Board Game Arena

RobertBr wrote: ↑21 October 2024, 13:45

BAJO Staff wrote: ↑20 October 2024, 00:20

In the situation in which player quits or is expelled, the advancing player is chosen by oldest account.

It seems not fair for the other player. Is there a suggestion that have been opened about this issue to improve the situation for 3 players tournaments ?

I have never seen Meeplelowda provide any evidence for their claim and people on the forums have certainly claimed counter examples that indicate it is randomly determined. Probably one of the biggest issues is clear information from BGA that specifies how various time breakers etc are resolved.

I think it's me that's usually pushing the theory that the newest accounts are eliminated first when an overdue player has been skipped. I base this on OVERWHELMING observational data, and am convinced that any exceptions are just weird glitches that occasionally pop up. If it was actually supposed to be random, I would probably join those random number conspiracy theorists that frequent the forums

RobertBr wrote: ↑21 October 2024, 13:45 I have never seen Meeplelowda provide any evidence for their claim and people on the forums have certainly claimed counter examples that indicate it is randomly determined. Probably one of the biggest issues is clear information from BGA that specifies how various time breakers etc are resolved.

In multi-player elimination tournaments, there is strong evidence that ties are broken by age of account. As Shivaware says, there is overwhelming observational data to support this, in the case of multi-player elimination tournaments. For example, here's a Catan tournament, three players per game of whom only one advances: https://boardgamearena.com/tournament?id=311268 Looking at the standings, I see 18 games listed with a tie for first place. (They must all have been cases of expulsion, since in Catan there's no other way to tie for the win.) On a computer's browser you can mouse over the name of any player in the standings, and the link will show the player ID number. I assume that the ID numbers are issued by BGA in sequence, so that a smaller id number corresponds to an older account. I haven't checked all 18 tied games in this tournament, but in the ones I looked at, it's always the older account that advances.

Are there any counterexamples from a multi-player elimination tournament?

In round-robin tournaments, BGA support has apparently said that if there is a tie in the tournament standings, the ranking of those tied players is determined randomly. Source: viewtopic.php?t=29137

But these are two different questions:

1. in an elimination style tournament (whether multi-player or single elimination), if the result of one game is a tie, how is it determined which player advances to the next "step"? Answer: age of account (per observation, see example above).

2. in a "groups" format tournament, if there is a tie in the standings of Phase 1, how is it determined which player advances to Phase 2? Answer: random selection (per BGA support, as reported in the thread linked to above). This presumably applies where Phase 1 of the groups tournament is either round-robin or Swiss-style. Edited to add: in Swiss-style tournaments BGA has documented tiebreakers of (1) opponents' tournament points and (2) opponents' average skill as measured by an archaic version of Elo. It seems like it would be exceedingly rare to still have a tie in the standings of a Swiss-style tournament, after applying those two tiebreakers.

Meeplelowda wrote: ↑21 October 2024, 10:30 This is correct if time simply expired.

But it also says "Player Phynec was out of time (or quit this game) and lost this game." The text saying "or quit" confuses things because actually quitting, i.e., going into the menu options and selecting "Quit this game," is different from merely running out of time and is handled differently.

I think in this particular game, a player was expelled and time expired. Because typically, after a player's turn is skipped, the other players can continue playing; they have a choice whether to quit without penalty or continue playing, but either way they will tie for the win.

The tournament conditions (https://boardgamearena.com/tournament?id=327881) specify a maximum game duration of 7 days. So each player would have 56 hours in their time bank.

Looking at the game log (https://boardgamearena.com/gamereview?table=575087186), the game commenced 2024-10-13 at 12:11:19 (in my time zone). At move 24 (four days later), Phynec was out of time, and at move 32 (2024-10-18 01:06:16) JR0275 skipped Phynec's turn. It looks like JR0275 and ElisabethBT continued playing , and were not finished by the seven day mark, so at move 71 the game was declared over. But move 71 has a timestamp of 2024-10-20 at 12:10:07, which is 72 seconds early! I don't understand that. I would have thought the server would wait until 12:11:19 on Oct. 20 before terminating the game.

RobertBr wrote: ↑21 October 2024, 13:45

BAJO Staff wrote: ↑20 October 2024, 00:20

In the situation in which player quits or is expelled, the advancing player is chosen by oldest account.

It seems not fair for the other player. Is there a suggestion that have been opened about this issue to improve the situation for 3 players tournaments ?

I have never seen Meeplelowda provide any evidence for their claim and people on the forums have certainly claimed counter examples that indicate it is randomly determined. Probably one of the biggest issues is clear information from BGA that specifies how various time breakers etc are resolved.

The best evidence would be a BGA admin actually stating definitively in the FAQ how it works. But because they refuse to give us this information that they seem to be protecting as if it's some kind of trade secret, we can only go by what we observe.

I used to be in your camp of believing we don't know for sure, but have finally been convinced that there are too many instances of the older account advancing for it to be coincidence.

Mathew5000 has given a more full response above that covers cases that are random (and confirmed to be random by someone official), and it's possible that people have been confusing those circumstances with the circumstance of someone being expelled.