Board Game Arena

Hi,

i just joined the tournament : itsawonderfulworld • Monthly Knockout IX - September 23
and noticed that there are about 10 players in the 1st round who are alone at their table (and so getting immediately in the 2nd round).

Is there any reason for this ? Why weren't these players recombined to form 5 more full tables ?

thanks for any insight ...

I would say it's because of the bracketing. If you have six players in a tournament and want to do matches of two with single elimination, you either have two matches and two byes (skips) in the first round, or if you do three rounds, then the next round has to allow someone a bye into the final. You're better off doing your byes as far out from the final as possible, so I would guess BGA says to get down to the right number of matches at the end, we need this many matches (and has to be a power of 2 - I'd guess it's done 128 matches, so 109 matches, 19 byes).

nik592 explained it correctly - one wants to have byes only in the first round. The reason was given by nik also.

I only want to repeat his arguments with two extreme examples with higher player numbers. Let us first assume that 96 players enter a tournament.

1. solution (implemented on BGA)
1st round: 22 games, 52 byes -> 64 =2^6 players advance
No more byes necessary, the number of players in each round will be 96 -> 64 -> 32 -> 16 -> 8 -> 4 -> 2 -> 1

2. solution: only award a bye if there is an odd number in a round. The numbers of players in each round will be
96 -> 48 -> 24 -> 12 -> 6 -> 3 (one bye) -> 2 -> 1

The 2nd solution has only 1 bye instead of 52 byes. But it is still considered inferior because at the end of a long tournament where hopefully the best 3 players are finally determined you do not want to favor one of these three.

Abother extreme example would be 65 participants.

BGA would let 2 players play an extra round (63 byes) and reduce to 64 players. The tournament in effect starts with round 2, only 2 people were unlucky and had to play an extra game. It is fair for 63 of 65 players.

If you were to award only 1 bye per round the numbers would go: 65 -> 33 -> 17 -> 9 -> 5 -> 3 -> 2 -> 1
You have a bye in every round till the finals.

Thank you for your explanations.

euklid314 wrote: ↑15 September 2023, 16:41 BGA would let 2 players play an extra round (63 byes) and reduce to 64 players. The tournament in effect starts with round 2, only 2 people were unlucky and had to play an extra game. It is fair for 63 of 65 players.

I wouldn't use the word "unlucky" there, because on BGA, who gets a bye in Round 1 of a single-elimination tournament is determined by elo (not by chance). You can see this in the tournament OP mentioned: https://boardgamearena.com/tournament?id=231030 Under "Participants", the players are listed in order of elo, and you can see that the players with the highest elo are the ones who received first round byes.

You can also see that the bracketing was set according to elo at the start of the tournament; the two players with highest elo (jsnoop36 and primoclark) cannot face each other before the final round.

In the hypothetical tournament with 65 entrants (or 33, or 129, or 257), it will be the two players with the very lowest elo who play in Round 1 while everyone else gets a bye. The winner of that game will face the highest-ranked player (by elo at the start of the tournament) in Round 2.

I don't know whether that is documented anywhere. Probably most players are unaware that single-elimination tournaments on BGA are seeded like this.

When I first started playing single-elimination tournaments on this site, it was pretty discouraging to always get clobbered in Round 1, and it took me a long time to notice that because my elo was so low, I would always be paired with one of the top players in the tournament in my first game. Now when I'm learning a new game, I don't play single-elimination tournaments until I get my elo up to 150 or 200.

euklid314 wrote: ↑15 September 2023, 16:41 nik592 explained it correctly - one wants to have byes only in the first round.

I wish it would work like this with multi-player games as well

I'm currently playing in three-player Azul tournament where two players advance. Now there's seven players left which seems to result in:

2 tables + 1 wildcard ->
1 table + 2 wildcards ->
1 table + 1 wildcard ->
final table

This should be quite easily preventable, right?

EDIT: For some reason I didn't think that the format in question will always result in a wildcard getting to the final table, so maybe the format itself is a little flawed or, at least, unfixable by the above standards.

pahakuukkeli wrote: ↑17 September 2023, 11:17

euklid314 wrote: ↑15 September 2023, 16:41 nik592 explained it correctly - one wants to have byes only in the first round.

I wish it would work like this with multi-player games as well

I'm currently playing in three-player Azul tournament where two players advance. Now there's seven players left which seems to result in:

2 tables + 1 wildcard ->
1 table + 2 wildcards ->
1 table + 1 wildcard ->
final table

This should be quite easily preventable, right?

Unfortunately not! The situation is much more difficult!

If you want to have 3 players on the final table there is no constellation in the semifinal without a bye. Proof: A round without a bye will always result in an even number of advancing players, thus never to an odd number like 3.

If you stick to "only 3-player tables" and "exactly 2 players advance" in each round, you will have many byes in almost every round - and you cannot help it. It is almost impossible to have 2 rounds in a row without byes. 9->6->4 is such an example but only a round with a bye leads to the number 9 and only a round with a bye will continue after the number 4.

euklid314 wrote: ↑17 September 2023, 11:38 Unfortunately not! The situation is much more difficult!

It seems I was too slow to edit my post, and acknowledge my brain fart. You are absolutely right that the specific issue in my post is 100% unfixable.

Actually that 2-out-of-3 advances format doesn't still work properly.

After 7 players were left, it should be:
2 tables + 1 skip
-> 1 table + 2 skips
-> 1 table + 1 skip
-> 1 final table

While in reality it is:
2 tables + 1 skip
-> 1 table + 2 skips
-> 1 final table (with only 1 player from the previous table advancing with both of the skips)

There should either be that one extra round, or the 2 skippers should play a 2-player game and the winner advances with the top-2 of the 3-player table. The current implementation doesn't make any sense.