Swiss system: match distribution problem!

[Tulivuori]

How I believe the algorithm works is that it starts from the leader, and tries to find an opponent that the leader has not yet been played against. When it finds one, or when it comes to the bottom, it stops. The latter part is the problem (if there are too many rounds). A better solution would be to ignore the "try to avoid same opponent" condition, and match with the player closest in the standings.

[Pinovitch]

Finally, understand that the issue you're encountering, which stems from your misunderstanding of how to set parameters for the Swiss system tournament and restrictions you choose to place upon your tournament organization are, let's be honest, quite an edge case. Basically, the vast majority of people that have taken part in this discussion are either players from your group, or myself... So asking BGA to change settings for roughly 200 players, while they have to satisfy a player base that can be counted in hundreds of thousands to a couple millions, you understand that you're not really a priority to BGA admins... They don't have to cater to your particular problem...

Et pourquoi ils ne devraient pas écouter mon problème ?
On est peut-être que 200 mais il y a une bonne proportion de premium et d'utilisateurs quotidiens et sprockitz ne fait pas partie de mon groupe (et je soutiens ses arguments d'ailleurs)
Tu restes cantonné à la théorie du système suisse mais as-tu déjà vu comment se déroule un tournoi ? Je vois que tu n'as pas beaucoup de parties à ton actif et tu n'es finalement pas concerné par notre problème...
Pour ma part je pense avoir organisé peut-être 300/400 tournois donc j'ai une bonne compréhension du système pour savoir qu'il ne fonctionne pas correctement. Mais je ne jette pas la pierre à BGA, parce que c'est cool la possibilité d'automatiser tout ça.
Pour rappel, les tours suisses permettent deux options :
1. Try to prevent same players to be opponents twice
2. Don't try to prevent same players to be opponents twice

Si je m'en fous que 2 joueurs se rencontrent plusieurs fois je choisirai la 2 !

Mais moi je choisis la 1. Et j'estime qu'à 21 joueurs pour 20 parties, on peut chacun avoir des matchs différents : c'est possible. La V1 la faisait et il n'était pas défectueux. Il était plus efficace !

[Pinovitch]

I'll be willing to admit that the algorithm is flawed when I'll have conducted an analysis that proves it, or when someone presents a complete and thorough analysis proving it. What you present for now are claims and nothing more, unless you're able to present otherwise. For now, I'm still analyzing the Patchwork tournament I mentioned, and doing it by hand is quite time consuming... Are you willing to admit that you've been in these situations because of the settings the tournament creator used ? That doesn't seem to be the case seeing your answers...

https://boardgamearena.com/tournament?id=113201

A partir du tour 10 j'ai systématiquement rencontré les derniers du classement... Équitable, logique ? Pratique, oui.. j'ai pu gagner le tournoi aisément !
Le système de calcul pouvait faire un autre brassage...

[Mathew5000]

How I believe the algorithm works is that it starts from the leader, and tries to find an opponent that the leader has not yet been played against. When it finds one, or when it comes to the bottom, it stops. The latter part is the problem (if there are too many rounds). A better solution would be to ignore the "try to avoid same opponent" condition, and match with the player closest in the standings.

Yes! The most salient post in this thread.

It looks like BGA developers are not prioritizing any improvements to tournaments, but one very easy change to make: Change the default to "Don't try to prevent same players to be opponents twice".

Edited to add: the options currently are:
- "Try to prevent same players to be opponents twice" and
- "Don't try to prevent players to be opponents twice"
but this is poor English. In each of those options, "from being" should be substituted in place of "to be". Or something like:
- "Try to prevent players from meeting any opponent more than once"
- "Don't try to prevent players from meeting each other more than once"

[Mathew5000]

They are playing bottom people they have already played (not bottom people they haven't played)...because the algorithm is broken...

I raised this issue a year ago in the "Bugs" section of the forum: https://boardgamearena.com/bug?id=56504

Anybody following this thread, please consider voting for that report.

[Pinovitch]

Histoire d'alimenter le sujet...

Tournoi de Potion Explosion : https://boardgamearena.com/tournament?id=187187

Journée 8 : on est 44 joueurs (2 ont abandonné), je rencontre pour la troisième fois MadNein (step 4,7 et 8) alors que l'option Try to prevent same players to be opponents twice est sélectionnée.

ça fonctionne toujours pas

[MadneiN]

I'll make my contribution. I confirm the last post from Pinovitch!
And maybe I didn't make so much online tournaments as him, but in real life I made a lot of chess tournament. And that's was for many years ago: no computer calculation, online per hand and head. I never met twice the same players during these tournaments!

I really think there is a problem with the current swiss tournament system!

[Superluminal]

I can attest to this issue. I am currently tied at the top of the leaderboard in a tournament that has been ongoing for over three months. This tournament features 3-player Tash-Kalar Deathmatch games. The pairing policy "Try to prevent the same players from being opponents twice" is activated for this event. Despite having played against each participant once, I've found myself repeatedly matched against the same players. My username is superluminal. You can verify the pairings here:

https://boardgamearena.com/tournament?id=184247

From games 12 to 16 (current game), I've been paired with one player 4 out of the 5 times. Another player has been my opponent in 3 out of these 5 games. Notably, both of these players are at the bottom of the leaderboard. I'm currently sharing the first place with rogernunn1. I fail to understand the logic of pairing players from the top of the leaderboard with those at the bottom, particularly towards the tail end of the tournament. This is the point where top contenders should ideally be pitted against one another to determine the final winner.

Moreover, the bottom of the leaderboard likely consists of players who have either abandoned the tournament or are frequently timing out. This means games are not always being decided based on skill, but rather by other factors, some of which may not be entirely fair.

Rogernunn1 faces a similar issue. He's been matched with the same player for the fifth consecutive game, a player who seems to have abandoned the tournament and has not done any moves since a long time.

It's deeply frustrating to experience such irregularities, especially after dedicating so much time to a long tournament. I have been desperately waiting for a challenging match against rogernunn1, eager for a good fight. It's only logical that those at the top of the leaderboard should play against one another to determine the ultimate winner. Ideally, players like rogernunn1 and I should be paired, or at the very least, against others with high scores. I suspect that players at the lower end of the leaderboard aren't enjoying this any more than we are, continually finding themselves outmatched by the top contenders.

[covadis]

Two year later we still have the same problem!
Getting new games is fine, but if you can't enjoy them in good tournaments, is it really useful to have more choice ?
Why BGA is never answering to request and did changes that are not requested ?

[BananaHarambe]

All bugs with Swiss tournaments with the “try not to pair same opponent twice” settings in the topic are clearly created because the tournament creator chose BAD SETTINGS rather than an issue with the BGA coding (even though the settings broke the BGA code).

I will not go on a rant about how to properly make a Swiss tournament because I see the previous pages already cover it very clearly. As an example, it is mathematically IMPOSSIBLE to prevent duplicate matches in a 20 round Swiss tournament with 20-30 players by using the proper “Swiss Pairing Algorithm “, which pairs from the current 1st player and works down, player by player. The Swiss algorithm should not need to pair everyone in a way to prevent ANY duplicate pairings, that’s not the priority. The priority SHOULD be to provide the best possible pairings for the top players from top to bottom. Bottom players may get some duplicate pairings, but that’s not important compared to best pairings for the top players, which are playing for final placement. If you are having issues with this then the player count is WAY TOO LOW, under 40.

If tournament creators insist on having 20 matches despite a low player count, then just select the do not try to prevent option and run a “hot seat” style Swiss tournament where players move up and down in table number based on winning or losing. If there’s high matches, low players then players should WANT to be allowed to play the same players multiple times if they have the same player strength rather than force the strong and weak players together, which DEFEATS THE WHOLE POINT of running Swiss. The goal would be to win until you reach the top table, then try to win/maintain your position there. This produces equal matches over time which is the point right?