Suppose, after three rounds, one of the score groups has 6 players in it, with “raw” (untransposed) pairings for round 4 as follows:
1900 BWB vs 1590 xWB
1800 WBW vs 1570 WBW
1700 BWB vs 1500 BWB
(Note: All pairings are shown with the higher-rated player on the left, not with the player to be assigned white on the left.)
Suppose further that nobody has already played each other. Note that the 1590 had an unplayed game (half-point bye or something) in the first round.
What transposition should be made to improve colors?
What statement, if any, in the rulebook supports your answer?
What statement, if any, in the rulebook tends to support another answer?
What wording change in the rules would cause the rulebook to better support your answer over another answer?
Per 29E4, equalization takes priority over alternation when determining due color.
Per 29E6a this priority is also considered when trying to avoid color conflicts (thus 29E6a kicks in when the 1900, 1700, 1590 and 1500 are all due white, and indicates that for 29E6a pairing purpose the 1900, 1700 and 1500 are due white, the 1800 and 1570 are due black and the 1590 is not due either).
Thus you get
1900 - 1590 (the 1590 does not alternate)
1500 - 1800
1700 - 1570
The 70-point transposition equalizes all players except the 1590, whose already matching count of whites and blacks makes equalization irrelevent.
If you use 29E6b (even though the rulebook says that 29E6a is more accurate, if available) then you have
1900 - 1570
1590 - 1800
1700 - 1500 (the 1500 does not equalize)
A person reading 29E4 may misread it to infer 29E6a without actually reading 29E6a. If that is done then, in this case at least, the resulting pairings would be the same as if 29E6a was read.
Bill, I am waiting for you to drop the other shoe.
P.S. 29E4c is irrelevent because there no two of these players are paired with each other while having the same prior color imbalance.
I had not considered the possibility that it would make any difference. Since you apparently believe it does, let me turn the question around, and ask you two questions: What pairings would you make if it is the final round, and what pairings would you make if it is not?
Please answer questions 1-4 (in my original post) for each of the two cases.
That’s certainly the way I see it. But the “top-down” crowd (translation: too lazy to look ahead) would notice bad colors in the top two pairings, immediately transpose them to fix both, then, when finally looking at the third pairing, throw up their hands and say, well, the top two pairings have already been made, can’t change them now.
I suspect that many top-down TDs are in the (extremely bad) habit of writing down each pairing (on either the pairing cards or the pairing sheet, or both) as they make it. This would explain their reluctance to look ahead, as it would result in a lot of scratching out and rewriting.
And then there’s the “minimize rating differences” crowd. In this example, they would notice that 20 points is a smaller difference than 70, and come up with the same pairings as the top-down people, although for different reasons.
Aw, shucks, you know me too well.
Okay, here’s another shoe. I won’t say “the other” shoe, because this animal may turn out to have several feet.
This time we are pairing round 5. After four rounds, our 6-player score group has “raw” (untransposed) pairings for round 5 as follows:
1900 WBWB vs 1590 BWWB
1800 BWBW vs 1570 BWBW
1700 WBWB vs 1500 WBWB
(Again, all pairings are shown with the higher-rated player on the left, and again, nobody has already played each other.)
Since four of the six are due for White, there is going to be one bad color. The smallest transposition that improves colors is 1590/1570. So we get
1900 WBWB(W) vs 1570 BWBW(B)
1590 BWWB(W) vs 1800 BWBW(B)
1700 WBWB(W) vs 1500 WBWB(B) (Identical color histories so the higher ranked gets his due color unless you toss.)
Darn Bill you changed the color history. Anyway, in your original, I get the following:
1900 BWB(W) vs 1590 xWB(B)
1800 WBW(B) vs 1500 BWB(W)
1700 BWB(W) vs 1570 WBW(B)
This is the Look-Ahead method. We have 4 out of six due the same color (white). The object is to avoid pairings in which neither player is due for that color (white). So, board 1 is ok as it is, since at least one is due for white. (I must admit, I still don’t understand the due for neither color in the 3rd rd when an unplayed game is involved).
Now, since on board 2 neither player is due white, we make a transposition between 1570 and 1500, and the group is paired.
Since I’m a look-ahead guy, I appreciate this quiz.
rfeditor was listing the pairings with white on the left. I had listed them with higher-rated on the left. The difference caused a little confusion, it appears.
Yes. We can use the same philosophy as in the original (4th round) example. (This is the discussion I was hoping would come up.)
Since one of the four players due white must receive black, there’s a lot to be said for assigning black to the player least due white among the four. That would be the player with BWWB. (See the latter part of 29E4 for an explanation of why WBWB is “more” due white than BWWB.)
This gives us:
1900 WBWB vs 1590 BWWB (1900 gets white)
1800 BWBW vs 1500 WBWB (1500 gets white)
1700 WBWB vs 1570 BWBW (1700 gets white)
Bill Smythe
It’s an interesting suggestion, but I don’t think I agree with it. I have two objections, one based on the wording of the rules and the other on my personal prejudices.
The first, of course, is 29E5c, which seems to say that the smallest possible rating interchange should be preferred. (I’m not sure it actually says that explicitly, but the examples given strongly suggest that this was the intent.)
My other problem is that your pairing puts the bad color on the highest board. I can’t quote a rule covering this, but it feels wrong. (This is mitigated by the fact that it is the lower ranked player getting the “wrong” color, but the principle is the same.)
I second John’s disagreement, but for a different . My reason (as stated in another thread) is that 29E4 is part of the rules discussing who is most due a color between two players that are already paired with each other, and thus I don’t feel it should be used to decide who plays who, but rather what colors they play once that decision is made. Thus the 20-point transposition trumps the 70-point transposition and only board three has an irresolvable color conflict (since the color history is identical the 1700 gets due color as the higher ranked player).
If the lower-rated players were 1590, 1520 and 1500 then transposition would be on boards two and three, leaving board one with the color conflict (1900 gets white because of the round two colors each player had, while if you reverse the color history then the 1590 would get white).
I wouldn’t have a problem with a rule change to promote what you would like to see (or a rulebook editor clarification that your interpretation is correct), but for the 1590/1570/1500 I would not make your pairing without that rule change/clarification. In the FKB thread I did consider it before deciding that it was not applicable as the rules are currently written.
This became an issue in the other pairing thread that I initiated recently. Is this just “common sense” or is there a rule in the rule book backing it up?
I quite agree that my suggestion might be difficult to justify based on a literal interpretation of the current rules. It’s another case of what the rule “should” be rather than what it is. Having been a member of the 5th edition rulebook revision committee a few years ago, I guess I’m still in rule revision mode.
Nevertheless, there seems to be an inconsistent philosophy at work here. In the example that started this thread (shoe number one), everybody agreed that the transposition giving the wrong color to xWB was preferable to the transposition giving the wrong color to BWB, even though the former transposition was 70 points while the latter was 20.
But now (shoe number two), nobody wants to give the wrong color to BWWB rather than to WBWB, under essentially identical circumstances.
The case of the U.S. Championship round 8 (shoe number three?) seems intermediate between these two cases, but closer to shoe number one. It involved a transposition giving the wrong color to BWBWBWB versus another transposition giving the wrong color to WBWBBWB. I thought the latter was highly undesirable, assigning black in 4 of the most recent 5 rounds, just a smidgeon less serious than assigning 3 blacks in a row.
Here I totally disagree. I have no problem putting the bad color on the highest board. Even if a pairing turned out to give the bad color to the higher player on the highest board, it still wouldn’t bother me. “Higher rated gets due color” is just a tiebreaker (the 5th tiebreaker, at that) and is quite arbitrary.
Sometimes, when I contribute to (or start) a thread, the discussion is lively and productive for a while, but eventually, one of my posts seems to make the thread stop dead in its tracks, with no further contributions from others.
Is this because:
A. My last post makes so much sense that nobody can possibly argue with it, so everybody just accepts it without question.
B. Everybody tires of the discussion, and just throws up their hands as if to say, “If you say so”.
C. My last post is so absurd and far out that everybody thinks I’m a nut, not worth arguing with any further.
As of June 4, 2007, the three most recent threads in Tournament Direction all have my name on them as the most recent contributor. One of these dates back a full ten days. The current thread, I had hoped, would last a few more days. Any further thoughts?
I am using this particular thread (at least your starting question) as part of the next TD Workshop I am running. Thanks for the material! Yes, you get credit for it in the workshop handout.
This is my first post on this board, and my first langage is french, so please be gentle to me .
My point of view on this case is the following.
I believe the general idea of the pairing rules, although probably not specifically written in the handbook, is to pair the higher boards (in descending order) of a bracket first. The “look-ahead” method, as it has been called here, certainly goes in that direction. Because of that, the above given case would give the following pairings:
1900 WBWB vs 1570 BWBW
1590 BWWB vs 1800 BWBW
1700 WBWB vs 1500 WBWB
(prioritizing bad colors (as anyway there will be a bad color somewhere) to the bottom board). Note that the fact it happens to be to the lowest rated player of this group is just a coincidence.).
That being said, I believe 29E5c, as it has been explained here, is not well understood. The exemple given in the handbook seems clear enough to me.
2000 vs 1800
1980 vs 1500
When switching the 1500 player with the 1800 one, the switch should not be considered (“evaluated”) as a 300 rating points swich but a 20 points switch (the smallest difference between the involved players). Nothing more and nothing to do with pairing priorities.
I see that I must not have confirmed my prior comment when I thought I posted it last week.
BWWB and WBWB are already equalized to the same extent and are both due the same color by alternation. xWB and BWB do not even have the same equalization. Thus the two cases are quite different. Transpositions to make color corrections for xWB versus BWB are already in the rules. At the moment transpositions are not warranted to adjust colors for BWWB versus WBWB.
I do have a concern if you use color history not only to determine colors between two players paired with each other (the current rule governing BWWB playing WBWB), but also to determine who plays who in the first place. You are pretty much forcing either use of computers to pair, or a lot of extra manual oversight. If a rule change is made then for those directors that have not computerized there may need to be an allowed variation to not use full color history to force transpositions.
.