First, congratulations on growing your event. You’re having what I would consider a “good” problem. It’s also good that you’re willing to tinker with your formats to look for something better. I think this is a trait that all successful long-term organizers have.
Second, I concur with the quoted line of reasoning. But, as a wise old consultant once told me about customers, “when they pay, they say.”
If you set the minimum rating for the top section sufficiently high, and don’t allow playing up as you stated, that should cure the top-player complaints of not playing sufficiently strong players earlier in the event, as it should have fewer players, and all such players should have fairly high ratings.
If the format change proves to be unpopular, I would try to sell people on the idea that multiple perfect scores are not a terrible thing in a single-section cash-prize event. (I would not accelerate a four-round event with divisible prizes; as Bill Smythe pointed out, round 3 under accelerated pairings just becomes round 1 under normal pairings, and there’s not enough rounds to give the higher-rated top scorers after acceleration good chances to meet.) If you go back to single-section events, perhaps multiple perfect scores could have a G/3+2i playoff, the winner taking first prize and the other prizes being distributed normally.
I scanned about six of those tournaments. They ranged from having 0 to 3 masters. Whenever there were two or more masters, two masters played in the final round. In one of those tournaments, there were two masters, and no one else above mid 1800’s. I’m not sure what their expectation is—if there’s one player within 300 rating points and you play him in the last round, what else could changing the tournament structure do? (If they play in the third round, then the fourth round becomes “pick a victim”.) The complaint about the perfect scores is more likely because they come about because of 3 3-0’s with the downfloat generally getting a relatively easy game while two of the masters play each other. If you can somehow entice a few more experts to play, you might change the dynamic. Six months ago, the masters weren’t playing. Now they are. Has the prize fund adapted to that?
I played in a couple of tournaments with accelerated pairings. The first two rounds were challenging. The round 3 “slaughter of the innocents” was not pretty. A bunch of players were really dissatisfied. It is not so bad when everyone is fairly close in rating, like a 200+ size class section, but with a wide dispersal of ratings, it just looks like you are switching rounds.
Is it really true that round 3 in accelerated pairings is essentially the same as round 1 in normal pairings? It may be true for the masses, but at the top and bottom it wouldn’t be. The 2-0 scores and 0-2 scores are playing people closer to their ratings than they would in round 1 of a non-accelerated event. Isn’t the whole point to reduce the number of perfect scores more rapidly?
When accelerated pairings first were popularized, I was a C player. I loved them because after munching on a D player in round 1 I got to play an A player who had already lost a game. About half the time I beat them. Then during the evil round 3 hour I got to play another strong player while I was feeling good psychologically. And if I did get crushed, well, that was the Saturday night round so at least I got a good night’s sleep!
If form holds, the very highest (or lowest) pairings would be among the very highest (or lowest) seeds.
In practice, however - and one sees this in larger events that use acceleration, such as the Spring Nationals - all it takes is one or two upsets early, and you get some serious mismatches in round 3 on the top boards. This is fine as far as the pairing system is concerned, because the goal of acceleration is still being achieved in either case.
The issue I was referring to is just not having enough rounds for top players to work their way back to the top boards after an early upset. This, in and of itself, is not an issue either, except that when shorter events get accelerated, those upsets produce some pairings with big rating differences on the top boards. Players tend to complain about this, because it is “wrong” to many of them to see those later-round pairings. There’s a clear and correct explanation, of course, but when players are agitated and confronted with something they don’t immediately recognize as normal, they often don’t listen.
That’s one reason I tend to avoid accelerating anything less than six rounds. If it’s not an open section, or another section with a wide and reasonably smooth distribution of ratings, I generally won’t do it then either.
One way to combat the 3rd round effect in accelerated pairings might be to continue the acceleration into rounds 3 and 4, except for the top and bottom groups.
In round 3, pair the 2-0’s against each other, half vs half. Ditto for the 0-2’s. Within the 1-1 group, pair by quarters, top quarter vs 2nd, and 3rd vs 4th.
In round 4, pair the 3-0’s against each other, half vs half, and likewise for the 0-3’s. Pair the top-half 2-1’s against each other, and ditto the bottom-half 1-2’s. Pair the bottom-half 2-1’s against the top-half 1-2’s.
It would probably be a good idea to do this only if there are at least 5 rounds altogether.
Doesn’t that just push the “round one” pairings deeper into the tournament? (That’s rhetorical).
If you’re doing that, why wouldn’t you do a full acceleration into round three? The fact that accelerated pairings tends (for most players) to permute the first three rounds isn’t (to me) as big a deal as the very uneven pairings for the 2-0’s resulting from bottom half players pulling round two upsets. In a trophy tournament, there is at least a tie break advantage to having a strong round three opponent, but in a money tournament (which is the situation in the OP), a point’s a point. If you accelerate round three as well, the only way a low-rated player gets to 3-0 is by pulling two upsets.
In regular (two-round) acceleration, I think the only way a low-rated player gets to 3-0 is with two upsets, right?
A player in the third quartile of the wallchart is expected to win round 1, but is likely playing someone from the second quartile in round 2, and then likely someone from the first quartile in round 3.
A player in the fourth quartile of the wallchart would likely have to pull three upsets to get to 3-0 with regular acceleration.
True, but beside the point. Bill’s suggestion was to apply a 3rd round of acceleration, but only to the 1-1’s. I don’t see any point to that since it only delays the mismatches for the bulk of the players to round four, while doing nothing about the round three mismatches on the lower boards in the 2-0 score group. If you really feel like you want to delay the return to standard pairings, just accelerate the whole section into the 3rd round and hope that you can further knock off the bottom half perfect scores.
I understood the suggestion. I was just questioning whether the bolded sentence above works as a justification for adding a third full round of acceleration, as the cited effect already exists in regular acceleration.
No, because any pairing of 2-0 with 2-0 doesn’t do anything special to reduce the number of perfect scores. The only pairings that (one hopes) will help reduce perfect scores are a low rated perfect score against a higher rated less than perfect score. A minimalist third round of acceleration would pair top half 2-0’s against each other, pair the bottom half 2-0’s against the highest rated 1’s and 1.5’s, and pair everyone else normally.
