English Chess Forum

This may be a recommended solution, but is not one I feel particuarly comfortable with. I would prefer not to have the draw for all five rounds made by the start of round 1, with it known in advance which players are having three whites in the competition.

My approach would be to make the pairings in round 2 based on their first round results. However I would then have to make sure in doing the draw for round 3 that if player 1's last 2 opponents were player 2 and player 3, that player 2 and player 3 had already played each other. This would effectively mean that I would have to make the draw for rounds 4 and 5 at the same time as round 3, but after making the draw after 2 rounds have been played, I feel better able to assess the fairest allocation of colours.

When little is known about the playing strength of the players the numbers 1, 2, 3, 4, 5 and 6 may be fairly arbitary.

More posts have been added sinse I was compiling this, and I have also seen Alex's recommendation for 8 players. However I personally would find that solution a bit on the random side.

The way to think about it is the principle that: The idea of the Swiss system is to find a fair set of opponents for each player in the tournament in a transparent manner.



Taking the above principle, an All-Play-All generates the same set of opponents as a Swiss. So there's no need to make life difficult by pairing them, with the risk that you may reach round 5 with no set of legal pairings. You just toss a coin to see whether you do the pairings Jack showed or reverse them all, and off you go.



Again, taking the principle, if announced in advance, this system is transparent. In both cases, you play all opponents bar one. I think that if you choose this person randomly via the APA minus one trick, or via convoluted Swiss pairings, I don't see how the fairness of the opponent selection increases by choosing one or the other.

_________________
February 16-17: 2013 BUCA Team Championship
April 20: 2013 British Blitz Championship
July 27-August 10: 100th British Chess Championships

At least my original post related to downward floats and medians, and I think we both agree that these are not relevant here.

I understand the point Alex is making, but providing you avoid the scenario I described I cannot see that is at all difficult for the method I use to fail to complete the competition in five rounds.

There is perhaps no perfect solution for organising a competition with eight players. All I can say is that I personally do not like the scenario when player A is much stronger than the rest, player B is much weaker than the rest. Player C qualifies with 4 points without playing player A, player D fails to qualify with 3.5 points without playing player B. If the strengths of players A and B are established early on, then at an early stage in the competition player D knows that it is definitely harder for them to qualify than it is for player C.

I agree that no perfect solution exists.

In the 10-player 7-round event I mentioned, I knew from the list of entrants that one player was much stronger than the others, and that aside from the group of players in the same year, they were all about the same. So when picking the initial PINs, I made sure that the players in the same year played each other, and I also picked a PIN that guaranteed they all played against the really strong player. Not only that, but I managed to give them the same colour against the player too!

This wasn't a perfect solution, but given the information, I thought this was the best way of doing it. It made sure that all of the players in the same year played each other, and they are the ones in direct competition. I thought this was better than a Swiss, which had no such guarantee.

_________________
February 16-17: 2013 BUCA Team Championship
April 20: 2013 British Blitz Championship
July 27-August 10: 100th British Chess Championships

There are other things you can do, as well: going back to the 6-player tournament, if you know that one of your six players is much stronger or much weaker than the others (doesn't matter which), you can make things fairer by ensuring he gets PIN 6, so that the other players get two whites and two blacks against each other.

_________________
http://westcountrychess.blogspot.com

I think that there are different ways at looking at this. As far as the strength of players is concerned with the majority of players not having grades or not necessarily having reliable grades if a much stronger player exists it takes at least three rounds to notice this. However I would say that my approach gives me a better chance of finding out this quicker, as in a 8 player tournament there would be no chance of 3 players having maximum points after 3 or 4 rounds or two players finishing the competition with maximum points.

I usually am given the draw for the first round without knowing exactly everything what is required of me. I do not know the players and personally feel it is better to have an arrangement where the options can be kept open, rather than having things pre-determined from the start.

Isn't there a cook you can use to avoid this? You always deem the tournament to consist of an even number of players. There's a Mr Doesn't Turn Up * with the lowest grade of the tournament only included in the pairings when there would otherwise be an odd number. That works for the first round because he gets the median player and for as many rounds as there are players with nil points. Beyond that, you might have to award him the odd half point bye to avoid any need for floats.

* aka Mr C Controller

After my recent dialogue with Alex and Jack on the subject it was interesting to reflect that our megafinal was today. I assisted with three sections, one with 7 players, one with 8 players and one with 9. I did it in the way I normally do it, but could see why some people would prefer a pre-determined format.

There are advantages and disadvantages of both approaches, but I think one key point is that people are given different degrees of overall responsibility at a megafinal, and certainly for me that is a definite factor in how I view things.

I think in some Megafinals these would have been merged into one 24 player section.

The Warwickshire Megafinal merged the U14-18 sections, such that they had 13 players in total. 13 players in 6 rounds meant crazy things happened: In round 5, I had to give a bye to someone on 2/4, because all of the players on 1/4 had already had the bye.

_________________
February 16-17: 2013 BUCA Team Championship
April 20: 2013 British Blitz Championship
July 27-August 10: 100th British Chess Championships

Of course when you have 7 and 9 players in a section the norm is for players to receive a bye, and you do also get some interesting scenarios.

The two things I do not like happening in an eight player tournament in particular is

1) For two players to both finish the section with maximum points.
2) For one player to fail to qualify with 3.5 points when the only player they have not played has 0 points, and for one player to qualify with 4 points when the only player they have not played has maximum points.

My method eliminates 1) and very probably gives me a better chance of minating 2), but the fact that it is more labour intensive might be a disadvantage in other ways.