Page 2 of 4
Re: LLM-Based Players
Posted: 12 April 2026, 20:40
by Strode
I don't particularly want to play against bots (though I play against them in some game apps), but I fear they're coming one way or another. I would favor official, regulated bots (forcing them to identify themselves, not letting them join without permission) if it kept out the rogue ones. But I doubt that would be the trade off.
Re: LLM-Based Players
Posted: 12 April 2026, 20:54
by Pocc
The only bot I'm aware of is the can't stop how-to-play bot. I don't believe there are non-official bots (that aren't banned), but someone correct me if I'm wrong.
> official, regulated bots (forcing them to identify themselves, not letting them join without permission)
Reading through the thread, I don't think anyone is in favor of bots that join without permission. This would degrade the existing, good experience of playing with other humans you want to play with.
Bots could be trivially identified with bot_{model}_{game} with a username like bot_kimi2.5_chess.
@cigma thanks for the guidance on asking via official channels - I've asked.
Re: LLM-Based Players
Posted: 13 April 2026, 07:28
by Ingoschi
Are there any examples yet where a LLM-based bot plays a boardgame better than e.g. a rule-based?
Re: LLM-Based Players
Posted: 13 April 2026, 10:00
by Jellby
Pocc wrote: ↑12 April 2026, 20:54
I don't believe there are non-official bots (that aren't banned), but someone correct me if I'm wrong.
There was a bot for Boop, created by a user for a project, and allowed by BGA. I don't know if it's still around.
Re: LLM-Based Players
Posted: 13 April 2026, 12:42
by KuWizard
GTSchemer wrote: ↑12 April 2026, 14:42
lefalaf wrote: ↑12 April 2026, 11:48
Glad to hear that BGA doesn't want to have AI on the platform. I hate how LLM AI slop seems to be infecting nearly everything.
I'm only one opinion, but I have zero desire to play with anyone but real humans.
I may be in the minority but I agree. I would be incredibly disappointed if I tried to play a game and ended up with a bot.
What if you'll never know? Any opponent can be a bot nowadays... And you might have already played with one.
And I don't think you're in the minority. Many large industries are suffering from understanding if something is AI-generated or not. Most people want to know if a human produced music/video/whatever they are digesting
Re: LLM-Based Players
Posted: 13 April 2026, 12:57
by tbhp
KuWizard wrote: ↑13 April 2026, 12:42
What if you'll never know? Any opponent can be a bot nowadays... And you might have already played with one.
This reminds me of the question "how do you know other people aren't just p-zombies?" We have no way to know, but we sure don't want it to be true.
Re: LLM-Based Players
Posted: 13 April 2026, 13:52
by Arjuna92
Ingoschi wrote: ↑13 April 2026, 07:28
Are there any examples yet where a LLM-based bot plays a boardgame better than e.g. a rule-based?
ZeusAI plays 7 Wonders Duel very well. It beat several top players in bo5 matches. Here's the
thread about that.
Re: LLM-Based Players
Posted: 13 April 2026, 14:07
by Andrei3009
Arjuna92 wrote: ↑13 April 2026, 13:52
Ingoschi wrote: ↑13 April 2026, 07:28
Are there any examples yet where a LLM-based bot plays a boardgame better than e.g. a rule-based?
ZeusAI plays 7 Wonders Duel very well. It beat several top players in bo5 matches. Here's the
thread about that.
ZeusAi is not a LLm, it's neural network crated using Deep Reinforced Learning
Re: LLM-Based Players
Posted: 13 April 2026, 14:15
by Arjuna92
Andrei3009 wrote: ↑13 April 2026, 14:07
Arjuna92 wrote: ↑13 April 2026, 13:52
Ingoschi wrote: ↑13 April 2026, 07:28
Are there any examples yet where a LLM-based bot plays a boardgame better than e.g. a rule-based?
ZeusAI plays 7 Wonders Duel very well. It beat several top players in bo5 matches. Here's the
thread about that.
ZeusAi is not a LLm, it's neural network crated using Deep Reinforced Learning
That means it wasn't given any strategy advice but simply told to play the game a huge number of times and find out when it happens to win against itself, gaining an immense understanding of game positions and strategy but not necessarily of tactics or logic, right? So would an LLM just try to find a legal move by parsing the rulebook?
Re: LLM-Based Players
Posted: 13 April 2026, 14:50
by cigma
Pocc wrote: ↑12 April 2026, 20:54
The only bot I'm aware of is the can't stop how-to-play bot. I don't believe there are non-official bots (that aren't banned), but someone correct me if I'm wrong.
I am not sure if it still exists, see also
https://boardgamearena.com/forum/viewtopic.php?t=30279 and
https://en.doc.boardgamearena.com/Bots_ ... telligence