Steve,
The programers use the rulebook and input from TDs to determine how their program pairs. Sometimes TDs contact them. Other times while they are on staff at some tournament they interact with the TDs working at the event. Still other times the programers contact TDs directly.
The TD is responsible for the pairings. The pairings progams are not responsible for the pairings. TDs need to review the pairings before they are posted.
Tim Just
Tim,
I think the light is finally coming on. I have to apologize for not knowing a lot of this because I am not a TD. But, based on your reply, the pairings can be generated by a program but the TD, using his knowledge of the rules and experience, must “bless” the program generated pairings before they can be posted. That makes sense. It also fits in with the complexity of determining which and what rule gets enforced when.
My understanding at this point is that any program that is capable of generating Swiss pairings can be used as ultimately, the TD must approve the pairings (and if need be do them manually if the program is not capable).
None of this helps in building a standardized pairing suite, as was originally proposed, but it does clariify how pairings are accepted.
Thank you very much.
Steve
[/quote]
Perhaps a standardized pairing suite could be developed with user options on which of the potentially conflicting rules would take precedence. Then the software will pair more similarly to the TD.
Steve Coladonato says:
All too often, TDs (even relatively advanced ones) will just rip the computer pairings out of the printer and post them, without even looking at them.
At an Illinois Open a few years ago, this happened, then a few of the players complained about the pairings. The TD staff, AFTER having posted the pairings, looked at them and agreed to re-do the pairings on the top six boards or so. This was a bad scene, especially since (in my opinion) the computer pairings were better to begin with.
It shouldn’t be necessary for TDs to examine ALL of the computer pairings, but they should at least look at the top 2 or 3 score groups, in the final 2 or 3 rounds.
On the other hand, most TDs below the Senior level probably won’t be able to come up with better pairings anyway, although they might come up with pairings they THINK are better. Club and Local TDs are probably better off, for the most part, simply accepting the pairings generated by the computer.
Bill Smythe
Bill,
Given that the pairings are the responsibility of the TD (even when a program is used to make them), should not the pairings, in the Illinois case, have stood unless there was a violation of the rules. And this is where the TD comes in. If the players that were making the complaint were also citing a rule violation then I can understand the TD agreeing to re-pair the round (or the top six in this case). But if there was no rule violation, perhaps just an interpretation issue, then I think the pairings should have stood.
Steve
Agreed. In this case, there was little doubt that the computer pairings were legal. The only debate was whether they were the most desirable possible pairings. So, given that the pairings had already been posted, they should have stood.
Agreed.
Bill Smythe
Bill,
Just out of curiosity, what would make the pairings more desirable? Something like who your opponent was, your opponents rating, your assigned color.
Steve
You have hit the nail on the head. From each player’s standpoint, the most desirable (and hence the only “correct”) pairings are those which give that player the easiest time of it in the next round, and give somebody else a harder time. Small considerations, like what’s best for the tournament overall, which pairings involve the least distortion, which pairings follow the rules, etc are unimportant when it’s ME-ME-ME who’s being paired.
In this case, the situation was difficult. As best as I can recall (and I may have a few details wrong), after 6 rounds of a 7-round tournament, there were four players in the top (5.0-1.0) score group:
– and players 1, 2, 3 had ALL already played each other. Therefore, it was necessary to pair player 4 against 1, 2, or 3, and pair the other two down into lower score groups.
Normally, the best candidate for staying in his score group would be player 1, but in the pairing 1 vs 4, the colors are bad – not just for alternation, but also for equalization (they both had had byes earlier). So the computer paired 2 vs 4 instead, a 41-point transposition. I think the computer did the right thing, but somebody howled, so the TDs paired 1 vs 4, causing one player to finish the tournament with two more blacks than whites, an imbalance that should almost never be necessary in a tournament with an odd number of rounds.
Bill Smythe
Some time back (several years ago), I read a case put forth by whoever it was writing the TD Corner in the rating supplements about Swiss pairings not being appealable at the local level. Players were to file appeals to USCF only. Whatever came of this idea? Does anyone remember the article? Maybe it was written by Jim Meyer, I can’t remember. Anyway, the idea was that the TD was supposed to be familiar enough with pairing rules, and that those rules were not subject to TD subjectivity. Personally, I don’t think pairings should be subject to appeal as it generally slows up the tournament schedule. Unless a player can point out a clear violation of the rules, such as a player receiving the same color 3x in a row (which even that, has a variation rule to it), or two players being paired twice, the pairings, as posted, should stand.
I’m probably way off here in my opinion, and am sure LTDF will set me straight, but in the meantime, any ideas?
The rules address the same color 3x in a row and it is allowed in certain situations (I don’t recollect the exact wording) and there are software pairing programs that will allow you to allow this. I don’t recollect the Swiss pairing rules allowing the same players to be paired twice though.
But, I agree with you in that, unless there is a clear violation, the pairings should stand.
Steve
Clearly, it would be impractical to appeal pairings to USCF, since there would be no way USCF could act in time. In most cases, even a local appeal would not really be practical, because it would take at least an hour to assemble an appeals committee and address the issues.
As for same color 3 times in a row, of course there are (rare) situations in which there may be no alternative. A simple example: What if there are only 2 players in the score group, both of whom have just had 2 blacks in a row?
Even playing the same opponent twice may be unavoidable if the total number of players is small enough. In pairing a 5-round Swiss with 6 players, for example, you must carefully calculate future possible pairings when pairing round 3, else there may be NO pairings possible in round 4. Even WITH careful calculation, if two players drop out after round 3, there may be no way to pair round 4 without pairing two players who have already faced each other.
Bill Smythe
See 27A1.
Acknowledged. But, is it not a bit complex? Why not just “Avoid players meeting twice.”
The rule goes on to say that the rule “…must be violated when the number of rounds is equal to or greater than the number of players.” So now you get into an interpretation problem. e.g. 4 rounds, 7 players, 3 players drop out of the 4th round and player A has already played the remaining 3 players. Does he get paired against a player he has already played, because this round has an equal number of players as rounds, even though the tournament started with less rounds than there were players, in which case he shouldn’t be paired against the same player twice? I think it would be better to just not allow pairing the same players twice and just address the bye score they will get in the round.
Steve
You’re making more of this than it deserves. The meaning of the rule seems pretty clear. When making pairings, you follow a set of procedures: Don’t pair players twice. Pair equal scores. Equalize colors. Don’t give three in a row. And so on. If one of these is impossible, you move to the next higher one. (E.g. you can’t equalize colors, so you pair the score group with some players getting extra Blacks. Or, you can’t pair within the score group because they’ve all played, so you pair each player against the nearest score he hasn’t played.) If a player has already faced everyone left in the tournament, you have to pair him against someone he’s already played (following the other rules as far as possible). I suppose another way to put it is that the “zeroth” rule, not stated, is “pair players.”
You suggestion that a player in that situation should get a bye (of whatever amount) has major drawbacks. The most likely reasons for having to pair players twice are a) the leader has beaten everyone else, or b) the players doing badly have dropped out. In either case, you would be allowing a bye to determine prizes, which is a really bad idea.
Yes, I am making more of this than it deserves but only to show that I think that in some cases the rules are too complex. If I were in the situation that I cited as an example, I would insist on not being paired to the same player twice because the tournament started with less rounds than players and that does not fall under the condition for violating the rule.
My intention is not to argue this point to death, but if you were faced with the example given, what would you do? If you decided to pair the players and the player asked for an appeal, what would you do? How would you refute the claim above? In either case, someone will be upset big time, either the player or the TD.
Steve
You can always end the tournament a round short of that which was advertised, Steve, if you don’t want to pair the same players more than once. Simply have a meeting with the players, address the situation with them and get their opinion as to which they prefer: a shortened tournament or pair two players more than once. When I get an 8 player turnout for an advertised 4 round tournament, I always let them know at the beginning of the tournament that the schedule has suddenly been reduced to 3 rounds (usually with the addition of slowing down the Time Control - i.e. a G/:70 will become G/:90). Unless there is a strenuous objection, then that’s what I do. Usually nobody objects.
The rules are fairly flexible, I think. The rules allow for TDs to arrive at reasonable solutions to problems not specifically addressed by the rules.
I would, as rfeditor pointed out, avoid the BYE situation at all costs, even if it meant one or more players objected to the point of filing a USCF inquiry against me.
I agree that the rule is confusingly phrased. If the situation you describe came up, I would tell the player that a) the first Swiss System rule is that everyone must be paired if possible, and b) b) there are several precedents, in tournaments directed by prominent NTDs. The latter ought to be less important, but many people are impressed by arguments from authority.
The latter is less important because sometimes prominent NTDs do establish wrong precedents 
I give you, as an example, the Blitz rules for National Scholastic Blitz side events, where an illegal move loses the game. This violates the main USCF Blitz rule, where 2 minutes is to be added to the opponent’s time. Variation 3A (p. 291) should, in my opinion, not be allowed since prior to the 5th edition, this variation was not mentioned for USCF events. It was a FIDE rule. There really shouldn’t be two separate reules on this. One is sufficient, and I can adapt to either. Another rule variation that I see unnecessary is the one which allows the TD to correct illegal moves. Why do we need these variations?
Now if that were a rule, we wouldn’t be in the predicament that I described. But I don’t see it. It may be implied or inferred because of the pairing rules themselves but it is not explicit. If it is, I just don’t see it.
Anyway, I think I have generated some interest into looking into the complexity of the rules and to perhaps consider some rewording in the next edition.
Steve
I think paragraphs 26A and 26B cover variations to the rules such as that mentioned.
Steve