I have a question of my own for the experienced TDs. Assume that the transposition in this example was within the 80-point limit. It is then allowed within the same score group. In this scenario are all four players considered to be in the same score group? The pairings include three score groups, and the pairing of 2 vs. 3 includes a score group (0.5) that is lower than the score groups in the board 1 pairing. I am inclined to treat the four players as being in the same score group for pairing purposes, but I can’t find support for this in the rulebook. For example, when a single player is dropped down to a score group that contains many players, he plays the highest-rated player from the score group - not the highest-rated player from the bottom half of the score group. This indicates that the dropped player is not treated as part of the same group, and if so then the first pairing might be correct even if a transposition of less than 80 points was available.
Assuming that the players are in the same score group for pairing purposes, I think that the second pairing is better. I support having no rating limits to correct color issues within a score group, and this is an example of where that produces a better pairing than does the current 80-point rule.
I am operating without my rulebook. But, if memory serves, rule 29D provides examples that essentially say the four players you reference would be in three different score groups, and the first priority is to pair the two players in the top group with the next highest score group(s). If my citation is correct, this would indicate that the first pairings would also be correct under your alternate scenario.
To answer Fishbag’s original question, I would be likely to go with the first set of pairings for the third round. The top two players are virtually certain to get their due color (especially player 3) in the last round, so there isn’t much fiddling to be done. It’s probably best to just go with the natural pairings, unless there is a significant defect that can be fixed with a legal transposition or interchange.
I don’t understand the point of this question, nor how its answer would be relevant here.
Not counting the bye, there are two players at 1.5 and two with lower scores. The two at 1.5 cannot meet again, so no matter how you slice it, each 1.5 plays an opponent with a lower score. Which 1.5 plays which lower-scoring opponent is determined by colors and rating differences, just as with intra-group pairings.
In this case the untransposed pairings result in bad colors, but the rating difference between the players to be transposed is 185 points (2090 minus 1905), so you don’t transpose.
On the other hand, the two players to be transposed have different scores, so it doesn’t really seem right to compare just the ratings. Perhaps the score difference is worth something. If you consider a half-point score difference to be worth 150 rating points, then the 185-point rating difference “feels” like 35, so you might want to go ahead and transpose. But if you consider the score difference to be worth only 100 points, then you might not.
On the third hand, this is a small tournament. Small tournaments are notorious for creating huge pairing problems in later rounds. Generally, with six players, if the colors work too well in round 3, you have painted yourself into a corner in rounds 4 and 5, and you will be forced to pair players who have already met.
That’s what happens in this tournament, too. Try it out in advance – you’ll find that the bad-color round 3 pairings will still let you pair rounds 4 and 5, whereas the good-color pairings will not. (You can still pair round 4, but only because of the bye.)
Summary for this tournament:
If this is a 3-round tournament, make whichever pairings you feel are better.
If this is a 4-round tournament, make whichever pairings you feel are better, but do the pairings for both rounds 3 and 4 before the start of round 3. You may want to reverse some of the colors in round 3 so that nobody gets 3 blacks or 3 whites out of 4 games.
If this is a 5-round tournament, definitely make the bad-color pairings in round 3. There will then be only two possible sets of pairings for the rest of the tournament. One of these becomes round 4, the other round 5 (take your pick).
Relevance goes to the original post, as the question was whether it is was more correct to improve color alternation despite the point difference being more than 80. Boyd Reed’s response (which I found convincing) was that the colors should not be improved even with a point difference of less than 80. For me this was useful as I pair a weekly small Swiss by hand, and often have players dropped from a higher score group. Your feedback on pairing future rounds is also useful. I don’t do this when I have eight players (I don’t think that I could do it accurately in the available time), but I’ve successfully done some advance planning with five players.
For small tournaments (say, less than 20 players) I think it’s good to avoid transposing for color alternation. Transpose for equalization only, or to prevent two players from meeting twice, or to avoid serious problems.
Making the colors work too well tends to divide the players into two camps, those who started with white and those who started with black. Good colors means primarily inter-camp pairings. In a small tournament, you will eventually run short of inter-camp pairings and be forced to make intra-camp pairings, where the colors don’t work. It’s better to have the bad colors in the odd-numbered rounds, when alternation, rather than equalization, is the only issue.
In SwisSys or WinTD, for small tournaments change the transposition limit from 80 to 0 for alternation, but leave it at 200 for equalization.
Small events can have interesting issues.
3 has yet to play 2 and 5.
Look at the pairings (ignoring colors) of 1vs2, 3vs5 and 6bye. If 1 and 5 win then the final round could have 1 (2.5-0.5) vs 5 (2-1), 2 and 3 unable to play 6, 6 having already had a bye, and thus skipping to 1vs6, 5vs3, 2bye - except 5 already played 3 and has to play 1 or 6 or get a bye, so it instead becomes 1vs6, 2vs3, 5bye giving 5 a chance to use the full-point bye to take first (5 thus both played 4 and had a bye).
As Bill Smythe often says, avoid splitting your players into two camps (such as having 1, 5 and 6 with only each other as remaining non-bye opponents while 2 and 3 have only each other as non-bye opponents). Note that a potential round five would require either 2 or 3 to play somebody a second time (both if neither is given a bye)
If you go with 1vs5, 3vs2, 6bye then 2 is available to receive a bye in the final round after a loss or even a draw. If 1 and 2 win then you could have a problematic (but less so than with the other round three pairing) 1(2.5)vs2(1.5), 5(1)vs6(1.5), 3(1.5)bye. For that matter, a possible fifth round could be 1vs6, 3vs5, 2bye (everybody active played each other, 1&5 played 4, 2&3&6 had a bye).
Another thing I often say is, avoid byes if you possibly can. They are ugly, they displease the players, and they skrew things up (although, in this case, they made one otherwise impossible pairing possible because a player who had already played the withdrawn player could now be given the bye, thus effectively “pairing” him against the withdrawn player again).
There are lots of ways to avoid byes. The TD can jump in or out of the tournament, round by round, to keep the number of players even. Or, some willing soul could play an extra game between rounds. Such a soul might even volunteer to play two games at the same time. All of this is great fun.