I thought about that, but there are problems with that as well. I don’t like the consequences for the person who gets stuck at the top of the bottom section. He doesn’t get to play evenly matched opponents. He doesn’t have much chance for a big rating gain because he is playing weaker opponents. He doesn’t get a chance to play someone good. In most tournaments, he is compensated by having a good shot a cash prize, but I don’t give cash prizes.
I suppose I could at this one. It won’t be part of the “charity” series.
With the three sets of quads, I think it very unlikely that the same three players will play each other three times. There will definitely be overlap. but there should be different players in each of the three quads.
But that passes the problem to the next guy down on the list. There’s a new top guy. (And it really isn’t just “the” top guy. It’s a handful near the top of the bottom section.) Even in quads it can happen.
I run Swiss during my regular tournaments, because I think that overall, it’s the best way of doing things, and one game is a wasted game, or for some people, two games are wasted.
Starting with my last tournament, I experimented with a new pairing system that was not exactly Swiss, and I mostly liked the results. If you look, you’ll see that the top player in a five round Swiss played against the five players with the highest ratings. That was what was intended. If he were playing for a cash prize, that wouldn’t be so good. The guy who starts at the top would have a penalty, having to defeat tougher opponents. As it is, he was there to play good games of Chess, and that system produced the best it could.
If I can get one hour earlier in the AM, I could see a quad - 5SS - quad format. The early birds could start at 9 and be done by lunch. The afternooners could play all day and have an hour dinner break. The crazies could play 11 games. I would expect 20 players or so for the afternoon session, so 4SS wouldn’t be as good as 5SS, but a two hour dinner break would be better than a one hour. So many options…and there’s the Shogi people to consider.
I think showing up for a 9:00 tourney is kind of obscene myself, but I hear that there are “morning people” in the world. Do they play Chess?
I think you might be missing a very important point about having one game where a higher rated opponent plays a lower one.
The main point that I experienced when starting was that this game, usually in the first round, was much a chess lesson for the lower rated player. Doing a post mortem with the higher rated opponent shows what the lower rated player might look at for improvement. And the post mortem is easy to do since that game is usually over in a quicker time than more evenly rated opponents would have.
Also, as one gets better this provides a good opportunity for an upset. I fondly remember being a 1380 player and beating a 1700 player in the first round.
As I got better, those first round games took longer and when I lost it was a more sophisticated, for me, mistake that I learned about.
If you want to not consider money as a factor in a tournament, you really should pay no attention to it in your design of the pairing system. If someone comes to play chess and they face a weaker opponent, that’s life and he should be happy to play that person even without some money as a prize.
I have run free tournaments with no prize fund, both Quads and Swiss style events. My players have ranged from 600 - 2080 in rating. All come to play chess and if a player is paired against another with a great rating difference, they simply play them. It’s still Chess after all.
The higher rated player will have an easier game in that first round and often times will use it as a warm up for the rest of the tournament. He can also play around with different openings. I know of some that use some openings only when they play a weaker opponent.
Weeks?? Various ways to (re-)do norm-based titles in the USCF world were floating around for over ten years.
Whenever you have a 0-1 decision like this (you can’t get .99 norm if you just miss), you have to draw bright lines. When you do that, there will always be examples on either side of the line that feel “wrong”. Is it possible to have a 3-round result that seems more meritorious than a marginal 4-round norm qualifier? Yes. We put up with the marginal 4-rounders as the price we pay for having a relatively simple criterion. If we had a more complicated criterion (I had one that required use of both the standard normal density and distribution functions), the effect would be to leave out the marginal 4 rounders and exclude 3 rounders.
In a multi-section Swiss as many players as want to play up can do so, subject to whatever restrictions the organizer imposes. It gives the players more flexibility than quads, which are usually set up strictly by rating. Some players like to play up and some don’t.
It sounds like you used 1 vs. 2 pairings (Variation 20L1, page 170 of the 5th edition rulebook). Although I haven’t used it myself I think it’s a reasonable system for a tournament without cash prizes.
True, but then what can happen is that there are three people near the top of the bottom. Two decide to play up. The third is now the one who is stuck.
I do agree with Ron’s point about the benefits of that first round throw-away. It’s not all that bad, having to play someone much better, or having an easy start. All of my tweaks are really just fine tuning to avoid situations that I see. The huge point spreads were things that people didn’t like about my tournament, and have been cited by a couple of people who chose not to return.
Plain old Swiss really isn’t bad, but I’ve been tweaking tournament rules in my “other sport” for 30 years. I’ve run a lot of tourneys there and learned a lot about what works under what circumstances. I’m trying to do the same with Chess. It’s interesting to see different challenges in the different games.
Not exactly. It’s a system of my own devising that is basically Swiss on the top, Swiss on the bottom, and 1-2, 3-4, in the middle. If I ever release my tourney software, it will be one of the options.
Ironically, the one option I am having a lot of trouble programming is USCF standard Swiss. I have to get back to it soon. I never got it quite right.
Another method to work with this might be 1 vs. 2 pairings. (Each round, closest in rating is paired.) I’ve done this with some of our club events, and it does shake up the usual pairings somewhat in early rounds and does a lot to avoid the trouble of round 1 throwaways/massacres. The trouble is that in an event of any size, in later rounds you certainly run the risk of having to drop multiple score groups to find opponents who haven’t met.
I’ve been running tournaments that are something like this for several years. The main difference is that instead of 3 sets of quads, we have only 2.
I originally began doing this because (1) it was the first tournament I had ever run, so I didn’t want to have to deal with the complexities of running a swiss, and (2) there was only one player in our club who had a rating, and one of the goals was to provide a way that all participants could come out of it with published ratings (which requires at least 4 games). But it quickly became clear that there were a couple of other benefits: (a) having 2 sets of quads provided a convenient place for people who couldn’t stay for the entire day to leave or join the tournament, and (b) since ratings assigned to unrated players had to be somewhat arbitrary, it was possible in the second set of quads to tweak the assigned ratings based on how people had done in the morning.
I arrange it in such a way that players who face each other in both sets of quads will do so with reverse colors, and I try, as much as possible, to insure that players who stay for both sessions will be color balanced, though there are situations where this is not possible.
True, but this doesn’t mean we shouldn’t allow norms at all from 3 round events. At the very least, a 4th game should be calculated with a loss to a 100 player, the worst case scenario. Bill Smythe previously stated that it would be stupid not to award the norm in a situation like this.
If you are worried about the “throw-away” first round game due to huge rating differences, then use accelerated pairings. That tends to shrink the rating gaps and lower the number of “throw-away” games in the tournament.
Another option to quads is the 3 round swiss done in rating groups. The Octogons were popular in the 70s & 80s, but have since fallen out of favor. Also, with an Octogon there is less of a problem with having an odd number of players.
You could try for “Quicker Quads” Where the first set of Quads were G/60, then your next set of Quads are G/45, and the final set of Quads are G/30. That might fit into your idea for 3 sets of Quads in 1 day. I do recall Tom Fineberg running a tournament where the time controls changed for each round.
Acceleration is not a panacea. In a 16-player event with no color transpositions and the higher rated player always winning, player 7 has standard opponents of 15, 3, 11, 4 and accelerated opponents of 3, 11, 13, 2. His closest opponent is farther away in acceleration than in standard. Have fun telling him you accelerated to get closer games.
I haven’t used accelerated pairings for many years. Getting “real” results seem more important than having only one winner. Something unusual always seems to happen and from the second quarter of the field. Had results in a club championship where there was an eventual tie between two IMs and a Class A player, the A player both winning on tiebreak because of accelerated pairings and sneaking into the winner’s circle because of the system. Just isn’t worth it, at least to me, the risk of weird results versus multiple “real” winners.
I pretty much agree with you, but there are people who think that accelerated pairings can be used to reduce the mismatches even when a maximum of one perfect score can be guaranteed by standard pairings.
For local scholastic tournaments there are not as many draws in the lower grade levels and there is less acceptance of multiple perfect scores with trophy awarded by tie-breaks or a blitz play-off. In such cases I generally look at using acceleration somewhere around 1.5 * 2^n, which would be 24 in a four-round Swiss. In this area that seems to balance the needs for a single perfect score against the likelihood of acceleration working against you and giving more perfect scores than standard (as a reminder, don’t accelerate single-class sections as it is very feasible to have acceleration increase your number of perfect scores).
Only in the first two rounds. In round 3, only the top eighth and bottom eighth are paired close to their ratings. The vast middle three-fourths experience pairings which feel like first-round “standard” pairings. The effect is similar to first playing round 2, then round 3, then round 1.
An improved version of accelerated pairings might be along the following lines. In round 3, pair those with 2-0 against each other normally, and ditto for those with 0-2. Pair players with 1-1 by quarters (top quarter vs second, and third vs fourth). In round 4, top-half 2-1s play each other, bottom-half 2-1s play top-half 1-2s, and bottom half 1-2s play each other.
Some details regarding draws and odd players would have to be worked out.
Looking at player 15 in a 32-player, 4-round event (with the higher rated always winning and ignoring colors), standard pairings have the opponents as numbers 31, 7, 23 and 9 while acceleration has 7, 23, 27 and 8. So accelerating made his closest opponent 7 pairing numbers away instead of 6.
Accelerating does a great job with closer games for the outer ends groups (1 plays 9, 5, 3, 2 instead of 17, 9, 5, 3 and 4 plays 12, 8, 2, 11 instead of 20, 12, 8, 2 and 31 plays 23, 27, 29, 32 instead of 15, 23, 27, 29).
The net effect of such acceleration is to reduce the 4-0 and 0-4 score groups by one person each (from 2 to 1) while increasing the 2-2 score group by two people (from 12 to 14). The single 4-0 versus two 4-0s is the primary reason for using acceleration (accelerating a 4-round, 32-player class section has a good chance of increasing the number of perfect scores to three).