Alpha wrote:*edit: upon further thought, working out some examples, I think that for a small number of players per game (3,4,5), it would be highly unlikely for the tournament to take more than triple the current number of rounds before a winner is decided so it may be that no maximum on rounds is needed.
Hmm that doesn't sound great... 3x the number of rounds could be 9 rounds for most tournaments.
And 'highly unlikely' means it's probably going to happen at some point... you could end up with a never ending tournament for a 10 player per game tourney (although admittedly this would require a huge number of players so could be some way off).
How about a cap with a some tie-breaker criteria?
tom wrote:
Hmm that doesn't sound great... 3x the number of rounds could be 9 rounds for most tournaments.
And 'highly unlikely' means it's probably going to happen at some point... you could end up with a never ending tournament for a 10 player per game tourney (although admittedly this would require a huge number of players so could be some way off).
I still think it is worth trying. Big tournaments (lots of players) should take awhile.
But, here's a related idea that may be more palatable -
Have a setting that allows the tournament organizer to set a maximum number of rounds (say up to 3 times the minimum). After the minimum number is played then the tournament would end at the maximum number of rounds or when someone has a better record than anyone else - whichever comes first.
Also, remember that people can drop out of tournaments if it is going longer than they like. Does doing so punish their overall tournament ranking score? I would say that it should not.
I would say that it should.
9 rounds for most tournaments doesn't sound at all like a negative to me. A 3 round tournament is a bit silly regardless of what board it's on or how many players or whatever else.
asm wrote: I would say that it should.
9 rounds for most tournaments doesn't sound at all like a negative to me. A 3 round tournament is a bit silly regardless of what board it's on or how many players or whatever else.
Yeah, I'm torn on this.. I've decided not to join any more team tournaments because if one of my partners decides to opt out early, I'm toast.
I agree with Amidon on "Big tournaments (lots of players) should take awhile."
And with asm on "9 rounds for most tournaments doesn't sound at all like a negative to me. A 3 round tournament is a bit silly regardless of what board it's on or how many players or whatever else."
I disagree with idea of having games played only by unbeaten players, hence less player than it was established for the tournament.
I'd prefer to see more rounds, as first solution, and eventually some other criteria, like head-to-head score for example (though I don't think it would work well)
(=
In a different direction, there are two issues that I think require attention to satisfy the "fairness" of the winner being decided. (I believe these are the only two that were brought so please correct me if I have forgotten something).
The first has been discussed at length, tie breakers. By extending the rounds of a tournament, this should be a moot issue as tie breaker should rarely be needed (a clear victor should be decided in at most 6 rounds for the vast majority of the tournaments that have so far been played).
The second (related to the first) is how the top n players are decided for round 3 (or future rounds for that matter). If there is a tournament with 40 players, playing 5 at a time, there can be either 1 or 2 players with a 2-0 record and there will be 14 or 15 players with a 1-1 record. How do you establish which 1-1 players should play with the 2-0 players?
For the second, I will without argument make the claim that for the 3rd round there is no "fair" method to decide that will always (or even most of time) work. There simply is not enough information within the tournament (just looking at schedule of first two rounds) to make any decision about how to rank players with the same record. So, I propose that it should be completely random (I think this is how it is currently coded). Furthermore, after the third round (if there will be future rounds someday), players should be ranked based on record, then score (as it is currently), and if there is not exactly a top 'x' players, then the decision will again be random after the top 'y' are entered in as described in the example above.
It was under this schema, that I reasoned it would be highly unlikely (very close to never) that a Swiss tournament would go on past 9 rounds without one player having a better record than any other. I also agree that 9 rounds does not seem unreasonably long.
Sorry, I said more than I intended to.
The only issue with the "9 rounds long tournament" is that it will happen to play against someone more than once.
So I think the number of rounds should always guarantee that it is impossible to play against someone twice.
(=
Tesctassa II wrote: The only issue with the "9 rounds long tournament" is that it will happen to play against someone more than once.
So I think the number of rounds should always guarantee that it is impossible to play against someone twice.
(=
This kind of scheduling problem is a very hard problem to solve in general, which is why round robin tournaments are only implemented for 2 or 3 player games. Additionally, for swiss system (extended rounds), this cannot be garenteed as after 3 rounds, there will in all likelyhood be two players with a 2-1 record who have already played each other and by design they will have to play each other again.
Alpha wrote:Tesctassa II wrote: The only issue with the "9 rounds long tournament" is that it will happen to play against someone more than once.
So I think the number of rounds should always guarantee that it is impossible to play against someone twice.
(=This kind of scheduling problem is a very hard problem to solve in general, which is why round robin tournaments are only implemented for 2 or 3 player games. Additionally, for swiss system (extended rounds), this cannot be garenteed as after 3 rounds, there will in all likelyhood be two players with a 2-1 record who have already played each other and by design they will have to play each other again.
And probably that would be a good tie-breaking system! What do you think?
If two or more players are equally good, they'll play each other more times during a tournament, and in case of tied score, the head-to-head score would work for them.
(=
I wouldn't mind playing against someone more than once, as long as I don't play against the same 3 people in a second game.
And 9 game tournaments sound much better to me than the current system. I don't want to wait a month for a tournament to start up only to have it finish in a couple weeks :(.
The thing is the whole concept of a Swiss tournament is around having the minimum number of rounds to select a clear winner. If we are talking about increasing the number of rounds (e.g. from 3 to 9) then it becomes a sort of hybrid round robin / swiss tournament. I've no real objection to that but we should be clear about what we want to achieve.
Once I've finished the Simultaneous game mode I'll add knockout tournaments.
tom wrote: The thing is the whole concept of a Swiss tournament is around having the minimum number of rounds to select a clear winner. If we are talking about increasing the number of rounds (e.g. from 3 to 9) then it becomes a sort of hybrid round robin / swiss tournament. I've no real objection to that but we should be clear about what we want to achieve.
I agree with you about the concept of the Swiss system, that's why I originally said that number of rounds should stay below the limit over which you play twice against someone.
That said, increasing the number of rounds is a must right now. Three rounds aren't enough.
tom wrote: Once I've finished the Simultaneous game mode I'll add knockout tournaments.
Really glad to hear it!!! Thx tom! =D
I wouldn't suggest a fixed number, but end only when there is a clear winner. This captures the idea of Swiss, but extends it past 2 player games or fixed number of players per tournament.
tom wrote: The thing is the whole concept of a Swiss tournament is around having the minimum number of rounds to select a clear winner.
(emphasis mine)
I agree with this. I don't think the current setup satisfies this requirement.
In reading about swiss tourney's I see frequently something like this from Wikipedia:
"An additional disadvantage is that, while the players finishing near the top are typically those with the best performances, and those finishing near the bottom are those with the worst performances, the players in the middle tend to be jumbled with little meaningful order. "
In trying to extend the swiss system from 2 player games to many player games one necessarily need sto interact with that middle to match players to games. I think the hope is that with more rounds the middle (and hence the top) will be better sorted out. Yes, this will require playing the same people again - but in many player games this is less of an issue because of group dynamics.
I think the easiest thing is for the players who are tied for the top number of scores to go into a single game to decide the winner.
If this exceeds the number of players allowed in the game then they are split across two games with a final decider.