|
||||||||||||||||||
|
Title: My First Impressions of Arimaa Post by Meneldor on Dec 2nd, 2009, 10:41pm I am new here, I was a former chess player and I was researching abstract board games on boardgamegeek and came across Arimaa, which apparently was invented in order to make it tougher to solve by Computer A.I. I am attracted to it as I wish to teach my son a board game with a chess set that is not as complicated as chess. So far I have only been successful at getting him to play a few games of checkers. Nevertheless, I was curious enough to take a more in depth look at it by reading the Arimaa wikibook on tactics and strategy. I have not had time to play a game of Arimaa, but here are my first impressions, and opinions. Game Theme: most abstract games are modeled after armies, this one uses animals: we have an elephant, camel, horse, dog, cat, and rabbit. So what strikes me is that these were probably chosen to make it more appealing to children rather than to be modeled after the game itself. But dogs and cats are predators: thus they should capture a piece by eating it. So I am wondering why predatory animals were chosen, when in fact all they do is push and pull. And the rabbits. Why rabbits? To me, rabbits should be soccer balls, the objective is to score them in the opponents goal. Thus Arimaa would be better themed as a soccer-football type game, with bigger players pushing at smaller players. And when I say soccer-football, I mean it: its a combination of European soccer with American football, with a line of scrimmage for the soccer ball. Game Balance: all pieces move the same, size is the only difference, thus short term tactical play in Arimaa is limited, similar to checkers. But one thing that bothers me is that the Elephant is invincible: it will never die unless it commits suicide down a trap door. I would have liked to have seen a piece that can counter-attack an Elephant. So it seems that the only strategy to nullify an Elephant is to pin it down in a defense, or blockade it with a swarm. In the wiki, there is an example of pinning down a piece by fencing in another. But the question this brought up for me was that if I can push and pull an opponent's piece, why cant I push or pull my own team piece? Perhaps this was done for game balance for blockades, but if not, its somewhat of a logical inconsistency in the rules. In the short history of Arimaa, one of the first strategies discovered was to take a camel hostage, and attack with a combined elephant and horse. For me, this reveals somewhat of a weakness in balance: besides rabbits, there is TWO of every animal, except for the elephant and camel. So, why not have TWO camels? Then a hostage camel would not introduce such a rapid game imbalance. If there are two camels, we would then need a 9x9 board, instead of the 8x8 chess board. And thus 9 rabbits (or better, 9 soccer balls). So, it seems that the 8x8 board was chosen to make this game readily accessible for those who have chess boards. A 9x9 Arimaa board will do something else: it will make it more plausible, perhaps, to advance rabbits up the center. However, it will also make it harder to shift pieces from one wing to the other. The other problem I see with game balance is that silver can set up his pieces in response to gold. What this may lead to is that when this game finally gets solved (which will take a while), it may lead to a second move advantage. To me this is a bit artificial, and was done for the sake of defeating Computer A.I. Chess960 seems more balanced as each player must mirror each other. The initial advantage this gains over computer A.I. is probably temporary. After numerous plays some Arimaa players are discovering advantages to certain starting positions, as this game evolves, it may turn out that there will be a standard starting position. So, those are my first impressions, from someone who is familiar with chess, and also inventing games and testing them out in order to improve game balance. Since I have zero experience playing Arimaa, I would be interested in hearing thoughts from experienced Arimaa players. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by RonWeasley on Dec 3rd, 2009, 5:42am Welcome Meneldor! I hope you enjoy playing arimaa. Many of your questions are about variant rules from the current set we are using. There is a discussion thread about variants elsewhere in the forum, to which I would provide a link if I knew the spell. Many of your suggestions would make it a different game playing similarly. Pushing and pulling one's own animals would make things very different, I think, making some blocking tactics less effective and prolonging some types of games. As to the pieces themselves, it might just be a matter of personal taste. In the abstract, the pieces are numbers since only their relative rank is significant. In my case, animals are much more convenient since Hagrid has a ready supply of the animals we need and there are very few wizards who can produce and control numbers well enough to keep them on a game board while the rest of the world wants them for clocks and addresses and such. Rabbits are especially good for arimaa since they are plentiful and they can't hop backwards because of the shape of their hind legs. I don't think the animals like being pushed and pulled, so it's just as well this is not done by animals on the same side, which might lead to animals just leaving the game on their own. Or, in choosing animals, maybe Omar didn't want a strong connection between arimaa and some real world artifact that might produce an unwanted side effect on the game. I think after playing for a while, many of your questions will answer themselves and you would have a good baseline from which to investigate the variants you propose. For many of us, it has been a fun experience. I hope it's fun for you too. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Fritzlein on Dec 3rd, 2009, 8:03am on 12/02/09 at 22:41:27, Meneldor wrote:
Thanks for sharing your impressions, Meneldor. It happens that Omar's desire to teach chess to his son Aamir was an integral part of the invention of Arimaa. At first the rules of Arimaa were even simpler than they are now, for Aamir's sake. But then in playtesting Omar discovered that he needed to add some slight rule complexity (dislodging/capturing/freezing) to make the strategy interesting. I am such an abstract guy, I strip away any theme immediately. It doesn't feel any different to me to play with chess pieces than with animals. I expect playing with pyramids of different heights wouldn't change the experience for me either. I take your point that some rules don't fit with the theme (e.g. carnivores pushing instead of eating), but soccer balls moving forward and changing direction of their own accord would likewise be anti-thematic. To me that just proves that Arimaa isn't an animal fight or a race or a soccer/football match or anything at all besides Arimaa. I know that for some people a good theme can greatly increase their enjoyment of a game. It is a matter of taste, so your opinion about what would be a good theme is authoritative for you. Maybe when the Arimaa craze is sweeping the nation, Z-Man will produce sets with different themes to satisfy a variety of tastes. Inevitably there will be a pewter Civil War set with generals as elephants, etc. :) Quote:
You are not the only casual observer to be bothered by the invincibility of the elephant. I'm not sure why that is intuitively troubling to people who read the rules. Perhaps it is simply too different from chess, where even a lowly pawn can take a queen? The relevant question for me is why the invulnerability of the elephant is a problem. What goes wrong with game because the elephant can't be captured? I think you will find from playing the game yourself, or from asking experienced players, that nothing goes wrong with the game. In particular, Arimaa doesn't bog down to a stalemate where neither player can make progress. It's odd how much actual game play can diverge from intuitions based just on the rules. From the rules of chess, I would expect a brutal slugfest because capturing is so simple. Any piece can take any other piece. The last thing I would be worried about from the chess rules is draws. Yet it turns out that over half of grandmaster chess games are drawn. Meanwhile Arimaa, which might seem drawish because elephants are un-capturable, turns out to be extremely decisive. Go figure. Quote:
Hmmm... when you play chess, does it seem logically inconsistent to you that you can only promote a pawn to a piece of your own color? Wouldn't the rules of chess be more logically pure if you could promote a pawn to any piece, yours or the opponent's? There are chess puzzles based on that rule inconsistency, if you want to call it that. Maybe it is just what I am used to, but I don't see any inconsistency in chess that you can only promote a pawn to your own color of piece, and I don't see any inconsistency in Arimaa that you can only dislodge the opposing color of pieces. For Arimaa the rule about not being able to push your own pieces only makes a difference for pushing your own rabbits backwards, because any other piece that you could push can also move there on its own volition. The inability to push one's own rabbits backwards may seem like a quirk of the rules or not, depending on what you are expecting, but it turns out to have a nice strategic side-effect of enabling some elephant blockades. It's a non-trivial difference; a 2007 World Championship game between PMertens and Adanac was decided because of it. Furthermore, the inability of rabbits to move backwards is the bedrock of much of Arimaa strategy in general. If my elephant could pull my rabbits backward after the opponent pulled them forward, it would be tough for him to give me a permanent strategic disadvantage. Omar was at first opposed to non-retreating rabbits because he didn't want to water down the purity of the rules, but eventually reconciled himself to the rule because it made the game play better. I love pure and simple rules. The easier the rules of the game are to state, the better I like it. For that reason Go will probably never be dethroned as the best abstract game of all time. However, ultimately the way the game plays trumps everything else. In my estimation, allowing a player to push/pull his own pieces would make Arimaa far less strategic, because forward rabbits wouldn't be stuck forward. Furthermore, it wouldn't really make the rules simpler to change it. Either way you would have to insert a clarifying sentence that you can dislodge your own pieces or can't dislodge your own pieces, because without the clarification some people will wonder. Given that the simplicity is a wash, I'm all for the version that makes for deeper strategy. Quote:
It would be interesting to test having two camels. I'm not sure how the game would change. However, I don't see how it is a problem in the current rules that taking a camel hostage gives the hostage taker an advantage. Is your objective in creating rules to make sure neither player can get an advantage? :P Quote:
When Arimaa was just a computer science experiment, Omar wanted Arimaa to be playable on a chess board. Now that Omar is trying to sell Arimaa sets, maybe he wishes he had made the board 9x9 instead. Then people wouldn't play Arimaa on their chess sets and pass on buying an Arimaa set! ::) Quote:
If the game gets solved, there is a problem. A game is equally busted from being a first-player win as it is from being a second-player win. Also a game is busted if it is solved to a draw. Folks seem to have strong intuitions about what they want the solution of a game to be, but to me it makes no difference whether it is 1-0, 0-1 or 0.5-0.5. What matters to me is that the game is so complex that nobody can solve it, and in the mean time the players have equal chances and few draws. It would be a problem for me if Silver setting up second meant that Silver was winning 60% of the games, but in fact Silver is winning 50% (rounded to the nearest %), and there are 0% draws (exactly). Quote:
On the contrary, Chess960 clearly favors white. The setups are mirrored, but white gets the first move. In some piece configurations the advantage is tiny, while in other piece configurations the advantage is significant, but in no case is it even. You picked a bad example for comparison, because experience shows that Arimaa is far more balanced than Chess960. Quote:
Yes, as we get more experienced with Arimaa, the openings may standardize. Only time will tell. For the present, however, the preferred openings are still very much in flux. For example, my current favorite setup as Silver was never played by anyone in any game before 2009. In six years of experimentation, it never occurred to us before. Our learning has a long way to run before any opening theory gets settled, if it ever does. I have two competing thoughts about standardized Arimaa openings. On the one hand we should expect that the more we know and understand about Arimaa, the more obvious it will be what moves are correct, which will lead to standardization. In this light, the lack of Arimaa opening theory is temporary and due to our ignorance rather than due to anything inherent about Arimaa. On the other hand, I note that out of only 20 possible opening moves for chess, the best players in the world can't agree which move is best between the top two, d4 and e4. For Arimaa we have 64,864,800 possible opening moves. What kind of amazing understanding would it take to get agreement on the best thousand setups, never mind the best two? The range of options is so broad that if I learn anything new about strategy it will probably change the way I play starting from the very beginning, which means there will never be standardized Arimaa openings. Both of these intuitions are powerful in my mind. I think we will just have to wait and see which turns out to be more in line with reality. But one thing I am quite convinced of is that even if Arimaa is played for centuries like chess has been, the opening theory of Arimaa will never be mapped out as deeply as chess opening theory has. Compared to chess, more moves of each game of Arimaa will always be fresh. Quote:
The only thing I know about inventing games is that the testing trumps the intuition every time. Everyone seems to think they can tell how a game will play just from the rules but they can't. Even Christian Freeling, the expert abstract game designer whose intuitions on this score may be better than anyone else's in the world, can't tell how a game will play just from the rules. He tried to prove that he could in the Arimaa Forum (of all places!), and to my mind convincingly proved the opposite. He invented a new game, couldn't tell from the bare rules that it would be too drawish, and had to fundamentally change the rules when playtesting revealed the potential problem of drawishness. I have never invented a game of my own, but from having seen others try their hand at inventing, I advise you to do three things: playtest, playtest, and playtest. :) |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Arimabuff on Dec 3rd, 2009, 8:36am on 12/03/09 at 08:03:31, Fritzlein wrote:
And here I thought it was: Location, location, location. ??? ;D |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Hirocon on Dec 3rd, 2009, 3:01pm Regarding arimaa with two camels: I've actually wondered the opposite. What would arimaa be like if none of the pieces except the "pawns" had duplicates? For example, you could play arimaa with 1 elephant, 1 camel, 1 horse, 1 boar, 1 dog, 1 cat, 1 rabbit, 1 mouse, and 8 crickets. All pieces can move backward except crickets, and the goal is to get a cricket to the other side. This variant would make the number of opening moves even larger. Of course, there is nothing wrong with arimaa as-is, but I've just wondering how this variant would change things. I'd like to test this variant some time. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Fritzlein on Dec 3rd, 2009, 3:50pm on 12/03/09 at 15:01:59, Hirocon wrote:
My guess is that it would make it somewhat less defensive and more fluid. If Arimaa ever seems to be heading towards drawish positions, this would be the first thing I would want to try to make it more decisive. Fortunately, there is no need so far! :) |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Meneldor on Dec 3rd, 2009, 4:32pm Wow lots of replies. Somehow one of my posts here got lost. Here it is again, so sorry if my post appears twice: on 12/03/09 at 08:03:31, Fritzlein wrote:
I used to just play chess as it was a good abstract game. But lately I have studied some ancient board games, and symbolism and theme have thus become more important for me. I meant to update my post to not include the soccer reference. The more I thought about it, Arimaa is a combination of football and Sumo wrestling. A more descriptive name of it would be "Sumo-Chess", as it has a unique way of movement using push and pull. Quote:
I would like to directly attack the elephant. Not so important, as you can pin it down or blockade it with a swarm. Quote:
The fact that Arimaa is not drawish, like checkers, or like chess, is one of the main features that attracts me to it. Arimaa removed many draws by simple rule changes, some of these rule changes can be applied to chess or checkers (e.g., some chess variants regard a stalemate as a win). Quote:
In the earlier versions of chess, there was no Queen, instead it was a general, and pawns would get promoted to a general. In the original game, the theme was consistent. Quote:
If there were two camels, it will make it easier to launch an attack with a camel. For me, it will not change gameplay much, but will simply remove an "imbalance" in the game where a hostage camel can tilt the entire game. I would test out this variant first on a 9x8 board, then perhaps a 9x9 board. Having a wider board will also make it harder to shift between left and right flanks. Keeping the depth at 8 means the rabbit journey will not increase in length. Quote:
Any abstract game, a game where all information is perfectly known, is ultimately solvable. This includes the game of Go. Like Go, Arimaa is hard to solve. But ultimately when solved, every abstract game played perfectly will always result in a draw, a first player win, or a second player win. Chinook plays Checkers perfectly so that it will always end in a draw, if not a win for Chinook. So its important that a game not be humanly solvable (which is what I think you meant). Quote:
Christian Freeling? Can you point out the thread for that one? Quote:
This is very true. At this point, my children can play checkers and Parcheesi. But I immediately noticed how Parcheesi has some ridiculous rules that makes it imbalanced. So, I decided to invent my own rules to the game, and now its become an enjoyable strategic game (at least more so), and one of my children wants to play it all the time now. Patterns get revealed in playtesting, and that how game imbalances are discovered. I think I would like to see how a two camel variant of Arimaa would play. The 9x8 (or 9x9) board would also make it more feasible to advance rabbits up the center. [b][/b] |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Fritzlein on Dec 3rd, 2009, 5:53pm on 12/03/09 at 16:32:11, Meneldor wrote:
http://arimaa.com/arimaa/forum/cgi/YaBB.cgi?board=other;action=display;num=1236541162 |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by camelback on Dec 4th, 2009, 12:52pm 4 trap squares on a 8x8 board has a symmetry. All trap squares are exactly 2 squares apart and from the edge. Any other size would lose this symmetry. There was a player IdahoEv, who play tested with different sizes. I believe his conclusion was "capturing becomes extremely difficult" with a different size board. I agree since lot of existing tactics will change when the symmetry is lost. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Meneldor on Dec 4th, 2009, 11:27pm on 12/04/09 at 12:52:49, camelback wrote:
I tried a few games on the bot ladder, and more than symmetry you have to account for the fact that each player has four moves. Widening the board will make it harder to defend and counterattack. So the two camels is not feasible, unless a rabbit is killed off. But having a range of animals that are unique, without duplication, that would offer an interesting variant. I honestly like the gameplay. But the choice of animals is inconsistent, and its going to be an immediate turn-off to new players. maybe have all of them be predatory animals except the rabbit, or at least animals with horns that can attack. To make matters worse, its designed to be played with a chess set, but oddly enough the recommended configuration from this web site is that the rook represent the horse, and the knight/horse represent the cat. This leads to IMMEDIATE confusion when introducing the game. Everyone here is probably an abstract game player that does not care, but representation and theme of a game is important. I am still trying to come up with a list of animals that would be better theme-wise. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by chimaera on Dec 4th, 2009, 11:55pm on 12/03/09 at 08:03:31, Fritzlein wrote:
This isn't quite so simple as just allowing rabbits to move backward. Check this out: Code:
Ignore the stupid layout. Just pay attention to Hd6 and Re6. If the rules were modified so that you could push/pull your own pieces, gold could win on this turn. Gold can pull his rabbit west with his horse: 20g Hd6s Re6w Rd6n Rd7n |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Meneldor on Dec 5th, 2009, 9:53am on 12/04/09 at 23:55:25, chimaera wrote:
Its at first admittedly non-intuitive. But if that rule was allowed, pinned pieces next to trap doors would be able to move without losing a piece on the trap door, and it will be easier to unfreeze frozen pieces. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Hirocon on Dec 5th, 2009, 12:10pm on 12/05/09 at 09:53:51, Meneldor wrote:
The rules in general are such that, when a piece is on a trap without a friendly adjacent piece, it is captured immediately. Even if you allowed pulling of friendly pieces, you couldn't pull a friendly piece off of a trap, because the piece on the trap would be captured immediately after the first step of the pull, making the second step of the pull impossible. Still, the example above shows that allowing pulling and pushing of friendly pieces would have another effect on the game besides providing a way to move rabbits backwards. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Hirocon on Dec 5th, 2009, 12:14pm This would create a paradox: what happens if a piece on a trap tries to push an adjacent friendly rabbit backwards? The pusher would be captured after the first step, making the second step impossible. But because it is a push, the second step is required. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by chimaera on Dec 5th, 2009, 12:30pm on 12/05/09 at 12:14:23, Hirocon wrote:
A pull can be completed even if the puller disappears into a trap on the first step of the pull. If pushing your own piece were allowed even if the pusher would immediately disappear into the trap, then I imagine the push would consist of one step which counted as two, and you would lose the pusher. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by Hirocon on Dec 5th, 2009, 5:55pm What if a piece not on a trap tries to push a rabbit backward, and the rabbit is the only thing unfreezing the pusher? |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by omar on Dec 6th, 2009, 11:02pm on 12/03/09 at 05:42:20, RonWeasley wrote:
Another classic; I was laughing so hard my stomach is hurting :-) Some day I'm going to have to collect all your classic responses and make a 'RonWeasley Arimaa FAQ' page. |
||||||||||||||||||
|
Title: Re: My First Impressions of Arimaa Post by omar on Dec 6th, 2009, 11:05pm on 12/05/09 at 12:14:23, Hirocon wrote:
Wow, that would look like a pretty strange move :-) |
||||||||||||||||||
|
Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1! YaBB © 2000-2003. All Rights Reserved. |