That was you and I, wasn’t it? You were #19 and I was #20 in round 1 of the 38-player Saturday schedule of the Master Challenge one year. Accelerated pairings were used. Neither of us was the TD.
As I recall, we eventually figured out what happened. The program implemented accelerated pairings by temporarily “adding” 1 point to everybody’s score in the top half in the first 2 rounds. So, going into round 1, you were the bottom “1-pointer” and I was the top 0-pointer. Voila – pairing. I seem to remember we both had a good laugh once we figured it out.
This sounds like a “let’s see the full crosstable” situation. You could start a new thread, one of those “why did the computer do it” discussions. Those threads are always entertaining and educational.
One good reason to check the computer pairings toward the last rounds - time permitting - is to be prepared to answer questions before they are asked. Shows one to be on top of the situations when you don’t have to go back and check, especially if you didn’t run the computer. And in the rare case where the computer decided to make a very bad pairing, you can change it before - not after - posting the pairings.
It is great to have someone like Bart Mebane working the back room who will let you know when there are “interesting” pairings so you can double check and be prepared to answer the question which did come.
It was more like ten years ago. I think it was the first Master Challenge at the Dominican Conference Center, after the event had lost its traditional home at the Mohr Community Center.
I’m actually not 100% sure you were my opponent, but I do remember that my pairing had both you and me scratching our heads, until we figured out the program had used the phantom-point method.
Yes, that’s the best reason for the TD to check the pairings, I’d say.
Worst case is when the TD rips the pairings out of the printer and posts them without looking at them, then gets complaints from the players, THEN changes the pairings. This happened a few years ago in a sizeable Illinois event. Worst of all, the TD’s pairings ended up being worse (IMHO) than the computer pairings. There were complications because of the small size of the section, the large number of rounds, and the top player having already played most of his “natural” opponents. It was a classic case of the program using “look-ahead” techniques better than human TDs do.
I guess this sort of thing happens even to the best TDs at the biggest tournaments if they’re running into enough time pressure. I think it was last year at the Chicago Open when the first round pairings were scrapped and redone – twice. I eventually heard that some of the “switch” settings were wrong the first couple of times. I don’t know how big a difference it could have made in the first round of a class tournament, but I’m glad they got the problem resolved quickly and with (relatively) little disruption.
If it was the first round, it was a very different situation than what we have been describing. Must have been something pretty basic, like having accelerated pairings turned on by mistake.
1 Pairing of MCC events shall be done by computer whenever possible.
2 Where questions arise in the accuracy and/or propriety of a computer-generated pairing, it is the policy of the MetroWest Chess Club to accept the program’s pairings in the absence of an obvious or clearly understood violation of USCF rules.
Feel free to poach them if you like.
I can’t recall us having changed a pairing under this rule. Usually it’s an error on our part (wrong results previously entered, missed 1/2 point bye, etc.)
Actually Steve Immitt was the director who looked at the pairing after I asked him about it. He was the one who said it appeared the program went out of its way to make a bad pairing. Having watched Steve spend a great deal of time double checking computer generated pairings, I trust him to know whether the pairing made sense or not. He didn’t think it made sense either.
I agree. Steve is one of the few TDs I would trust to make a better pairing than a good program.
But I’d still like to see the full crosstable! Maybe we can figure out why the program made the pairing, even it was a bad pairing. I’m secretly hoping it fell into the “defective example 5” trap, which I mentioned in another thread. THAT would make for an interesting discussion.
