Page 6 of 22
Re: BGA has a serious bully problem
Posted: 03 February 2026, 18:02
by stillframe
Elo works fine for games with luck. The effect is that the total rating spread of the population, from beginner to expect, is less in total. If a particular game has enough luck that there is never a 90% chance to win, then the ratings will not reach a 400 point spread.
This is not a problem at all. The problem with the Elo rating system as applied on BGA is not that it was designed for zero luck, but rather that it was designed for 2 player games. The way BGA assigns rating changes at higher player counts was never intended. And so you get occasional situations where players are not trying to win the game, for rating reasons. It seems to work well enough usually, but not always.
Another possible problem with Elo ratings is that it works best in a large and active player population with well-distributed skill levels. This is often not the case for games on BGA.
Re: BGA has a serious bully problem
Posted: 03 February 2026, 18:09
by MarkSteere
ChatGPT...
Elo does not perform as well for luck-based games because it was designed around a skill-dominant outcome model. When variance dominates, the Elo probability distribution systematically overstates how predictive rating differences really are.
The issue isn’t that Elo “can’t handle luck,” but that it assumes luck is small relative to skill.
Re: BGA has a serious bully problem
Posted: 03 February 2026, 20:22
by FrankJones
MarkSteere wrote: ↑03 February 2026, 18:09
ChatGPT...
Elo does not perform as well for luck-based games because it was designed around a skill-dominant outcome model. When variance dominates, the Elo probability distribution systematically overstates how predictive rating differences really are.
The issue isn’t that Elo “can’t handle luck,” but that it assumes luck is small relative to skill.
Well, with all due respect to chatGPT, its answers are not always correct.
Re: BGA has a serious bully problem
Posted: 03 February 2026, 20:25
by FrankJones
stillframe wrote: ↑03 February 2026, 18:02
Elo works fine for games with luck. The effect is that the total rating spread of the population, from beginner to expect, is less in total. If a particular game has enough luck that there is never a 90% chance to win, then the ratings will not reach a 400 point spread.
This is not a problem at all. The problem with the Elo rating system as applied on BGA is not that it was designed for zero luck, but rather that it was designed for 2 player games. The way BGA assigns rating changes at higher player counts was never intended. And so you get occasional situations where players are not trying to win the game, for rating reasons. It seems to work well enough usually, but not always.
Another possible problem with Elo ratings is that it works best in a large and active player population with well-distributed skill levels. This is often not the case for games on BGA.
Another issue is that once players get to 100 ELo, it does not go down. And, we all start at zero. This skews the rating system to some extent.
Better would be to start everyone 1000 or something like that. And not have some floor below which the Elo calculation stops happening.
There are some games in which the better player would win in a large enough sample, but no one is interested in playing a real sample of hands. For example, Open-face Chinese poker. People play matches of 6 hands. Anything can happen in 6 hands, thus increasing the luck factor by a means having nothing to do with the actual game.
Re: BGA has a serious bully problem
Posted: 03 February 2026, 22:00
by Meeplelowda
FrankJones wrote: ↑03 February 2026, 20:25
Better would be to start everyone 1000 or something like that. And not have some floor below which the Elo calculation stops happening.
Long ago (before I joined BGA) it used to be like that. But you are assuming people really want to know what their true rating is. Some people can't handle that truth.
Re: BGA has a serious bully problem
Posted: 03 February 2026, 22:25
by RicardoRix
yeah it was 1500 IIRC (maybe 1200). I think they changed it around to make people feel better, you can only go up.
The ELO luck games thing has come up before, something about reducing the k-factor for these specific games would be the simple fix, although I think the range would still be narrower, but it's all relative.
Taking away ELO for co-op games was meant to help with the toxic.
Re: BGA has a serious bully problem
Posted: 04 February 2026, 00:30
by MarkSteere
FrankJones wrote: ↑03 February 2026, 20:22
MarkSteere wrote: ↑03 February 2026, 18:09
ChatGPT...
Elo does not perform as well for luck-based games because it was designed around a skill-dominant outcome model. When variance dominates, the Elo probability distribution systematically overstates how predictive rating differences really are.
The issue isn’t that Elo “can’t handle luck,” but that it assumes luck is small relative to skill.
Well, with all due respect to chatGPT, its answers are not always correct.
Oh definitely. I've seen ChatGPT be totally wrong. But on a pedestrian topic like Elo, on which a lot has been written, this seems fairly plausible, at least to me.
Re: BGA has a serious bully problem
Posted: 04 February 2026, 01:03
by MarkSteere
If you're asking ChatGPT to be creative on a topic on which little or nothing has been written, that's where it falls flat. But if you're just asking it to summarize or synthesize tons of existing material, it's remarkably adept.
Re: BGA has a serious bully problem
Posted: 04 February 2026, 01:04
by Andrei3009
MarkSteere wrote: ↑04 February 2026, 00:30
FrankJones wrote: ↑03 February 2026, 20:22
MarkSteere wrote: ↑03 February 2026, 18:09
ChatGPT...
Elo does not perform as well for luck-based games because it was designed around a skill-dominant outcome model. When variance dominates, the Elo probability distribution systematically overstates how predictive rating differences really are.
The issue isn’t that Elo “can’t handle luck,” but that it assumes luck is small relative to skill.
Well, with all due respect to chatGPT, its answers are not always correct.
Oh definitely. I've seen ChatGPT be totally wrong. But on a pedestrian topic like Elo, on which a lot has been written, this seems fairly plausible, at least to me.
Even if chat gpt is right, there's no point in using it in a conversation like this. I want to discuss with a person, with his own opinion, which he has built himself and fully understands. I could have typed the same thing into the gpt chat myself or asked him to issue counteraguments to yours (I partially disagree with the answer), but then I could just create a chat for bots and watch them...
Re: BGA has a serious bully problem
Posted: 04 February 2026, 01:07
by ChiefPointThief
FrankJones wrote: ↑03 February 2026, 20:22
MarkSteere wrote: ↑03 February 2026, 18:09
ChatGPT...
Elo does not perform as well for luck-based games because it was designed around a skill-dominant outcome model. When variance dominates, the Elo probability distribution systematically overstates how predictive rating differences really are.
The issue isn’t that Elo “can’t handle luck,” but that it assumes luck is small relative to skill.
Well, with all due respect to chatGPT, its answers are not always correct.
Well it is very accurate on this topic. Me & others have shown examples and given data on this topic on several threads here.