I can see the potential need for further acceleration when
A) There are more than (2n) + 1 scores of X-0 where X is the number of rounds played and n is the number of rounds remaining (i.e., more than five 5-0 players in a seven round tournament or more than nine 2-0 players in a five round event). If it is no more than 2n then no acceleration is needed. If it is exactly 2**n+1 then the extra person is getting the top player in the next scoregroup anyway and acceleration won’t do any better.
and
B) The reasonably expected number of draws (or ties in team matches) in the top score group would not be enough to reduce the number of perfect scores to one.
My prime example for this has been an accelerated four round weighted board team event (8 boards worth 68 points) where ties are very rare (2% would be high) with more than 40 teams. Four high-rated teams were 2-0 as expected and two middling-rated teams scored upset wins in round two and were also 2-0. Normal round three pairings would have left three 3-0 teams with a strong likelihood that the third 3-0 would stand a good chance of beating its potential 2-1 opponent in the final round and the tournament ending up with two 4-0 teams. I paired the two four teams against each other and put the bottom two 2-0 teams against the top two 1-1 teams (they lost). That left a natural final round pairing of the only two 3-0 teams and finishing with only one 4-0 team. I had no qualms doing that in a non-USChess-rated event, and a rule change would open the possibility to US Chess rated events.
One suggested modification would be to strongly advise against acceleration in the final round (one upset could cause a lot of complaints about a team getting an undeserved trophy).
In general, top half players will face stronger opposition (and bottom half players weaker opposition), so this will distort the standings for any rounds for which you have applied acceleration.
I suspect this is one of those statements that will hardly be read but, more likely the result will be noted by TDs and players. The procedures of this rule in general are often not noted at all. Why? Just fire up WinTD (or whatever)–point–click–press a few buttons and the round is accelerated. No rulebook reading required. So, a warning(?) like the one above would probably be missed by most folks.
You might want to make a change similar to changing
to
For those people that just click on settings and don’t bother to read the rulebook, when a TD clicks on the WinTD final round flag you could have a follow-up verification as to whether or not the final round would still be accelerated (assuming you program that into WinTD). If you don’t program latter-round acceleration into WinTD then a TD using latter-round acceleration would have to know enough about it to do it with manual pairing overrides (not a bad way of limiting it to when it is really needed and making sure it is done by somebody who has a better chance of knowing what is involved).
My limited experience with accelerated pairings was with several 5 round 50 to 80 player opens. In a practical sense, it seemed to move the round 1 mismatches to round 3.
It had two “advantages”:
I got to go home earlier in round 3.
The middle players had a better chance to score points in rounds 1 and 2. (This may be a better “marketing strategy” for chess.)
Although I understand the rationale behind accelerated pairings, as a player, I would much rather get the mismatches out of the way early and play increasingly competitive games as the tournament progresses. Although this is partly because I’m used to such a progression, it also reflects the way I generally approach things. For example, I will generally start a weekend by getting all my chores done and out of the way first – and then I can enjoy the rest of the weekend. And I would much rather start a bicycle ride with a headwind and finish with a tailwind, than the other way around.
When acceleration works perfectly, round one will seem similar to a standard round two, round two will seem similar to a standard round three, and round three in the middle will seem vaguely similar to a standard round one (perfect acceleration of 64 players will have 9-32 playing 33-56 instead of the normal round one having those players playing an opponent a mere 8 pairing numbers farther away) while round three at the top and bottom will seem similar to a standard round four.
The players in the middle 3/4 will feel like they are kind of playing the round one mismatch and it may well make for an earlier night for them. They get an experience that is a little disjointed in comparison the a standard first three rounds and the later rounds have slightly larger mismatches than under standard pairings (the round three accelerated mismatch is not as much as the round one standard mismatch but the somewhat larger mismatches in later rounds may mean the the average mismatch under acceleration ends up to be more than under standard). The people at the ends get closer matches faster.