I saw some tournaments using Swiss normal pairing, lower rating players did not have a chance to play in their own rating group until the last round. SO I asked the TD why not accelerated pairing for the first 2 rounds, and normal Swiss pairing from round 3. His reply is USCF does not encourage that. Is the statement correct? what is the disadvantage of accelerated pairing?
Chiiwen
The only disadvantage, if it is a small turn-out. If you do use accelerated pairing, need to tell the players before the start of the first round. Having a small turn-out and not telling the players is the only clear disadvantage.
The USCF does support accelerated pairings, if they did not they would not have rules about accelerated pairings.
Accelerated pairings aren’t designed to help match up lower rated players against each other, in fact I’m not sure that would be the result of accelerated pairings.
I once played in an accelerated pairing event where I was below the break and won my first 2 games. For my 3rd game I got the highest rated player in the state. Accelerated pairings are not friendly to lower rated players.
This is largely a tournament design issue. Was this intended to be an ‘Open’ event?
What percentage of the field was low-rated players? If they made up a small percentage of the field and it was a 5 round event, it is both reasonable and normal for them to play 3 or 4 games against higher rated players.
Suppose you have 6 players under 1000 in a 5 round event with 120 players. In the first round the bottom player gets #60. In the second round he gets #90 (more or less), in the 3rd round he gets #105. in the 4th round he gets #112. He doesn’t get a player from among the bottom 6 until the last round.
What if you accelerate the pairings for two rounds? Then player #120 gets #90 in round 1 and #105 in round 2. But in round 3 he’s likely to get a player higher than #90 again
Perhaps the budding sabermetricians here could discuss other acceleration methods that pair like-rated players against each other quicker, but I’m not sure how much that helps in the long run. (It might be interesting to experiment with a two-round acceleration that gives 1 point to the top 1/4, 1/2 point to the middle half, and 0 points to the bottom quarter, or perhaps 1, 1/2 and 0 by thirds.)
If the low-rated players were being paired way up for most of the tournament and were complaining about it, then next time think about having an ‘Under’ section for them.
The biggest advantage of acceleration is reducing the number of perfect scores. This is particularly useful when there are trophies and you want to reduce the chance that first place is decided by tiebreaks between perfect scores.
Depending on your point of view another result that might be perceived as an advantage is that it speeds up the pairing of high rated players with each other. If you are looking to publish games that other players may find interesting then this is an advantage.
You asked about disadvantages, and are probably most concerned about them from the point of view of a lower rated player. I’ll use quartile +1 acceleration (Harkness odd man) in the examples. The disadvantages would stem from the widening of the middle band of scores (as the extreme scores get squeezed). If the higher rated player always wins, the third-quarter players can actually get stronger opponents than without accelerating (the widening effect in rounds 3 and 5).
A) The higher-rated player always wins. 120 players and 5 rounds.
Player 120 plays 90, 105, 113, 116, 118. (without acceleration player 120 would play 60, 90, 105, 113, 116). With acceleration player 120 essentially skips his round 1 game against 60 and adds a game against 118.
Player 80 plays 110, 50, 35, 107, 29 (without acceleration player 80 would play 20, 110, 50, 103, 57). With acceleration player, on average player 80 actually plays slightly stronger opponents than without acceleration (29 is 9 players back from 20 and 107 is 4 players back from 103, but 35 is 22 players ahead of 57.
B) The player is the only one scoring upsets (in rounds 1, 2, 4). 120 players and 5 rounds
Player 120 plays 90, 60, 8, 35, 20 (without accelation it would be 60, 30, 15, 38, 23).
Player 80 plays 110, 50, 8, 35, 20 (without acceleration it would be 20, 30, 15, 38, 23). Round 1 accelerated isn’t actually an upset for this player.
In both cases, a couple of upsets pairs the player against top players quickly, but acceleration with only some upsets can actually result in weaker average opposition for the lower-half players (rounds 3, 4 and 5 are somewhat stronger while the opposition is much weaker in rounds 1 and 2), which is not what you might expect.
I’ve said it before, and I’ll say it again:
Accelerated pairings are highly overrated.
For the top 1/8, it has some point, in that the strongest players will play each other faster.
For the bottom 1/8, I suppose it also has a similar point, although at that level randomization of results tends to mask the effect.
For the vast majority in the middle 3/4, about the only effect is to make 1st round pairings feel like 2nd round pairings, 2nd round pairings feel like 3rd round pairings, and 3rd round pairings feel like 1st round pairings. After round 3 the difference is negligible.
It’s way better to simply divide the tournament into two or more sections. Depending on the size and strength of the event, three sections might work, such as Open, Under-2000, and Under-1600.
Bill Smythe