British Chess Championship Pairings Question

[Brian Beattie]

and I would have anticipated some intervention.

The problem is that, unless I have missed something, neither you nor anyone else has come up with an idea as to how the pairings might have been different.

I was hesitant to make any recommendation earlier today as I had was unaware of the intricacies of how the pairing system worked. Knowing a little bit more (albeit not much) both children could have played the next opponent the next score down.

[NickFaulks]

both children could have played the next opponent the next score down.

You are correct that this would be the only solution. However, it goes against everything that is sacred in Swiss pairing systems - doubly so, I would think, in the final round.

This question of scores v colours does in fact have some history. It is a little known fact that Swiss pairings lead to the higher rated players getting more than their share of Whites. This can be demonstrated theoretically and a very large scale analysis of German tournaments backed it up. I estimated the effect on the rating system and concluded that it might stretch the rating difference between two players by something of the order of 1 to 2 per cent over ten years, which didn't seem worth bothering about.

I tried to raise this issue, just as a talking point, with members of FIDE's Swiss Pairings Commission, but they refused to discuss it on the grounds that it was inconceivable. It is of course the opposite, being a natural by-product of prioritising scores over colours. You can't optimise both at once!

[Ian Thompson]

and I would have anticipated some intervention.

The problem is that, unless I have missed something, neither you nor anyone else has come up with an idea as to how the pairings might have been different.

I was hesitant to make any recommendation earlier today as I had was unaware of the intricacies of how the pairing system worked. Knowing a little bit more (albeit not much) both children could have played the next opponent the next score down.

If it hadn't been the last round, or the players hadn't been doing very well in the tournament, something like that would have happened. For better or worse, the pairing rules say that in the last round playing someone on the same score as you is more important than avoiding three games in a row with the same colour. See Basic rules for Swiss Systems and FIDE (Dutch) System Pairing Rules.

That's understandable because it wouldn't be very desirable for two joint leaders both to get downfloats in the last round and then finish the tournament tied for first place not having played each other.

There might be a case for tweaking the pairing rules though, and changing "Topscorers are players who have a score of over 50% of the maximum possible score when pairing the final round of the tournament." to something like "Topscorers are players who are in with a chance of winning a prize". That wouldn't have made any difference in this case though because your son could have finished in equal second place if he'd won and other results had gone his way.

[Brian Beattie]

The problem is that, unless I have missed something, neither you nor anyone else has come up with an idea as to how the pairings might have been different.

I was hesitant to make any recommendation earlier today as I had was unaware of the intricacies of how the pairing system worked. Knowing a little bit more (albeit not much) both children could have played the next opponent the next score down.

If it hadn't been the last round, or the players hadn't been doing very well in the tournament, something like that would have happened. For better or worse, the pairing rules say that in the last round playing someone on the same score as you is more important than avoiding three games in a row with the same colour. See Basic rules for Swiss Systems and FIDE (Dutch) System Pairing Rules.

That's understandable because it wouldn't be very desirable for two joint leaders both to get downfloats in the last round and then finish the tournament tied for first place not having played each other.

There might be a case for tweaking the pairing rules though, and changing "Topscorers are players who have a score of over 50% of the maximum possible score when pairing the final round of the tournament." to something like "Topscorers are players who are in with a chance of winning a prize". That wouldn't have made any difference in this case though because your son could have finished in equal second place if he'd won and other results had gone his way.

This makes sense. I admit I wasn't aware that under the right circumstances that the result of the match could place the winner into 2nd or 3rd place. I hadn't bothered to make any calculations as it seemed most logical they were playing for at best somewhere around 5th which is around the placing my child was targeting from the start. But when you end up 3 whites and 6 blacks losing out the colour pairings based on FIDE ranking twice it doesn't help matters. That coupled with the fact that since it was FIDE ranked insted of ECF ranked (like most junior tournaments in the UK) he ended up playing top players as soon as R2 instead of R3.

It makes me question though why the British does not include a tiebreak system. With a 9 round tournament with only 60 kids there are bound to be colour issues at some point in the later if not the latest round. It would make sense to at least reward the player with better tournament performance which a tiebreak system at least attempts to quantify. It is certainly more palatable to accept even if it goes against you. As it is if my child were to perform in a similar fashion next year as to this year there is a good probability he will end with 4 whites and 5 blacks even if he starts with white. That is unless I go for a Pay-To-Win approach spending a bunch of time, effort and money to ensure he competes in enough FIDE events to get a more appropriate rating. But that is not a path I am going to be taking and my child will simply have to decide whether playing in what is supposed to be the premiere flagship event is worth doing if it means he will likely be taking what might end up around a 0.5 point penalty to do so.

[Alex McFarlane]

Oh dear, where do I start with this one.
Starting with white means that you are more likely to end up with 5 whites and 4 blacks rather than 4 whites and 5 blacks. Your son, if the colours had been reversed, would have ended up with 3 whites in the last round. The higher rated does not get white, they get the colour they are most due. Your son was very unlucky in this case. He was not victimised by the system. Your feeling that he was will not help him to continue playing chess. Perhaps you could take an arbiter course and help out at his events. You might then understand the situation more clearly.


Can I also point out that, with the exception of a few tournaments in Wales, ECF rated tournaments are only held in ENGLAND. The statement, even if applied only to England, is bound to be correct as all FIDE rated events will also be rated by the ECF.

[Brian Beattie]

Oh dear, where do I start with this one.
Your son, if the colours had been reversed, would have ended up with 3 whites in the last round.

If my son had 6 whites and 3 blacks I would be equally disappointed in how the system works. And my son would have to take his results with a huge grain of salt. As such I’d still be here complaining about and asking how this happens and if there is a better way.

Starting with white means that you are more likely to end up with 5 whites and 4 blacks rather than 4 whites and 5 blacks. The higher rated does not get white, they get the colour they are most due. Your son was very unlucky in this case. He was not victimised by the system. Your feeling that he was will not help him to continue playing chess.

Well, the most likely occurrence for a colour change would be R9. And you have essentially 4 scenarios: expect black get black; expect black get white; expect white get white; expect white get black. I’d argue the last one is the least favourable to deal with psychologically. And, if you have control over it based on a higher rank it can’t happen. So, the issue at its core is still whether the player themselves has an opportunity to influence their rank by play over the board at the tournament. So the question I have asked that still remains unanswered is why doesn’t the British have a tiebreak system.

Perhaps you could take an arbiter course and help out at his events. You might then understand the situation more clearly.

Since parents aren’t allowed in the same room as their children when playing chess I would certainly consider this a huge conflict of interest no matter how benevolent the intention. And arguably younger children would be better served in their parent is accessible to them between rounds.

Can I also point out that, with the exception of a few tournaments in Wales, ECF rated tournaments are only held in ENGLAND. The statement, even if applied only to England, is bound to be correct as all FIDE rated events will also be rated by the ECF.

What is the inherent problem with utilizing the higher rating of the two systems for initial ranking purposes? I mean surely this has been suggested before.

[Alex McFarlane]

6v 3 is extremely rare regarding colours. It was (partially) caused because of the size of his last round score group. Most players would welcome 6 whites! Especially since white has a slightly higher win rate.

I dispute your 4 scenarios for the last round as you cannot ignore all of the previous rounds. Regardless, in an odd numbered round you are more likely to get the colour you got in round one in the final round. that is the way the pairing system works. Your maths is flawed.

There are a large number of events where parents are involved. They are assigned to a section in which their child/children are not playing. . There is, therefore no conflict of interest.

Some events, for example the 4NCL Weekend Congresses did take the higher of the two. FIDE has made that less desirable by raising its minimum rating to 1400. Overnight a 1000 FIDE rated player became 1400. This has caused a problem for British (and here I do mean British) events which used to take the higher of the two if they have sections for players U2000. (The increase in rating only affected players under 2000.)
Therefore there is now an inherent problem.

Why did you not ask these questions at the event? They could have been easily answered.
I will not comment on the wisdom of ECF officers not being allowed to answer points made on this forum. In cases such as this I think it has more minuses than pluses.

[Brian Beattie]

...I dispute your 4 scenarios for the last round as you cannot ignore all of the previous rounds. Regardless, in an odd numbered round you are more likely to get the colour you got in round one in the final round. that is the way the pairing system works. Your maths is flawed...

I wasn't implying that there were 4 possibilities going into the last round but there are four possibilities at the start of the tournament. Some of which are more palatable if they occur. However, from what I can ascertain from what has been discussed above by the last round there will be two possibilities as you are going to be due a certain colour. I am willing to posit that in a 9 round tournament with 60 players that there is a good probability that someone due their colour won't get it. The question is whether the person who should get the colour they are due be the player with the higher FIDE rating or the one with higher tournament performance (as determined through tiebreaks). A player with an artificially low FIDE and particularly one who only plays a few tournaments per year is disincentivized in playing such a tournament because they have no agency over the outcome not only in the final round but in pairing decisions throughout each round. As such they will be disadvantaged more times than not.

There are a large number of events where parents are involved. They are assigned to a section in which their child/children are not playing. . There is, therefore no conflict of interest.

My children don't play all that many tournaments and to the extent they do all the sections are typically held in one big playing hall. And like I previously mentioned they are better suited having my around between rounds.

Some events, for example the 4NCL Weekend Congresses did take the higher of the two. FIDE has made that less desirable by raising its minimum rating to 1400. Overnight a 1000 FIDE rated player became 1400. This has caused a problem for British (and here I do mean British) events which used to take the higher of the two if they have sections for players U2000. (The increase in rating only affected players under 2000.)
Therefore there is now an inherent problem.

I could see how that could be a problem. Yet for the U12 British Sections they seem to have remedied it as there are plenty of players with ratings below 1400. So it doesn't explain why the higher of the two ratings couldn't be used for those above 1400.

Why did you not ask these questions at the event? They could have been easily answered.
I will not comment on the wisdom of ECF officers not being allowed to answer points made on this forum. In cases such as this I think it has more minuses than pluses.

None of this became readily apparent until R9 so asking at the event would be akin to closing the barn doors after the horse has fled. I am sure the tournament officials/arbiters had much better things to do then respond to inquiries which can have no immediate impact or change any outcome. It would be a waste of everyone’s time. I thought asking here would be the most logical next step.

[Alex McFarlane]

You keep saying that you should use tiebreaks to determine seeding each round. Which tie-break would you use? That would cause further argument. As I always say at arbiter seminars, there is no perfect tie-break other than a play-off over an even number of games. Every other tiebreak will disadvantage someone.

If you did volunteer at an event you would have a much better idea of what really happens. Many of your questions would be answered.

At the event is the time to ask questions, particularly if you think a pairing is wrong. If there had been a mistake, how can it be corrected retrospectively? You had several days of doubt over the last round pairing. By asking a simple question you would have had an answer within minutes.

As a further aside, the number of comments I get when someone gets a second black in a row is at about 30 times the number of comments for a second white in a row.

I repeat, if you have suggestions to make about the British, this forum is not the place to get it actioned. Contact the organisers or the ECF Office.

[Brian Beattie]

You keep saying that you should use tiebreaks to determine seeding each round. Which tie-break would you use? That would cause further argument. As I always say at arbiter seminars, there is no perfect tie-break other than a play-off over an even number of games. Every other tiebreak will disadvantage someone.

I'd be happy enough with a split FIDE/ECF rating and any Buchholz system. Had this been applied at the tournament the overall complications with round pairings and colour choice my son ran into would not have happened. And to the extent they would have potentially happened to anyone as best I could tell would have been somewhat related to tournament performance which I would argue is a fairer decider. Now I am happy to learn how this is incorrect but the answer seems to be that it is not something that can be learned on a chess forum but one has to attend an arbiters training course.

If you did volunteer at an event you would have a much better idea of what really happens. Many of your questions would be answered.

I doubt it would teach me how a FIDE ranked tournament is better than one with an ECF/FIDE ranking and a tiebreak system which is really the only question I have been asking.

At the event is the time to ask questions, particularly if you think a pairing is wrong. If there had been a mistake, how can it be corrected retrospectively? You had several days of doubt over the last round pairing. By asking a simple question you would have had an answer within minutes.

All I can say is my perspective is different. Once the pairing was listed it was time to go in and play chess. That's it. Everything else is irrelevant at that point and is a distraction.

As a further aside, the number of comments I get when someone gets a second black in a row is at about 30 times the number of comments for a second white in a row.

This is irrelevant as are your consistent snide remarks towards me in this regard.

I am happy to continue discussions of pairing systems with you because I have a lot to learn. But if you are intent on continuing to insert various logical fallacies into the discussion that are irrelevant to that discussion then I am going to bow out. If that is the case then you are free to have the last word if you so choose.

[Christopher Kreuzer]

Not to derail the current, er, discussion, but is there a case for anticipating such scenarios and doing the "double downfloat" (or upfloat) method to reduce the number of players that may face this scenario (have a double colour run and ending up with a triple colour run), in an earlier round, before the final one? Or can that create more problems than it solves?

EDIT: I should add I think I have seen this done in small tournaments that were probably following their own idiosyncratic pairing rules...

[Paul Cooksey]

Alex is undermining his point that this forum not the best place to ask questions by responding. I'm not sure if Brian realises this is like complaining about a handball decision at his son's football game and getting a response from Howard Webb.

The one area where I doubt I will agree with Alex, an eminent FIDE arbiter and Chairman of Chess Scotland, is in the use of English Ratings at the British Championship, It is an entertaining discussion though. If we exhaust it we could move onto the ECF using the Union Flag, whether Scotland have any promising juniors who might be interested in changing federation, and DCMS funding of the ECF.

As a player, I would have queried the pairing before sitting down at the board. Concur with the view at that point only the game matters. But occasionally mistakes occur. In my experience only where a mistake in recording has occurred in a previous round rather than a fault in the pairing itself. But it can happen.

I'd be extremely surprised if a mistake was made in the pairing methodology itself or if there were easy fixes to pairing systems, particularly those used by a wide variety of events worldwide. Arbiters really, really care about pairings and have done for decades. Possibly only the location of restaurants that are open late is more important to them.

As a veteran player, I've come to accept Opens are inherently a bit unfair. There are many more things that can go wrong than in an all play all. I'd certainly be telling my friends about a 3/6 colour sequence over diner and expecting sympathy. But I wouldn't really consider it anything more than bad luck equivalent to walking into preparation or similar.

[Roger de Coverly]

Not to derail the current, er, discussion, but is there a case for anticipating such scenarios and doing the "double downfloat" (or upfloat) method to reduce the number of players that may face this scenario (have a double colour run and ending up with a triple colour run), in an earlier round, before the final one? Or can that create more problems than it solves?

There might be a case for avoiding BB (or WW) sequences for leaders towrds the end of tournaments. Defining when to do it by applying excessive floats would no doubt be a problem. There also has to be a unique solution to faciliate computer pairings.

I seem to vaguely recall a similar issue coming up on the forum many, many years ago. That particular context was two players with 0/2 who had both had BB sequences. Should you pair them or float both? There were no other players on 0/2. That was so long ago that it was in the era of pairing cards and CAA rules.

As Paul notes, a nearby Wetherspoons with food orders until 11pm can be a consideration in venue selection.

[Jonathan Rogers]

I am aware that Mr Beattie came to this forum not looking for sympathy but for an explanation as to how his son could play three Blacks in a row at the end of a tournament. And he certainly has come to the right place for that - every one has answered with their own understanding of swiss pairing rules.

And if he had been looking for sympathy, then from what we have seen so far, he would certainly not have come to the right place.

For what it is worth though, I for one am highly sympathetic. Of course, having three Blacks in a row, and especially at the business end of the tournament, is a great disadvantage, and it is no consolation to learn that it is the chess equivalent of being struck by lightning, depending as it does on a unique set of numerous different results in the preceding rounds. Nor is it anything to the point that no one would complain about having three Whites in a row. When that happens, we can rest content that the standard swiss rules have been applied and that someone got lucky. It need not follow that we should similarly shrug our shoulders when someone has three Blacks. And it is ultimately a decision made by the arbiters (or required by the pairing rules) and not to be equated with the sort of bad luck which Paul mentions.

No doubt sometimes it really is necessary for someone to play three consecutive Blacks, and in fact that must be more of a possibility at the end of the tournament, by which times players are divided off into smaller clusters of scores and the options for avoiding it accordingly reduce. Still, I should think there is room for recognising that having three Blacks is a peculiar competitive disadvantage (which leaves the recipient feeling rather victimised to boot) and perhaps avoiding it should trump other pairing principles unless, say, when that player has a chance of finishing first or first equal and fairness to the other players competing for first place (as opposed to the sanctity of ordinary pairing rules) requires it. (Certainly, in Roger's example with two players on 0/2, float them! Life will continue!).

[Ian Thompson]

perhaps avoiding [having three Blacks in a row] should trump other pairing principles unless, say, when that player has a chance of finishing first or first equal and fairness to the other players competing for first place (as opposed to the sanctity of ordinary pairing rules) requires it. (Certainly, in Roger's example with two players on 0/2, float them! Life will continue!).

That's pretty close to what the pairing rules already say - see links above.

As it's presumably not the last round the two players on 0/2 would be floated instead of playing each other. Even if it was the last round they still wouldn't be paired against each other because they've scored less than 50%.

The pairing rules don't refer to finishing first as a criterion, but they do say the possibility of three games in a row with the same colour only applies to players on more than 50% which I assume is an attempt to do something similar (i.e. only do it when the result matters).