I think knowing that discobot would give duplicate numbers @yebellz couldn’t roll just 7d7. I assume 20 was picked as probably large enough, but maybe there was some maths thought behind it 
I think just read left to right and find the first occurrence of each number in sequence to get a permutation.
First is 2 then 5 (twice but that doesn’t matter) then 7. The next number not appearing previously is 4 and so on.
Yes, that’s what I did
By the way, 20 is the maximum number of dice that disco can roll in one request.
Here’s a bad idea for a future game: always determine who moves next with a die roll 
A more sane variant of that would be to re-randomize the order after every player has played another stone.
I wonder if some sort of bidding system could be used?
The weird thing is that in the end game, one may be happy to pass, while in the early to mid game, one would might like another more sooner rather than later.
An earlier move also lets you account for the move order, whereas a late move makes you too the dice on who’s moving next.
I also considered that variant, but I honestly don’t mind the risk of more “unfairness” - these multiplayer games are never really fair anyways, but adding more randomness could make the unfairness easier to handle psychologically, I think.
Yes, part of the fun would be that a given “end” state can’t be solved by just counting territory - even between just 2 players it would be something like “player A has a 20% chance to win this territory filling stage”. My experience from other variants is that this uncertainty makes it easier to enjoy the game casually, since playing perfectly is even less realistic than in normal go 
Ooh; I like the rationale.
It’s easier to accept being betrayed by the dice than a person.
Other ideas (not necessarily to be applied simultaneously), organized with sub-options for each concept:
I like keeping things simple (the “roll after every turn” is actually simpler than fixed turn order in a way, since there is less state to track) - buuuut let me add this idea to the mix of complications anyways:
If there are some chains in atari on the board, the next player is picked at random only from those players. (otherwise picked randomly from all players)
The main idea here is to make capturing by teaming up harder, as soon as you atari someone they get a chance to defend themselves. Still you won’t survive long if multiple players are out to get you, but you will have more of a fighting chance.
A bug neat side effect is that there could be som fun tactics where you put your own chain in atari.
Black could play G2 here for a 50% chance to eliminate white:
…the drawback is that there is a 50% chance that it backfires.
Here is a deterministic one - black could safely play D3 and then C3:
(assuming no other chains on the board in atari)
Black could safely play D3, then proceed to play 100 moves in a row without putting anybody else in atari.
Good catch!
You could mostly fix this by never giving a player two moves in a row (so in the case where you put yourself and someone else in atari, the turn will always go to the other person).
It would still be problematic if two players can set it up such that they both have a chain in atari, and they just alternate free moves between them. Once they have a solid advantage they can save their chains and let the game continue (they will want to do that eventually, because they shouldn’t atari someone else while they’re still in atari themselves).
But how would you set this up in the first place? I think it would always involve a considerable risk to leave yourself in atari hoping for this scenario.
And assuming you do get into this scenario, what do the incentives look like between the two players? Once they have both gained a small advantage, they are each others biggest threat in terms of winning the game. So it would be desirable to eliminate the other player pretty quickly. However, you don’t want to capture their chain if this would leave your own stone vulnerable.
So this further restricts the “setup” you need to have your own chain in atari in such a way that the player you pair up with can’t capture it safely.
All in all, I think this modified version (where you never get too moves in a row) is hard enough to “break” that it would be pretty impressive if someone managed to set it up. (Especially if playing with zero communication)
Let’s say we have something like this and it’s black to play:
Black could play E1, hoping that white gets the next move and that white then chooses to play E9:
From here, black and white could safely alternate for a while. It wouldn’t be in blacks interest to capture E9, because that would leave both E1 and F9 vulnerable.
However, black playing the initial E1 stone is a big gamble - if red gets the next move black will lose immediately. And even if white gets the next move, white could play some other random move, making black spend a slow move at D1 to save the stone.
To more thoroughly “patch out” these strategies, would it help to make a rule that if you have one or more chains in atari, you have to add a stone to one of those chains? (normally this would just mean “you have to escape from atari”, but it could happen that you’re still in atari after playing your move)
I’m sure even that patched version has some problems though, so I’ll stop here 
I mean one can come up with all kinds of rules, like players get points based on the position they place and then points for secret shapes they have to make given prior to the game.
I don’t think I’ve as much interest in all these versions more than the current capture go, minus the diplomacy aspect.
I think the diplomacy might be interesting in one way, but it’s also slightly tedious to have to clear your actions with most players for fear of being eliminated in one move.
It would probably be more enjoyable with just chat but no direct strategic plans, collusion, pacts etc. It would mean one could invade attach etc without immediate fear 3-4 players have made a non aggression pact. I mean sure it might make sense to remove the aggressive player anyway, but it won’t be by prior agreement, and there’s no confirmation of others ideas until they put a stone down.
I think people will probably play for themselves more in the end, joining in on an attack if they can profit, avoiding helping for a similar reason.
Exploring variants that play with variable move order is certainly worthwhile, and I believe there are definetely some gems waiting to be found, but one has to be careful not to venture too deep into the thicket of broken combos 
Yeah, I wanted to have some fun exploring the implications, but it definitely got too complicated for my taste along the way. It still feels worthwhile to me to think about this kind of thing though, it makes it easier to come up with and analyze new variants in the future when you have a sort of “toolbox” of mechanics that you have some experience with
Inspired by the Blooms game mentioned in another post (2-player 4-color go ⚫ ⚪ 🔵 🔴 - #18 by LuisB), another interesting variation upon the rules would be to eliminate players only when X of their stones have been captured.
I think it would change the dynamics of the game quite a bit, with different natures across different sizes of X. Fundamentally, it is still a territorial game, but now the players would have more freedom and incentive to make a smaller sacrifice to help capture a larger group.
How about combining the concept of Battle Royale Go with Fractional Go? Where a player is eliminated as soon as for each of their colours, a stone featuring this colour has been captured.
For example if all players place stones of two colours, and two players share a colour. Then a player gets eliminated when one or more of their own stones get captured, or when both of their “allies” have had one or more of their stones captured.
(Or alternatively, the elimination rule stays the same. We don’t need to complicate the rules too much )
