Well this system is overcomplicated. You can't understand how this works with just a glance. It's hard for the casual viewer to follow what's going on. Some teams will play 6 matches and some will play 12 matches. 3/7 teams won't get a chance to get a norm. Some teams will thus have huge amount of rest days. It would be too expensive to buy flight tickets only right after being eliminated. Teams go to the olympiad to mainly play. Nobody even cares about the other divisions. You could have a top division and swiss for the scrubs, but there's also the world team chess championship which is just for the top dogs. I vote no for this change.
Admittedly the strength differences are huge between teams so some kind of division system might be a good thing. 32 team divisions sounds good enough, but you could make it bigger or smaller if you wanted. There's no relegation or advances because every team will have different ratings every olympiad. After 9 rounds only the top division will take the top 4 teams and let them play ko. The leader gets to choose its opponent. If the match is a draw, ko rapid will ensue and then blitz. The last round will decide the medals. All the other teams continue playing swiss. There won't be too many meaningless games because people care about the ratings. No matter what system you choose, it will have drawbacks.
Following Craze argumentation, it could be great if someone is able to locate and inform on the discussions leading to "Olympiad" format changes in the past.
For the "Olympiad", I see the following problems:
- No norms from most/many countries with elimination/divisions
- Early exit of weakest teams
- No amateur players playing against their idols
These could lead to less federations playing at the "Olympiad", which is its biggest asset.
I would suggest some hybrid system,like, for example, a Swiss for classification of the 8/16 best teams and then just pure KO system. The rest of the teams continue playing the Swiss, mainly for norm purposes. The Swiss could end the day before the final, so that a final with public can be played. You can replace the Swiss for whatever system ensures that the above problems are not an issue
GSL groups are the best group phase I've ever seen in tournaments, chess should absolutely steal it. Currently the major starcraft (the "S" in GSL) leagues do two sets of gsl groups into a top 8 single elimination bracket.
I really like that format. I know it's worse for determining the best player, but single elim brackets are exciting. I think starting the knockouts at the quarter finals is the best balance of getting a consistently loaded top 8, but a sufficiently unpredictable champion.
I like the pasta.
Where is the concern for the expreience to be a learning experience beyond the high emphasis of spectator sport, and podium target.
round robin in principle is more mixing, less stratified mixing. why should there be on one format and same size gladiator sport formats.
I did approve of some of the strating questions, but I found it was going too heavy on the entertainment priority dimesion. Did not keep reading. I think starting a debate should not ask and then hammer and answer to the point we forget what the question possible other answers might have been.
In an ideal OTB chessworld, no online mixing rated and not rated experience, where tournament are really the only experience where to learn from mixing with different biases of previous experience sets being paired, isent streamlining the level of random mixing going to allow conservative sugjective line for risk assessment in game diversity?
could there not be more small scale events, not necessary with money or industrial profitability or world wide shining scope, where perhaps there could be some particiapting requirements to a minimal amount for any big and high stage event entry requirement.
ensuring that at the global and local scales, there is sufficient mixing across all strates of chess experiences? not become like the engine pools, in danger of getting stuck in some style, and not having any means to ensure diversity, for at least making we don,t assume chess is done deal too soon?
I get the size of population argument toward praticality of mixing styles in tournament format conceptions.. been a trend all the extra social competition chess rule piling up in the ruleset along the demographic changes....
but maybe there was some other measures or reasons than the one I could read so far in the blog.... I am using a theory of learning type of thinking here, insipired by the machine models of learning that we have come to not find weird anymore over the past five year that i am being a tourist in the lichess version of the chess community... I am not of the OTB subpopulation though. at least the tournament subpopulation . I am thinking of OTB larger populatoin but non tournament or non club and non rated OTB players. just in case my nesting of pppulatoin seemed like self perpetuating "russian dolls" image..
For me, a simpler approach is:
1) RR + KO like the FIFA world cup. It has some flaws, but it's easier to understand.
2) The RR phase splits 192 teams into 32 groups of 6, with the top two teams advancing to the KO portion.
3) Teams that finish in places 3-4 in the RR enter a division 2 KO and places 5-6 enter a division 3 KO.
4) In each KO division, eliminated teams continue to play each other to determine the exact places (1-64) for each division.
This would consist of 11 rounds (exactly as the same as the previous Olympiad) and give a perfectly timed rest day after 5 rounds of the RR section. Every team ends up playing 11 rounds.
Some games may lack meaning in the RR section, they are relatively early in the tournament. Teams are still fighting to make it to division 2 to somewhat mitigate this issue as well
Even though i would prefer other formats other the one you proposed, I think it's very important to discuss this issue. You did it in a very thoughtfull way, that I really appreciated. Thank you and hopefully soon GM draws will be history...
I think you raise some very interesting points.
I'm not sure I like the idea of a single grand final in a double elimination playoff bracket for chess specifically. It works in a lot of eSports because there is some room for asymmetry which can create incentives to win the upper bracket final. In Valorant for example, you have seven maps. The grand final is played as a best of 5, and the team that came from the upper bracket can essentially select which 5 maps are being played, which can be a big advantage. Not big enough to make the match super one sided (we've seen the lower bracket team win in a lot of instances), but enough to incentivise winning in the upper bracket. In chess you really don't want to do this. The symmetry is part of its appeal. Sure, the Armageddon exists, but that's a last resort tiebreaker. I don't think we want something like a priori time odds in the final of a big tournament.
I'm also not sure I necessarily agree with your take on Candidates. Say what you want about Round Robin, but it is the best format at determining the best player (or the player with the best performance during the tournament). For the World Cup or the Olympiad it's cool to have upsets, but I think the WCC should be between the reigning champion and the best, most consistent player besides the champ. RR does ensure that the WCC map will be of the highest possible quality and we won't see some fluke winner getting absolutely destroyed in the biggest match in all of chess with a lot of eyes on it.
Other than that though, I agree with most of what you say. Thank you for bringing this up, we should be having these discussions more.
