whatever you like.

But you need to document your rules in a detailed way and publish.

This ensures a level playing field, makes sure everyone knows the rules, and should help to avoid arguments about the legitimacy of the resulting pairings.
This is a detailed written description I'm not prepared to write, and indeed, one that the CAA pairing system doesn't seem to prescribe either. So I avoid it!

Alex McFarlane
Posts: 1757
Joined: Sat Aug 02, 2008 8:52 pm

Re: Are Open Tournaments really interesting?

Post by Alex McFarlane » Thu May 03, 2012 10:09 am

For clarity, in a Championship event I would (with possible first round exception) pair people regardless of club, etc.

In other tournaments I would be more flexible.
The problem with computers is that they tend to pair by ranking rather than grade. This would mean that the order set up for round 1 would apply to all rounds unless the program was subsequently re-ordered. Not a major task but one which is easily forgotten (and impossible if manual wallcharts are being used!).
We have already established that Swiss Master does not always do as it should. It has been programmed to follow FIDE procedures from B onwards. Unfortunately this means that it can do stupid things such as floating two people down from a score group that it cannot pair rather than change the downfloat into that group.
In the FIDE system if we had joint leaders A (colourBWBBWW) and B (-BWBWW) then the program will not pair AvB but will instead downfloat both players, even if this a full point downfloat. At the bottom it would upfloat both players. (B has first round bye but could have had either colour)

I think that the British rules should be checked for clarity and then submitted to FIDE for acceptance.

As Sean has said, provided you explain to FIDE the pairing system used (I usually say British Pairing System and have never had it queried) then you don't have a problem. FIDE are only really interested in whether you are changing the draw to benefit players seeking norms or trying to raise their rating artificially. Provided you are doing neither of these things then you should have no problems.

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Are Open Tournaments really interesting?

Post by Roger de Coverly » Thu May 03, 2012 10:22 am

Alex McFarlane wrote: In the FIDE system if we had joint leaders A (colourBWBBWW) and B (-BWBWW) then the program will not pair AvB but will instead downfloat both players, even if this a full point downfloat. At the bottom it would upfloat both players. (B has first round bye but could have had either colour).
If I understand it correctly, the FIDE system has two absolute rules
(1) that you never play the same person twice
(2) that you never get the same colour three times in a row.
The British system doesn't have rule (2).

I would be inclined to agree with the FIDE position that 3*colour should override playing the same score as yourself. The most likely scenarios are that you award a tournament leader a third white in a row, or you award a tournament back marker a third black. I know the three blacks issue arose in a 5 day tournament at the British. Has the three whites issue for the leader ever arisen?

Locked