RazorOz wrote: ↑06 September 2025, 01:13
Mathew5000 wrote: ↑05 September 2025, 17:36
So to be clear, you're saying that in tournaments of type Multiplayer Elimination, BGA also uses the typewriter method to seed players based on Elo? Does it do that in each round or just the first one?
Group stage and multiplayer elimination both use the same methodology yes. In that specific example the group stages are just multiplayer elimination both stages so may influence that.
It uses this formula in R1 then in the rounds after it appears to be random, but I can't be 100% sure.
Based on a test I am running, I would tentatively agree with you, that for multiplayer-elimination tournaments, round 1 is seeded by Elo, following the typewriter method (which is flawed), and for round 2 the matchups are likely at random (at least, I see no pattern). My test tournament is
https://boardgamearena.com/tournament?id=433038 (Wingspan 3p, 27 entrants, keep 1 player per match). In round 1, I verified just before starting that the matchups were based on Elo ranks as follows:
• game 1: #1, #10, #19
• game 2: #2, #11, #20
• game 3: #3, #12, #21
...
• game 9: #9, #18, #27
But in round 2, which is just about to begin, the Elo rankings of the players are:
• game 1: #8, #3, #9
• game 2: #7, #4, #5
• game 3: #2, #1, #6
It's weird, right? You'd think that since they have a system in place for round-1 matchups, they'd use the same system (or a similar one) for subsequent rounds. I also considered that maybe the round-2 matchups could be based on performance in round 1, with the players seeded by either their raw score in the round-1 game or by their score differential (how much they beat the second-place player in their game by). But when I checked, I did not see any such pattern.
One thing that I also wonder is, in a tournament where two (or more) players per table advance, is there any effort to keep players from competing against players they have already met?
