|
||||||||||||
|
Title: Fairy Pieces Post by IdahoEv on Apr 10th, 2006, 2:00pm As an amateur game designer myself, it doesn't take long with any new game before my brain starts musing on the topic of "hey, I wonder what else we could do with this...?" (Note that this is no way disrespectful of Omar's beautiful creation. As someone who's designed several games, I can say that Arimaa is practically the holy grail: it manages to be both fun and strategically deep in an extremely simple set of rules that can be learned in minutes. And, it's an abstract strategy game - the hardest kind to design; flavor and story are crutches that nearly all game designers depend on.) Anyhow, on an airplane this past weekend, I couldn't play, so instead my mind was mulling over Arimaa. Practically the only thing that has ever nagged me about the design is the near invulnerability of the elephant; I feel like the strategic depth would increase if the elephant weren't so invincible. So, tapping into a particular cultural legend about elephants, I came up with two possible fairy pieces for Arimaa: The Mouse: The mouse is an officer, moves like any other non-rabbit piece, and cannot score goals. For purposes of pushing, pulling, and freezing, it is "smaller" than all Arimaa pieces, including the rabbit ... except for the Elephant. It is accounted larger than the elephant, and can freeze, push, or pull the elephant. The Rat: Identical to the mouse, except that it is accounted as the "same size" as a rabbit, i.e. it cannot push, pull, or freeze a rabbit, and vice versa. I came up with the rat when I worried that it might be too easy to just push the enemy mouse near your rabbits and never worry about it again, making it fairly impotent. Playtesting would be necessary to tell which of these was the more interesting piece. A few thoughts on possible ways you would use these:
The first two options retain the feature that the game can be played with a chess set. Just thoughts ... hope the purists aren't too offended. :-) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 10th, 2006, 2:36pm Your propopsed rule ideas are very interesting, and it would be fun to experiment with them. I remember as a child when almost the only game I knew was war (a card game of no skill), I thought it was dumb that nothing could take the top card, so you could only lose that card as a concealed card in a war. I always preferred to play that an ace beats everything except a joker, a joker beats everything except a deuce, and a deuce loses to everything except the joker. Thus my deuce was your mouse. This change made the game much more dramatic, and made a comeback possible from the weakest of positions. on 04/10/06 at 14:00:37, IdahoEv wrote:
I think that a rodent or two in Arimaa would increase the drama of the game. However, I suspect it would correspondingly lower the strategic depth, and make it rather easier for computers to beat humans. At present humans can outplay bots at all phases of the game except goal search, which is the most drastic and tactical part of the game. As bots improve, I can imagine them also surpassing us at wide open games when many pieces are threatened, although we can presently dominate such positions. What I have the hardest time imagining is how computers will beat us in games where the elephants become deadlocked around the same trap. The very fact that the two elephants are equal and invincible slows the game down and makes it more strategic. No matter what you threaten to capture with your elephant, I can make it safe with my elephant. This means that to make progress when our elephants are deadlocked, you have to overload my elephant by operating in antoher quadrant to create a second threat. Computers are just hopeless at this. Either they are too passive, and you can increase the pressure on them until they crumble, or the bots are too aggressive and you can pick off a piece they unwisely expose. The fact that any piece in chess can kill any other piece makes chess fast and tactical. Chess wouldn't be strategic at all if facing pawns could capture each other too, but luckily for chess the pawns can deadlock. In Arimaa, the fact that neither elephant can hurt the other and they can stalemate each other makes Arimaa much more strategic than it would be if, say, equal pieces could push and pull (but not freeze) each other. I admit, this is all speculation on my part, and we would actually have to playtest the different versions of your rodent rules to see what they did to the game dynamics. I have, however, seen the balance and depth knocked out of a few games by players adding rules to make the play more dramatic and interesting. Things that make a game "cooler" for the first dozen times you play usually seem to make you not want to play more than a dozen times that way. Just my intuition. |
||||||||||||
|
Title: Re: Fairy Pieces Post by jdb on Apr 10th, 2006, 3:21pm An interesting idea! I think that a rodent would dramatically increase the value of a hostage. Lets say gold takes a silver horse hostage. Normally silver sends in his elephant and creates a stable position. If gold had a rodent handy, silver's elephant would need some help to protect the hostage. The lone elephant play would also be very difficult since the rodent could venture about and freeze it. However there would certainly be new strategies that make use of the rodent too. Two or three rodents could attack an elephant and cause alot of trouble. It would also be easier to blockade the elephant since the rodents could assist by positioning the elephant on the desired square. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 10th, 2006, 3:41pm on 04/10/06 at 15:21:44, jdb wrote:
Good point. Lone-elephant attacks are definitely out if there are two rodents on defense. I wonder what sort of attacks are still feasible. I was blathering about rodents making the game faster and more tactical, but suddenly I wonder if they would make it hard to go on offense at all. Perhaps some kind of swarming attack could still work if it involved a rodent in the attacking force, but it isn't clear to me why the more-numerous defenders wouldn't always win. |
||||||||||||
|
Title: Re: Fairy Pieces Post by IdahoEv on Apr 10th, 2006, 7:14pm I think the lone-elephant sort of position was the item that inspired the concept. The fact that an elephant can hunt completely alone without worry feels "too simple" to me. Fritzlein's points on how the balance of the elephants creates strategic depth are excellent, and no doubt demonstrate a more thorough understanding of the game than I possess. Intuitively, to me, the struggle to create two simultaneous threats doesn't feel particularly deep, since the game essentially always comes down to that particular goal: make two threats, since the enemy E can always keep one safe. My thought was to 'complexify' the situation by requiring the player to pay attention even to the "automatically safe" trap normally defended by the E; all positions require defense at all times. But I do see how that may qualify as 'more complex' only to a human. The rodent concept may be adding tactical complexity at the expense of strategic. To a relative noob like me (especially since I think I'm better at strategy than tactics), that would seem on the surface like a deeper game. I do think, however, that the option to replace N rabbits with N rodents would add depth (but of which sort?) in the sense that a game with 8 rabbits and 0 rodents is a fundamentally different game than one with 4 and 4; material-based strategic evaluators would need to be very different. I suspect in a game with 4 rodents, the camel is probably more valuable than the elephant, but rabbits are suddenly priceless. Computer strategies would need to become at least a little more adaptive, and opening books would get an additional factor of N larger. Maybe this is worth investigation in Zillions of Games at some point. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 10th, 2006, 8:44pm on 04/10/06 at 19:14:15, IdahoEv wrote:
I see the problem there. If it had turned out (or still turns out) that lone-elephant is the dominant opening strategy that squeezes out all more ambitious strategies, that could sap Arimaa of some of its current excitement. It was good news for Arimaa that I didn't win the World Championship with my lone-elephant openings, and that Robinson, Adanac, and PMertens, who took the medals, all play more varied and aggressive openings. At present the balance between lone-elephant opening and elephant-horse openings seems very interesting to me, and very unclear. The elephant-horse attack is the more forcing of the two, and compels the lone elephant to hurry home, but who stands better in the resulting position? If the top players don't agree, it's hard to call it "simple". Indeed, there even is a persistent fringe that is willing to attack with the camel in some way, including Adanac and Blue22 in my postal tournament games. This is further evidence that we haven't by any means plumbed the depths of Arimaa opening strategy, never mind the strategy of the full game. Perhaps I try so hard to be clear in my writing about Arimaa, I give the false impression that the game is pretty well understood by now. But it is definitely not the case that all of us are trying to do the same thing, and whoever does it better wins. On the contrary, we're often trying to do very different things. In many of my games I discover that my opponent and I are angling for the same position! For example he will give up a horse hostage and think he is winning, whereas I will take the horse hostage and think I am winning. Or perhaps he will give up a camel hostage for a cat capture and think to have gained, whereas I will take the camel hostage for a cat capture and think to have gained. It seems to me very exciting that semi-stable positions (such as frames and hostages) can tip in either direction. I love it that I can sit down to either side of a position in which an elephant holds a horse hostage while the other elephant guards, and from either side thrash the best computer players. The excitement inherent in semi-stable positions is likely to be more enduring than the sort of excitement that prevents semi-stable positions from arising in the first place. Quote:
Wait, wait, there a certain amount of complexity involved in even creating the first threat. And after the first threat is created, the one which ties both elephants down near the same trap, the second threat is vastly more subtle. If you think the second threat is boring, take a look at game 24224, the fifth game in this year's Challenge match. Adanac attacked c3 with EH, so Bomb took the horse hostage. That created a first threat on each side and an elephant deadlock by move seven. Then Adanac attacked f3 with his camel for the second threat. Would this game be a rehash of his crushing win in game 23918? No! Bomb switched wings with its elephant, throwing the board into chaos and gaining an advantage in the process. The "create a second threat" phase of the game is usually intense because the deadlocked elephants are free to run across the board at any time and smash up the second sphere of action. The elephants are preoccupied, not frozen. In fact, it is basically only computers that let you execute the second threat. Human opponents will see the writing on the wall in advance, and complicate the situation because it is their only chance. Add to this the common possibility of undermining the first threat (e.g. breaking a frame, or flooding an elephant that holds a camel hostage), and you've got a lot going on. Quote:
I can see the urge to 'complexify', but if your new rule means that the primary situation never stabilizes, you might just have simplified the strategy by eliminating the second threat. If Arimaa contained only one hot zone at a time, it might become primarily a tactical game. Particularly if there were continual capture threats, it might become hard to build up incremental (strategic) advantages. Arimaa might devolve into a situation where you either have a forced capture or you have nothing to hang your hat on, because the situation is too unstable. By all means, if Arimaa is boring to you as it stands, try to inject some excitement. I don't think the rodent rules will necessarily make a less deep game. We would have to play it to see. I certainly couldn't have predicted the subtlties inherent Arimaa just from reading the rules, so I shouldn't pretend to know how altering the rules would alter the game. But for my money, it will be soon enough to tweak the rules when everyone agrees on the best strategy to the standard game, and when winning at Arimaa becomes only a matter of blunder-free execution of things that everybody knows. For as long as we have World Championship games in which the spectators can't agree who is winning, why they are winning, and what a good move would be, then I say Arimaa doesn't need to be spiced up. ;) |
||||||||||||
|
Title: Re: Fairy Pieces Post by IdahoEv on Apr 11th, 2006, 2:40am Just to clarify - I don't have any feeling whatsoever that Arimaa needs to be tweaked or complexified; as I said up top, I think Omar has quite achieved the holy grail in game design here. Mostly I just think about variations because it's an interesting exercise (to me) to do so. (I'm the sort of person who will often play a game just ten times and then switch to designing maps/levels/mods/variants/bots because the creative challenge interests me even more than playing ...) For toying with Arimaa's gameplay, I was drawn to the elephant because it felt unusual to me that it was so nearly invincible, and i liked the elephants-are-scared-of-mice flavor as a conceptual approach to that particular issue. I don't in any way think that's something the game needs, and I'm not suggesting this as a replacement for the official rules any more than fairy chess pieces like the Archbishop or Nightrider are intended as rules changes to be considered for adoption by the FIDE. I do think it's quite likely your analysis of the rodents' effects on the game are correct; you pointed out some consequences of the Elephant's invulnerability that I hadn't thought of, and your arguments are pretty sound. This is just an entertaining intellectual exercise. ;) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 11th, 2006, 3:10pm Designing variations is indeed an interesting exercise. In fact, I understand that nearly all games (except more modern games) evolved and improved over time. Poker used to be a rather dull game, and not very deep, compared to the subtleties of seven-card stud and Texas hold'em. In fact, Arimaa has already evolved slightly. The rule for breaking ties in elimination tournaments has changed, and the repetition rule has changed. Also the victory condition has been clarified. By the way, new user Unic posted over at BoardGameGeek about the rules of Arimaa. (I believe he hasn't yet joined the Arimaa Forum.) His beef with the rules is that repetition should be a draw, not a loss for one player, since it is hard to keep track of who it is going to be a loss for. I like Arimaa being essentially drawless, but I would support changing the rules in this related way: rather than making 3-fold repetition a loss for the repeating player, make it an illegal move, just like it is illegal to move into check in chess. (For the ultra-nitpicky, if your only legal moves would cause threefold repetition, then you would lose.) Rather than have 1 or 2 percent of games ending due to repetition, I think we should force those games to continue, with the server/interface enforcing the rule. That way players wouldn't have to do the bookkeeping to avoid repetition; instead they could deviate whenever required to do so by the rules. |
||||||||||||
|
Title: Re: Fairy Pieces Post by RonWeasley on Apr 11th, 2006, 3:43pm If the third repetition were an illegal move not allowed by the game interface, there would need to be some indication so the player would not mistake this for an error in the game interface. In wizard's chess, if you try to make an illegal move, the offending piece has to clean up after the knights' horses when the game is finished. |
||||||||||||
|
Title: Re: Fairy Pieces Post by omar on Apr 12th, 2006, 1:04am Thanks for bring up this suggestion Evan. I think we think alike :-) While developing the rules I also felt that perhaphs the elephants should not be so invincible and in fact experimented with a variation in which the eight rabbits/mice could push/pull/freeze the elephant but not any other piece. I don't remember now the exact reason, but I rejected that version and felt that the additional exception it introduced to the rules was not worth it. I never tried the variations you proposed. Whenever a rule is changed it can significantly change the dynamics of the game. For me it is a lot easier to play lots of games against myself and get a feel for the nature of the game then to simulate in my head what effects the rule changes would have. I am very interested to try out your first two proposals since they can still be played with a standard chess set. However, I won't be able to in the near future. I would encourage you and others to try it out if you have the time. Perhaps you could report back on what you thought of it and if it seemed interesting more people would try it. If many people feel it improves the game, we could change the official rules to adopt it. I am always open to changes which could improve the game while maintaining the original objectives (playable with a chess set, difficult for computers, simple rules, interesting for humans (very subjective)). |
||||||||||||
|
Title: Re: Fairy Pieces Post by omar on Apr 12th, 2006, 1:16am on 04/11/06 at 15:10:56, Fritzlein wrote:
But this would create a loop hole for mutually agreed draws which we try to forbid in the match rules. Quote:
I like this much better, because I perfer to see games continue to a natural finish (goal or immobilization) rather than have an abrupt ending. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 12th, 2006, 1:49pm To further expand on my proposal to make a third repetition illegal rather than have it be a loss: There is often some confusion as to "whose fault" the repetition is. When I am attacking, I naturally feel that my opponent shouldn't be able to undo my moves. I feel I should be able to make the attacking move again, and my opponent should have to deviate. If I could just use this rule of thumb and not worry, there would be no problem. Unfortunately, it isn't so easy to keep track of. Sometimes there are two slightly different ways to undo my move, so the defender can avoid repetition, but there is only one way for me to redo it, so I have to deviate to avoid losing, even though I'm on offense. It's a bit of a hassle to keep track of which of those two situations I'm in. If I lose by repeating, it is not a strategic error, nor even a tactical blunder, it is merely a bookkeeping error because I thought my opponent was obliged to deviate instead of me. In the game of go there is a rule that forbids you from undoing the opponent's move. The rules don't say, "If you undo the opponent's move, you lose." Here's a case where I can see the wisdom of the ancients. For Arimaa, imagine a ridiculous rule that said, "If you move a frozen piece, you lose the game instantly." It would have no effect on strategy whatsoever, but it would require people to waste their attention making triple-sure the pieces they were trying to move were indeed eligible. The fact that the Flash interface doesn't let us move frozen pieces frees us from this hassle, and allows us to concentrate on strategy instead. Therefore, if it isn't too hard, I'd love for the interface and/or gameserver to simply disallow three-fold repetition so that I don't have to think about it any more. There's enough to pay attention to in Arimaa without it! |
||||||||||||
|
Title: Re: Fairy Pieces Post by IdahoEv on Apr 12th, 2006, 7:29pm I concur with the thought of simply making repeats illegal; bookkeeping is a pain and should be minimized. Also, IIRC, the repeat rule is not just consecutive repeats, but rather takes action on the 3rd repeat of any position regardless of the time during the game. If that's so, then it could be 3-4 turns down the road and you could be completely surprised when what you think is the 2nd repeat of a position (forcing your opponent to deviate) turns out to be the 3rd repeat, and you lose! Not sure how best to do that in the interface, though. Certainly not by just eliminating the blue arrow if the position would be a repeat ... that could be quite confusing. ("Why the *&#@$ can't I move my camel there? It ain't frozen!") |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on Apr 13th, 2006, 11:18am *waves* unic here. Making repetitions illegal moves wouldn't reduce the book-keeping for games not played on the server... or is the game designed only to be played using computers, and not in real life? (I'm currently looking for suitable toy animals to make a physical set that's nicer than a chess-set and specific to Arimaa...) If neither player can deviate in a profitable manner, to me, a draw seems like the most logical outcome. (On the other hand, the situation where all rabbits are captured being a draw feels counterintuitive - wouldn't it make more sense to have the one who first loses all his rabbits lose the game?) While I'm posting anyhow - why restrict the access to the bot-developing tools and having so strict rules around bots? I've started working on a bot, but as it is now, it's unlikely to ever play on the server (unless I operate it manually) or take part in the yearly competition. Giving the game's stated goal of being difficult for bots, I would have thought developing a bot would be encouraged, not curtailed by lots of conditions and requirements? |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 13th, 2006, 1:56pm on 04/13/06 at 11:18:14, unic wrote:
Hi, unic! Welcome to the wild, wonderful, wacky world of Arimaa. Quote:
True, but making threefold repetitions into draws also wouldn't reduce the bookkeeping for live games. You would still be in the same situation as a live game of chess in which (unless the players record the moves) it's hard to know whether a position is occurring for the third time with the same player to move, especially in some endgames. Then consider the bookkeeping needed to enforce the fifty-move rule in chess. Somehow live games of chess get played anyway... To have the server do the bookkeeping for Internet games is a small convenience to make playing more pleasant. Quote:
Are you sure a draw is most logical? If so, does it bug you that the game of go has a rule forbidding you from undoing the opponent's last move? Do the rules of go seem inelegant, because they force games to continue which would otherwise be "naturally" drawn? Go as it stands is a drawless game, but without the ko rule (i.e. no undoing the opponent) there would be draws, because ko fights crop up which are worth more than the margin of victory, i.e. which neither player can concede and yet win the game. Arimaa too, as it stands, is essentially a drawless game. I like that feature. I have yet to encounter an Arimaa position (either real or composed) in which it was true that neither player could profitably deviate, but if there were such a position, I would rather have a rule which forces deviation than introduce draws. Quote:
The rule you suggest, i.e. that losing all rabbits is a loss, was in effect for the the World Championship and the Computer Championship because they were elimination tournaments. I like the rule a lot. I personally wish it were the standard rule for all games, which would make Arimaa completely drawless. However, the practical consequences are tiny. There have been more than 1300 rated human vs. human games on arimaa.com, and none have been drawn. The human vs. bot statistics suggest about 1 in 5000 serious games will be drawn. (Incidentally, 2 human games have been decided by repetition of position, games 11661 and 14990, but they don't help answer a pertinent question: Are there games in which both players are stuck in repetition, so that either player loses by deviating? Chess has such postiions and declares them drawn; Go has them and forces deviation; Arimaa might not have them!) Quote:
Way cool. If you ever get an Arimaa-specific board cobbled together, please post pictures here for the rest of us to admire. Furthermore, if you actually play Arimaa live, please modify the official rules however you like, and tell us all if your rules work out better. As Omar said just a few posts above, "If many people feel it improves the game, we could change the official rules to adopt it. I am always open to changes which could improve the game while maintaining the original objectives." Quote:
In the beginning, there were hardly any people making bots, and Omar wanted to know any time someone started to develop one. The only restriction for getting all the tools is to let him know that you are developing a bot, right? I imagine that Omar will eventually ease up on this, especially when there are other places to play Arimaa online (GGZ Gaming Zone?), and when there are so many bot efforts he can't keep track of them all anyway. In the mean time, to tell him you are developing a bot doesn't seem like a huge concession to make, given that he invented this great game, and built a great place for us to play, and hosts the site without charging us for playing, and put up his own money for the Challenge prize, and yearly puts up his own money for the Computer Championship prizes. I wouldn't have trouble listing a dozen things where I think, "Why does Omar have policy X? It would make more sense to have policy Y." Yet somehow, I spend half my life on Arimaa anyway. :-) |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on Apr 13th, 2006, 4:35pm on 04/13/06 at 13:56:17, Fritzlein wrote:
... that's why in the comment you originally alluded to, I didn't state that such positions should automatically be drawn - instead that a draw could be claimed (which is how I believe IRL chess tourneys work - you have to actually claim a draw due to repetition - it's not automatic (unlike on the chess servers)). If the repetition is non-obvious, a draw is unlikely to be claimed (and the repetition is unlikely to keep going on without becoming obvious). Quote:
Which of the 20+ different sets of rules for Go are we talking about? Go with positional superko seems inelegant, due to bookkeeping required. Go with japanese professional rules, where immediate Ko is forbidden, but some positions with multiple Ko are allowed and result in a drawn game also seem inelegant - but that's more due to how difficult it is to know when to apply which rules. Go in general does not strike me as elegant - a scaled down rules-set (Trump-Taylor is close, but not quite there, in my opinion) might make it elegant... getting rid of the "agree which stones are dead, and if not, have a complex protocol to settle who is right and who isn't without changing the score"-part of Japanese rules is certainly not elegant in my opinion. Anchor, on the other hand, is an elegant territorial game, far superior to Go, I think, much due to the elegance of its rules. Quote:
Again, there are lots of variations of the Ko rule, and some of them allow complex repetitions, and some say that the game then ends in a draw (and usually have to be replayed with whatever time the players have remaining). Go is also only a drawless game due to the Komi - surely a game-theoretically "correct" Komi would allow drawn games. I also see nothing wrong with allowing draws. What hurts a game is if draws become too common. E.g. Chess or Checkers. Quote:
If I get it all cobbled together, I'll probably upload a picture to BGG. (Part of my inspiration to want to make a Arimaa-specific set comes from the picture of a homemade Congo set with toy animals that is on BGG.) Either way - I am very much enjoying the game (even though the computer bots slaughter me in tactics so far... though I am slowly working my way up the bot ladder... beat another one for the first time today) - as can be seen by me playing quite a few games online here :) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Swynndla on Apr 13th, 2006, 4:49pm I agree that the 3rd repetition shouldn't be allowed on the servers (ie not getting a blue arrow, or maybe something more obvious). But while we're at it, why allowed the 2nd repetition? ... why not make any repetition illegal? |
||||||||||||
|
Title: Re: Fairy Pieces Post by frostlad on Apr 13th, 2006, 5:25pm The server could return that it is an illegal move due to repetition, although I know that when the server has returned illegal move notifications in the past that some bots won't recognize this and timeout. But it could be an option, that way you would know why you can't do it. |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on Apr 13th, 2006, 6:14pm on 04/13/06 at 13:56:17, Fritzlein wrote:
Replying to this in a separate post, as it's not really related to the repetition discussion. The web page says: "Since we want to know who is developing Arimaa bots and add them to the participants page, the bot interface kit is not available for direct download. However if you contact us with some information about yourself, we will send you the interface kit." What information is wanted is not specified - plus the long license agreement scares me. Most game designers of abstract games seem happy to give support for free without a long license with conditions attached - indeed, most abstract game designers are happy if you pay attention to their game in the first place. Plus, it seems unfriendly to me when people expect me to give them my email-address, buy hide their own beind a form to fill out... if I had an actual email address shown as the one to contact, I'd consider doing so. Filling in a form (which also means I don't get a copy saved in my sent-folder) is unfriendly in my eyes. Also, I would definitely rather not appear on the list of people developing programs... perhaps once I have a program I actually am happy with - but to get to that point, I would need access to the tools first. The main stumble point with the bot challenge, to me, is the requirement that the program should run on Linux. I don't use Linux, I never have, and I'm not going to learn about it for something which I am doing in my spare entertainment time. "During the month of November and December the programs that wish to qualify must be made available in the Arimaa gameroom for anyone to play against at the time control of 2/2/100/10/8 for at least 40 days." ...is also troublesome - even if I had a reasonable program by winter, I don't have a computer I can dedicate solely to Arimaa for 40 days out of two months. (I could of course run it as a low-priority process, but that would be against another one of the rules - and my cpu is often busy - I usually have one or two other (game-AI) programming projects running in the background while doing things.) So, for now, my program will just be for me to play with, and with the occasional manually operated game against the bots here, I guess... |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on Apr 13th, 2006, 7:06pm on 04/13/06 at 16:35:09, unic wrote:
Yay - make that two in a day :) Managed to beat both GnoBot2006P2 and GnoBot2005P2 today :) ... now, the question is how to a/ be able to repeat this on a regular basis - I think I mainly need to learn to avoid tactical blunders for this to happen. b/ win in a more efficient manner - both the wins were long and drawn-out grinds, with me gradually reducing the computer's number of pieces... I need to learn how to clear the way and make a goal earlier. No clue how to go about learning this though. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 13th, 2006, 7:11pm Unic, I'm sure Omar will eventually reply here to your post on the hurdles to bot development, but he doesn't read the forum every day, so there may be a small delay. I'm equally sure he wouldn't mind you sending him an e-mail directly to his Yahoo account "osyed" to discuss anything that is on your mind. When I first learned the game (and didn't yet know about the forum) he was kind enough to exchange many personal e-mails with me. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Ryan_Cable on Apr 13th, 2006, 7:31pm on 04/13/06 at 18:14:15, unic wrote:
The purpose of this rule was mostly to make sure the bots have a record on the server for the Computer Challenge defenders to study. In the Deep Blue v Kasparov match, the IBM team had a full list of games by Kasparov from his play against humans, but nearly all of Deep Blue's games were kept private. There was a long discussion of other ways to accomplish this goal, and this rule is likely to be changed for the 2007 Computer Challenge: http://arimaa.com/arimaa/forum/cgi/YaBB.cgi?board=talk;action=display;num=1131752654 |
||||||||||||
|
Title: Re: Fairy Pieces Post by omar on Apr 14th, 2006, 3:20am Thanks for the comments Ola (unic), I really appricate it. I didn't realize I might be scaring off bot developers by requiring them to contact me for the bot interface kit. At least you took the time to post about it, many others may have just left. I've changed the page now so it can be downloaded right away. Im sorry the license agreement has scared you. I guess people normally tend to get cautious when they see any kind of license. I've tried to keep the Arimaa license short and easy to understand. If what the license says scares you, that's a different story :-) But if that's the case please let me know your concerns and I'll try to explain it. You don't have to develop your bot on Linux to interface it to the gameroom, but for the computer championship I wanted all the programs to run on the same hardware and OS to keep a level playing field. So if you decide to enter in the computer championship you would have to port it to Linux. Myself or some of the other developers could probably help you with that when the time comes. Also I provide the computer championship participants with an account on a Linux system, so you don't need to worry about installing or setting up Linux. I chose to use Linux since it is easily available and is free. If you suggest another alternative I will certainly consider it. The rule for requiring the bots to be online for a fixed amount of time has become obsolete and will not be needed in the next computer championship. I think the Arimaa community came up with a much better solution to the underlying problem. |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on Apr 14th, 2006, 6:58am on 04/14/06 at 03:20:56, omar wrote:
Thanks - I've downloaded it (and already spotted a few things I need to change in the code I've written so far). I have a few more queries caused by reading the information that came with it - but I'll start a thread in the bot development forum later to ask these questions. Quote:
If it comes to that, I might take you (or some other developer) up on that - but that worry is still far into the future - for now, I first need to get a working bot :) (I'm using the MinGW compiler, and the Allegro graphics library, so as far as I know, porting should not be too much trouble.) Quote:
I glanced through the thread somebody else had linked to and it seems like a better solution has been reached :) Glad to see it. Now, I only need to finish up my bot, and get good enough at Arimaa myself that I can actually spot what errors it make ;) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 14th, 2006, 11:13am on 04/13/06 at 16:35:09, unic wrote:
Hehe. I see your point. The most elegant Go rules do allow for draws due to triple-ko, whereas the Go rules which globally forbid repeat positions are a bit awkard to enforce, and do require bookkeeping. Nevertheless, I hold up Go as a standard example where even the simplest, most elegant ko rule forbids most types of repeating positions in an effort to force games to continue rather than let them be drawn. In other words, forcing a continuation is an ancient and respectable tradition. You may find it counter-intuitive, but I expect as least as many people find it natural to forbid undoing your opponent's move, and reasonable to forbid repetition in a more global way. Quote:
Sure, it isn't necessarily a design defect for a game to allow draws on the board, and when it seems inevitable that there will be a draw on the board, it is reasonable to allow the players to agree to a draw before it happens. I think what Omar is most trying to get at with the repetition rule is banning "artificial" draws, such as short agreed draws between grandmasters in chess. Lots of chess games are agreed drawn, not because the position has no play in it, but because players are conserving their energy for future rounds. Arimaa, unlike chess, has hardly any positions which are inevitably headed for draws on the board. Indeed, I haven't seen a single game in which an agreed draw would make sense. Most games never become drawish, and in the few that do, it always seems that one player has significantly better winning chances than the other. In this context, allowing players to repeat the position and claim a draw would not have the practical effect of saving time and sparing the boredom of playing out a game that is going to draw anyway. Rather it would have the practical effect of allowing players to agree to a draw when the position isn't played out. I can see why Omar wants to avoid this possibility. It's interesting, by the way, that Americans seem to be more averse to drawn games than Europeans. I believe that in the game of Lines of Action the American inventor explicitly ruled that simultaneous connection of both players would be a win for the moving player, so as to prevent draws, but under European influence it has become standard in Lines of Action tournaments that simultaneous connection is a draw. Quote:
I'm glad you're having fun just playing Arimaa, whether or not your Arimaa-specific set and Arimaa-playing bot materialize. I have the problem that my Arimaa-related projects (such as annotating games and writing about strategy) suffer because I can't stop playing Arimaa long enough to finish them. :-) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Apr 14th, 2006, 2:34pm on 04/13/06 at 16:49:48, Swynndla wrote:
Good point, Swynndla. I think the threefold repetition probably comes from live chess games where people wanted a margin of error. A second repetition could have been accidental, but the third means you really aren't trying to vary. With the server keeping track of repetitions, there would be no need for that safety catch. There would be no punishment for accidentally creating a repeat, except that you would have to pick a different move instead. So we might as well forbid any repeats, not just the third. In light of your comments and unic's, my new favored idea is to have two slightly different sets of rules: A) For arimaa.com games, forbid any repetition of position, and have the server enforce it. B) For live games, simply forbid undoing your opponent's last move. This takes care of the majority of repetition of position. If there happens to a more complex repetition, then the players can either play on and deviate or they can accept a draw. Yes, this introduces draws, but it doesn't require any bookkeeping, so it is more suitable for live games. Also I think Omar cares more about what happens in games on his server than he cares how people play in the privacy of their own homes. :-) |
||||||||||||
|
Title: Re: Fairy Pieces Post by Swynndla on Apr 14th, 2006, 3:45pm I like rule-set A, although my first impression for rule B would be to make any repetition a case where a *win* can be claimed (but not enforced, ie it's up to the opponent of the person making the repetition to claim the win). This requires book-keeping in live games, but so does chess (where games are recorded). This would make it more dangerous playing bots live, unless it was done through an interface like arimaa.com. I say this is my first impression, and I'm not closed to the idea of making it a draw instead. It's just that for the moment I'm leaning towards the side that if a player does a move to get into a strong position, but where their opponent was able to undo that move, then since the player reached that position first, then they shouldn't be disadvantaged. But I'm still pondering about this, maybe it should be a draw. It's amazing that we can talk about the finer points of arimaa rules, where the outcome could really change the official arimaa rules ... scary! It's a good point about Americans not liking draws as much Fritzl ... it goes right back to Bobby Fischer days huh? |
||||||||||||
|
Title: Re: Fairy Pieces Post by omar on May 4th, 2006, 7:42am on 04/13/06 at 16:49:48, Swynndla wrote:
As I was reading this thread I was also thinking of proposing not being allowed to undo the opponents move or repeat the same position. The rule can be stated as: a player is not allowed to make a move which brings the board back to a previous position with the same side to move. When playing online or with a local program, the client/program should prevent such a move from being submitted. If such a move is submitted the server must reject it and notify the client that it is an illegal move and allow a different move to be submitted. When playing OTB games the players should try not make such a move. If a player makes such a move and it is noticed by the opponent then the opponent can require the player to take back the move and make a different move. Or the opponent can accept the move and allow the game to continue. The OTB games are not much of a concern right now, but in case we ever have in person tournaments the rules should be clearified. So we might as well take care of it now. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Janzert on May 4th, 2006, 9:11am I mentioned my concern to Fritzlein when the topic first came up but I guess never posted on it. So since it looks like it might be under more consideration I'll get my thoughts out there now. ;) While I'm not sure whether it makes much practical difference, I would rather make the third time a position is going to be seen illegal rather than the second. Let's say play goes from position A->B->C->D->E->F and then has the potential to loop back to A again. By disallowing the first repetition it forces the play to immediately diverge. But the player might have been in a better position to diverge at one of the positions in the middle of the loop, say at position C. But it may not be apparent at C the first time that by playing to D the loop is going to occur. Basically not allowing positions to be seen a second time forces a player to either see the loop coming, no matter how far in the future it will occur, or the loop to be broken at the first move. Allowing a position to be seen twice allows a loop to be broken at any point along it's path. Janzert |
||||||||||||
|
Title: Re: Fairy Pieces Post by unic on May 4th, 2006, 10:01am on 05/04/06 at 09:11:08, Janzert wrote:
I'd like to say that Janzert's thoughts make excellent sense, and would definitely agree with them. ... the second time a position comes up, it might not be obvious that you're heading into a repeating loop. By the third time, it should be. (Also, I still strongly think repetitions should be draws, rather than losses or illegal... seen too many bot-games where it's basically a lottery who wins/loses due to repetition (and me as a watching human usually has no clue either who'll win or lose on repetition until the server says it has happened)!) |
||||||||||||
|
Title: Re: Fairy Pieces Post by chessandgo on May 5th, 2006, 7:52pm Yep, I strongly agree with Janzert, I think it's also the point of the chess 3rd repetition rule. I would be very bad for the play to end the game at 2nd repetition ... Moreover, there might exist positions where the winning side would depend on the past of the game, even if no repetition as occured yet and with the 3rd repetition rule. Making the same position have different winner depending on such trivial thing looks very strange ... but I guess these positions almost never happen anyway ! Last thing, even if preventing abusive draw is a good idea, it might be the case that perfect play would yield to really balanced game ... forcing it somehow not be draw is not logical ... at go, the .5 komi indeed ensures no draw appart from 3 simultaneous ko, but this komi has appeared only quite "recently" in the rules ; before that "jigo" (draw) games sometimes happened ... and as said before, the appropriate komi should indeed lead to a draw :) the same wouldn't be bad for arimaa, ie if to players play really equal, draw is not a bad end .. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Swynndla on May 5th, 2006, 11:35pm on 05/05/06 at 19:52:59, chessandgo wrote:
But Omar's suggestion would mean that the game wouldn't end at the 2nd repetition, right? |
||||||||||||
|
Title: Re: Fairy Pieces Post by chessandgo on May 6th, 2006, 1:48pm ... Or to disallow it all the same ... |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on May 8th, 2006, 7:22pm Getting back to the original subject line of fairy pieces, there is one change that I am curious about, and which I think might improve the game. The biggest potential problem with Arimaa in my book, is not that it lacks strategic depth, but that it is a bit too defensive in the opening. A long time ago I proposed that, if we ever needed to loosen up the opening, we could simply subtract a horse and a dog from each side. With only six officers on each side instead of eight, and with fewer pieces matching each other in rank, the balance would tip perceptibly towards offense. An alternative to emptying out the board would be to convert one horse on each side into a lion. The pieces would then rank elephant, lion, camel, horse, dog, cat, rabbit. Then you couldn't put simply horses on b3 and g3 to block the opposing horses on b6 and g6. There would be more tensions with different unequal pieces staring at each other across the field than there are now with horses (equal value). Not only does this keep the full set of pieces, it also still plays with a standard chess set by means of inverting one rook on each side: elephant = king lion = queen camel = rook horse = inverted rook dog = bishop cat = knight rabbit = pawn That said, I have no desire to tamper with the rules of Arimaa unless extensive experience shows Arimaa openings devolving into maneuvering with no progress. As long as a determinedly defensive player can't shut down all of the opponent's offense strategies, Arimaa doesn't need to be modified. |
||||||||||||
|
Title: Re: Fairy Pieces Post by chessandgo on May 9th, 2006, 6:43pm yeah, looks like the "only forward" rabbit ability to move is a kind of entropy, and just enough not to make arimaa looks like abalone ;) There are so much strategies others than the quite conservative you reprensent with brio, Fritz, that it doesn't seem likely that defense only is a viable strategy ... ;) Have fun ! |
||||||||||||
|
Title: Re: Fairy Pieces Post by omar on May 14th, 2006, 1:31pm Discussion of the repetition rule should be moved here: http://arimaa.com/arimaa/forum/cgi/YaBB.cgi?board=talk;action=display;num=1142809371 |
||||||||||||
|
Title: Re: Fairy Pieces Post by Adanac on Jul 20th, 2006, 6:48am See what happens when professional commitments get in the way of your hobbies? I was extremely busy during April and May and I’m only reading this very interesting thread for the first time. I completely agree with IdahoEv and Fritzlein that Arimaa is a fantastic creation and the “holy grail” of strategy games. It’s truly remarkable that Omar was able to make a game that is more enjoyable than chess, with none of the glaring flaws such as pre-arranged GM draws, memorization of opening theory, etc. However, I do have one long-term worry about this game… on 05/08/06 at 19:22:20, Fritzlein wrote:
Boy, I hope someone can perfect a convincing attacking strategy that tilts the advantage towards aggressive play. Right now, I see lots of players trying aggressive opening ideas but it seems that defensive play/lone elephant attack is more “correct”. Time will tell for certain, but for now my hunch is that Arimaa is an inherently defensive game during the opening phase. Only after a few each player captures a few pieces does it become advantageous to attack aggressively. In the event that Arimaa needs a minor rule tweak to encourage more exciting openings (like others have said, I really, really hope that the rules of Arimaa never need to be fundamentally changed – it’s an amazing game as is), I would suggest a change that relates to the first article in this thread. IdahoEv suggested that the strategic depth of Arimaa would be increased if the elephants weren’t invincible. I had once believed that as well, but I came to realize that the invincibility of the elephant is essential for offensive-style Arimaa. Also, since each player is virtually guaranteed to keep his/her elephant alive until the end of the game, it makes endgames much more exciting and unpredictable than they would otherwise be if one player lost an elephant early in the game to a rodent piece, thus ending all suspense. In fact, my own suggestion for a rule change (subject to the caveat above), if it became necessary, would be to increase the strength of a camel to match that of an elephant. Or, more simply put, to give each player 2 elephants and no camel. Although it intuitively might seem that this could turn Arimaa into a stodgy, low-capture game, I think the more likely effect would be to turn it into a wide-open attacking game right from the start. Any tentative or defensive play would easily be smothered against a risk-free double-trap-swarm by an aggressive player. Thus, each player would be virtually compelled to aggressively attack at least one enemy trap in the opening phase, and the game would quickly turn into a wild affair where all 4 traps could be simultaneously contested. Some of the dynamism of the game would be lost since both the east & west wing would probably have 1 elephant, 1 horse, 1 dog, 1 cat & 4 rabbits on each side, with lots of congested positions. However, it might also turn Arimaa into a more human-friendly, bot-unfriendly, strategic-manoeuvring type of game. The most glaring flaw in the proposal might be the terrible endgames such as EERR vs eerr which could be a very dull affair, indeed. Again, I don’t want to see the rules changed, and Arimaa would definitely lose some of its strategic charm according to my suggestion. Hopefully, our worries about the domination of defensive Arimaa openings will be as baseless as the concerns of the “draw death” of Chess during the 1920s. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Jul 20th, 2006, 9:02am It's amazing how our intuitions differ. While I agree that promoting a camel into an elephant on each side would make the game more strategic and more computer-unfriendly, I am almost positive it would worsen the problem you are trying to solve. Offense in the opening would become impossible. People could feel free to attack with abandon, but it wouldn't accomplish anything if each side had two elephants to defend with. Every position would become a quagmire. As hard as it is now to overload one defensive elephant, how hard would it be to overload two defensive elephants? And a player worried about getting smothered wouldn't have to counter-attack, he could just advance some pieces one or two rows to claim space. I guess this whole discussion proves the importance of play-testing. If we can't agree at all about the effect of various rules changes, then any rule change is extremely dangerous, and reasonably likely to make a great game worse. You can tell that we collectively have no idea how rule changes will affect the game when the two of us make essentially opposite suggestions to address the same problem. If the opening is too defensive, you want to change from EMHH on each side to EEHH, whereas I want to change to ELMH. I'm trying to create fewer "ties" in strength, whereas you would introduce more! But one thing I think we both agree about is that aggression becomes more worthwhile after there have been a few exchanges. Maybe the first thing we should playtest if we want to soup up the offense in the opening is to simply start with 14 pieces on a side instead of 16. |
||||||||||||
|
Title: Re: Fairy Pieces Post by chessandgo on Jul 20th, 2006, 10:18am I don't really see the problem : lone phant attack is no defensive play, it's an attack, isn't it ? As long as arimaa is not like abalone, where one player staying in defense can hold forever it's alright : the right strategies might be different from the very opening to the rest of the game, but even if it were true, then it's all the better ! This said, I hope that one day arimaa will get as widely-played as chess, and I'm sure this day will see a lot of players playing variants, like what you propose, just for fun ... |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on Jul 20th, 2006, 11:13am on 07/20/06 at 10:18:43, chessandgo wrote:
Yes, precisely. As long as the lone-elephant attack is effective, then there is no need to change the rules. (Unless, like IdahoEv, you think that makes the game too simple.) My theoretical concern is a bit more subtle. I'm afraid that the lone-elelphant attack is only effective if there is a lurking threat of an elephant-horse attack. If we someday decide that the elephant-horse attack is always unsound, then the defender won't have to keep horses on b6 and g6. If the defender doesn't have to keep horses on b6 and g6, then it may be possible for the defender to keep everything safe, and not even allow a rabbit pull, which would render the lone-elephant attack ineffective, and make Arimaa like Abalone after all. However, at present it seems to me that some elephant-horse attacks are effective, and in particular that the defender really does need to keep horses on b6 and g6, and that in turn gives an attacking lone elephant enough targets that something has to give. So I don't expect that rule changes will be necessary, I just reserve it as a theoretical possibility. |
||||||||||||
|
Title: Re: Fairy Pieces Post by Adanac on Jul 20th, 2006, 1:19pm on 07/20/06 at 11:13:58, Fritzlein wrote:
Agreed, it's an interesting theoretical discussion for what could be changed, but I like Arimaa just the way it is now. I still have some long-term concerns about how opening theory will play out, but for now all is well in Arimaa-world and we still have lots of unexplored attacking ideas. It's somewhat counter-intuitive that aggressive attacking occurs at all in Arimaa given that there is absolutely no urgent need to advance one's army, no intermediate objectives in the central squares and the high risk involved with multi-pieces attacks. In Go, accurate play is required right from the start and Chess players immediately fight for control of d4, e4, d5 and e5 in a sharp struggle, rarely with the luxury to use even a single move on perimeter play during the opening phase. Arimaa, by contrast, seems to reward defensive play, has no central squares that need to be controlled or "fought over" and multi-piece attacks seem to be strictly optional. And yet, last year's World Championship and this year's Postal Championship strongly suggest that top-level human games very often become exciting tactical and strategic battles. It's quite remarkable, I dare say 8) |
||||||||||||
|
Title: Re: Fairy Pieces Post by LiquidTester on May 6th, 2007, 5:00pm To me the the things which move Amiraa along are 1. The invincible elephant. 2. Entropy due to rabbit pulls Is it possible to stop a lone-elephant attack without even a rabbit pull? I don't see how this is done. Obviously some attacks aiming for a rabbit pull would just put your own elephant out of position for a more threatening attack. I like the lion, upside down rook idea, just because the Camel seems "weak", this is of course only due to the invincible elephant, with a mouse, the camel would gain the ability to be used more activley. With a Lion, Camel, Horse setup, there would be a little more dynamic setup. My question is, then, why not just go 1-8 so there are very few ties, and are more freezes / hostages. Compare the game between a Elephant, 2 horses, 5 cats setup, vs. just giving a number 1-8 to the officers. I think the fewer ties available, the more congested and tactically complex the game gets (not necessarily more strategic). To reduce symetry (like always choosing 1 horse for left, 1 horse for right), but still keep some ties to keep the board open, why not a 3 cats-2 horses-1 camel-1 lion-1 elephant situation ( I think this would do better at forcing imbalances on each side and the center ), can we figure out what is best without extensive playtesting? |
||||||||||||
|
Title: Re: Fairy Pieces Post by Fritzlein on May 6th, 2007, 7:41pm Hello, LiquidTester. Welcome to Arimaa! on 05/06/07 at 17:00:12, LiquidTester wrote:
It seems to me that the invincible elephant rather slows the game down, since the two elephants can deadlock each other. If it weren't for rabbit-pulling making irreversible changes in the board, we would devolve into an Abalone situation where you could play for a draw by keeping everything safe. Quote:
Based on earlier experimentation, I believe that, if you started with all eight rabbits in the back row, and the other player attacked with only an elephant, you could keep everything safe, including from rabbit pulls. The theme of getting a stronger central attack when the opposing elephant decentralizes too much might not even be necessary. To test this it would be interesting to give Gold only an elephant, and give Silver the full box, and say that Gold wins by pulling a rabbit one step, or by capturing anything, or by pulling any piece across the mid-line. Silver wins by lasting fifty moves without Gold winning. If you would like to try it, invite me to an unrated no-time-limit postal game. I'll take Silver. (Hmmmm... this could be an interesting way for the top-rated to get schooled by a noob: I just have to make broad, untenable claims. ;-)) Luckily, the defense for Silver seems to require keeping MHHDDCC all on the second row. Advancing the horses to b6 and g6 makes the horses themselves into possible targets. So, why doesn't someone try the "home rows" defense? Because it seems to invite a successful elephant-horse attack. When the defending elephant is preoccupied with the attacking horse, the attacker is able to get some kind of advantage. Quote:
Thank you! Quote:
One of Omar's design requirements was (is) that Arimaa be playable with a standard chess set. The trap squares stretch this rule a little bit, and an inverted rook would stretch it further. Having eight distinct ranks above rabbits would definitely make Arimaa more dynamic, but playing with a standard chess set would then be impossible. Quote:
Ah, there is the trouble. A bad game can be busted with a short amount of playtesting, but a fairly good game could conceal its weaknesses for a long, long time. We have collectively played over 50,000 games of Arimaa, and we still don't know whether lone-elephant openings are mandatory, or whether some sort of elephant-horse opening will trump it and become standard. There's even an outside chance of it being correct to give the camel an attacking role, and/or it being correct to make a swarming attack. On the other extreme, there's an outside chance that purely defensive play is effective, and an attacker can't make progress without using moves that lead to disadvantage. It will be a shame if Arimaa peters out after 500,000 plays because we understand by then that offense is futile. Think of all the time we will have invested to kill our hobby! But the good news is that, even if Arimaa isn't an excellent game, it is so close to being excellent that a very small change could tip it in the right direction. The simplest way to make Arimaa dynamic without invalidating anything we have learned: remove one horse from each side, playing 15 vs. 15. Recent experience strongly suggests that no changes are necessary to ensure dynamism. If you look at the great variety and complexity of top-level games these days, it seems there is no shortage of excitement in correct play. On the contrary, my current suspicion is that dual-lone-elephant openings would never have had an hour of supremacy if it hadn't been for anti-bot play. Given that most games are man vs. machine, and a lone-elephant attack is safe and effective against every bot, then we can understand why lone-elephant play is popular, but strategic understanding didn't have to evolve that way. It is possible that, if there had been only human vs. human games from the start, the top players would all have been multi-piece attackers from day one. |
||||||||||||
|
Title: Re: Fairy Pieces Post by NIC1138 on May 6th, 2007, 8:07pm on 05/06/07 at 17:00:12, LiquidTester wrote:
|
||||||||||||
|
Title: Re: Fairy Pieces Post by aaaa on May 6th, 2007, 9:28pm on 05/06/07 at 17:00:12, LiquidTester wrote:
I don't know, can you place your knights and bishops upside down as well? Perhaps we should lobby chess piece producers for this feature. ::) |
||||||||||||
|
Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1! YaBB © 2000-2003. All Rights Reserved. |