Board Game Arena

The normal matchmaking in single-elimination tournaments always follows the same pattern, the player with highest Elo is matched with the player with lowest Elo, 2nd-highest Elo is matched with 2nd-lowest Elo, and so on. But you may wish to also have tournaments where the matches are completely random. I noticed you can abuse the multi-elimination framework of BGA to create such KO tournamets with random match draws. The way to do that is to choose Group stages as tournament format, then you can choose "multiplayer elimination" for one or both of the stages, and you can leave the players per game at 2! The advantage of choosing "multiplayer elimination" is that you can set the match assignment to "Random".

Voilà, you should have 2p KO tournament with completely random match draws. I want to test it out in this Can't Stop tournament: https://boardgamearena.com/tournament?id=560013

I was wondering what happens with seeding in uneven player groups when you insist on 2p games. Usually, the players with highest Elo are seeded to the next round, however, with the setting "Random", I would assume that the seeded player(s) would be chosen randomly. I want to test it out here:
https://boardgamearena.com/tournament?id=560016

As a comparison, I want to see what happens when I use typewriter or serpentine. https://boardgamearena.com/tournament?id=560019

I am not sure if it all works out or if there are some weird bugs with these settings. Let's find out. One consequence should be that you don't have the nice complete tournament bracket like in a usual 2p KO tournament, however, we should have the multiplayer elimination interface.

Maybe someone tried these things already and can link to it. I am just curious if and how it works out. And to be clear, I like it that a randomized version of 2p KO tournaments seem to exist!

So, after my test tournaments finished, I can conclude that random matchmaking for 2p KOs does work in principal.

The first tournament above (with 32 players) worked as expected, matchmaking seemed completely random overall. In the others (with 9 players), I expected the byes to be given either at random (with Random) or to highest Elo (with Serpentine or Typewriter). However, after closer inspection, it seems that byes were always given to the lowest player ids (oldest accounts). In the Serpentine setting, it happened that the byes were given to the lowest Elo which happened to also be the lowest player ids. In the Random setting, it could be that the lowest player ids were chosen by chance. However, we have overall 8 out of 8 instances where the byes were given to lowest player ids.

In order to confirm this, we would need to play more such "odd" tournaments, however, I am not interested enough in these odd ones to set up more of them.

Conclusion:
- You can set up single-elimination tournaments with random matchmaking by use of the multi-elimination framework (as described in the previous post).
- Not yet confirmed but plausible assumption: You can compromise how much random matchmaking you want. For instance, have group assignment at random but within each group, you set matchmaking by the usual "Serpentine". That means there is some randomness in the matches but still high-Elo players get matched with low-Elo players first. Or you can set random matchmaking only in the first or only in the second stage of your tournament.
- You have to live with the multi-elimination interface.
- If there is an "uneven" player count, byes seem to be given to lowest player ids (oldest accounts).

I would love to see more single-elimination tournaments with some amount of random matchmaking! Your tournament will have a different flavor.