> Capturing rules are pointless because you wait a turn and the captured stones would have died anyway. > There is no way to score at the end, because the concept of surrounding areas becomes broken when the surrounding bits either die or explode. > It's not a playable game because you are unable to read more than two turns ahead in any sensible fashion (on a related note, ask me about my travelling salesman game)
Well OBVIOUSLY it's gonna need a few rules changes over either conway or go. Let's see now. The concept of surrounding an area still works - you'd need a much bigger grid for it to be practicable.
Yes, but it's nonsense because the walls will just die. You need to ignore the surrounding bit.
You also, fairly obviously, need to be able to place more than one stone per iteration of the life count. When a new stone is born due to a life step, it becomes the colour of the majority of its parents (three parents exactly, so that isn't a problem)
Not convinced by the former, but I took the latter as given.
I suggest that captured stones not be removed, but instead be turned grey (or something) - no longer permitted to be used as part of a capture, able to be replaced by a new stone, but still 'live' for the purposes of creating new stones (though those stones would also be dead)
The stones will die immediately anyway, by virtue of having more than three neighbours. Furthermore, I see little gain in adding a colour as this will break the ability to tell what colour a new stone is if we use the majority system. The alternative, that any stone born which has a dead stone as part of it means that grey will sweep the board and take over; if the game is played anything like towards conclusion, then the board will be filled with grey stones, with any black and white ones left over more by accident than by design.
Interestingly, I think this would allow some patterns of captured stones to break down their walls by overcrowding them.
I don't think that you will be able to create any walls in this variant, because one stone can break the lot.
Next up - the grid. Conway is played on an infinite grid. I don't see any particular reason to change that. I suggest seven go moves per conway move (maybe more for the first turn) - since this is enough to build a glider.
Hmm. Five is enough to build a glider. I'm not sure why you want to build a glider, though. Some explosive shape
like the following is more than enough. And to be honest, I think that with that many moves, you *really* have no control, as opposed to the limited control that you have with one move per life step.
Objective of the game becomes to place the most stones of your colour - non captured - on the board.
Ah, so we are ignoring the areas part of go. Good.
Either a turn limit, or a number of go turn limits, then run conway to convergence (or proved divergence - though setting up a glidergun without the opponents co-operation is gonna be a toughy.)
Does that seem workable? Any obvious flaws (other than ClinicalInsanity?, I mean) --Vitenka
My main concerns are your ever multiplying zombie army of dead pieces, and the fact that you have too many piece placements per turn. Plus my concerns about unreadability (see ClinicalInsanity?), and the fact that it's not Go except in that we place pieces on the board. --Angoel
It's only not go because you keep removing the 'capturing areas' thing - if you still deny being able to put pieces in a captured area (which was all I meant by greying them out) then yes, the stones inside will die quickly, but also you can't place a piece to just bust them out. We may need to rethink the definition of a wall though, since corners explode. Mind you, many of the shapes of walls explode into other walls (for at least a few iterations) - so it might survive long enough to capture the pieces inside. I also meant that the play order would be:
(n(Black places, White places),conway iterates) - making building anything really nasty a matter of the other player letting you.
I don't keep removing it. I merely keep asserting that I see no way that you can get any rules for it that work. And, indeed, I have seen no evidence yet that you can do so. As an example, say in one game, the following situation comes up. White asserts that the black group in the top left corner is dead, and that white has therefore won. Black asserts the opposite. If you give me a clear rule to say whether the black group is alive or not, then I'll grant the point. --Angoel
Maybe instead of greying out an area, we make it that only the player owning the area can play into it? That ought to allow them to kill off the pieces still inside it easily enough. Area control remains important. --Vitenka
I think we need to redefine 'wall' - since whilst yes, it is easy to break basic structures, there are other structures which although they have apparent gaps in them are fairly stable. If we make those be the walls (uh, I dunno - use a 'zone of control' - each piece controls the intersections next to it as well, and if those controlled intersections form a wall then the area is dead?) then maybe something more go-like can be salvaged? --Vitenka
I think this can be made into a workable game. You do need to keep some concept of area control, but equally not require solid walls. I think Vitenka's idea works, where you forming a zone of control (you control empty squares adjacent to your pieces), and groups of yours whose zones of control touch are considered to be touching. Territory as enclosing areas with this zone of control. Dead pieces inside are captured, and you can't play stones into a region that is purely controlled by someone else. I think you do need to be allowed to place more than one stone per turn, because otherwise you can only play near existing groups which loses a lot of the Go feeling. But equally you shouldn't be able to scatter stones all around the board first. I suggest allow placement of up to three stones per turn (the minimum to form a stable group, and either a block or a spinner may be formed with three), but only within a 3x2 rectangle of each other. Turn sequence goes White places up to three, Life iterates once, black places up to three, Life iterates, etc. Best played using a computer. A good UI for this game would have two windows, one showing the current game state and proposed placing of your three pieces, the other being a preview of one timestep's time if the current chosen placements were accepted. Note that one can create a glider in two turns if the other player doesn't interfere: create a spinner on turn 1, then next turn place two further stones (decline to place the third) to convert it to a glider. --AlexChurchill
(AlanLawrence)I think maybe we need to change the rules of the automaton a bit - atm we are keeping ConwaysLife intact, and losing most of Go: groups, territory, capture, eye space. Life models a single organism - what do the two colours represent? Do we really want to treat them the same (so the colours are just used for determining who wins at the end), or maybe we can do something different for each - e.g. white plays a place, we do an iteration of Life for white's pieces only (ignoring black), then repeat? Do we want special rules for groups of the same colour, perhaps? Or a combination rule - new cells are born as in Life, but die as in Go (doesn't work well - things don't die often enough). Also, Go doesn't use diagonals, whereas Life does - one could resolve this by converting to a hexagonal board (valence of 6 being halfway between 4 and 8), but Life doesn't work as well there (have a play at http://www.hexatron.com/hexca/ - Callahan seems quite interesting). OTOH a "good" automata is one in which there's lots of interesting behaviour if the players dont interfere - not necessarily what we want at all; robustness might be better....
Good points - we may be taking too much conway here. We had one example of a rule for colour interaction given already. One important thing here is that the automata iteration must be simple enough for the players to predict it at least a few steps ahead. conway has this, where many others do not. --Vitenka
Huh? By my suggestions above, we have concepts of territory and connectedness, which extend straightforwardly to groups and capture. I agree that those are vital if this is to claim to be Go-like. Playing with the automaton rules is possible in theory, but ConwaysLife has the advantage that the rules are much more familiar to most potential players of this weird game we're concocting. There is potential for keeping the Conway iteration rules but doing other things with them, as you suggested: iterate black pieces after black's plays, for example. The hexagons are interesting but lose a *lot* of what makes both ConwaysLife and Go "feel" the way they do. --AlexChurchill
Agreed ConwaysLife rules are familiar - but otoh, I don't know any more shapes than Glider, box, and spinner, and that seems enough :-). I'm not sure that we can make a game that *is* either of Go or Life (i.e. a player from either would be able to use their existing tactics/structures), tho - we just want the funky behaviour/possibilities of both (strategical depth of Go, and...emergent complexity of Life?). But ok; maybe we want to leave the rules intact - I'm not convinced by the hexes, by any means, no.
Also, I'd point out that territory in Go isn't "you can't play into your opponents territory" - it's just that if you did, you'd die; territory doesn't exist in the rules until you come to scoring up at the end, and until then it's a consequence of the rules, rather than actually being defined in them. This is nice, because it makes the rules much simpler, and....
I have an applet which lets you play around with the rules of Life! and Go! http://www.srcf.ucam.org/~acl33/applet.html . Feel free to make suggestions, and to paste in rule sets for discussion (?!) here....
Pallando writes erk! What about Conway Sparklies ? Where the colour of unowned pieces mutates each turn according to some function of the colours of the four neighbours!
![[Home]](/img/text_logo.jpg){kind=logo}
ConwayGo
