Arimaa Forum - Print Page


    
      
        Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> General Discussion >> Choice of server, amongst others...
        
(Message started by: Farkov on Oct 13^th, 2004, 2:03am)

Title: Choice of server, amongst others...
Post by Farkov on Oct 13^th, 2004, 2:03am

Hi all, this is my first time at the message board, which may be a bit premature since I've only played two games of Armiaas... Firstly I'd like to congratulate the developers of this game. I think it's a fine idea to have a new game devised and built around deloping truly interesting AI. My experience lies in programming chess bots that play all the different wild variations, so it was nice to find another 'variant' to mess with ;)

Now that the niceities are out of the way I have a few of questions that I find truly puzzling. The main one is about the server. Is there a reason the server is in html form? This must be the one of the most annoying forms for any experienced computer user, and it must make connecting bots a real headache as well. I'm also yet to find a way to ask a simple question to someone in real time. Are there any plans to change the server into something one can telnet into and write decent interfaces for...like ICC and FICS? The advantages for noobs and experienced users alike are too many to list.

Secondly...who thought of the move notation? :o I might be missing something but isn't it possible to use standard algebraic notation like chess? a8b8, b8c8, c8c7, d7d6. That's at least as short as the current notation and a lot more readable.

The seemingly over-complicated time controls are last on my gripe list...I appreciate the need for bots to have time to think in a game where 2 ply is not exactly trivial, but surely a simple starting_time/increment system would have been enough?

Please don't think I'm dumping on this site...I see a lot of people have put in a lot of effort and it's great to see. I just feel that in a couple of areas the efforts are being wasted. These are just opinions I have as an experienced bot operator, and having spent more time than I should have playing chess online ;D

I look forward to your replies (and the inevitable correction of my mistakes) :)

Title: Re: Choice of server, amongst others...
Post by Fritzlein on Oct 13^th, 2004, 3:36pm

It's never too early to join the message board. I regret playing so long before I bothered to read the interesting discussions and post to the message board myself.

Not having programmed a bot, I can't respond to the point about the server interface, but I have an opinion on your other two points.

One feature of the time control I love is having some bank of time in reserve, but not too much. This keeps the game moving at a nearly constant pace. For example, the EU-US games are a minute per move, plus a bank of 3 minutes. You don't have to make every move in exactly a minute, but then again you can't save up and suddenly stop and think for twenty minutes. Furthermore, if your bank of time is full, you have no incentive not to think for the full time allotted, which discourages people from being overly hasty and blundering.

I think it is a feature, not a bug, to take away some of the control over time management that chess players have traditionally had. I was following the Kasparov-Karpov championship match in 1987 when Kasparov thought for (iirc) 83 minutes on one move. That sort of thing kills chess as a spectator sport, and probably isn't too fun for the opponent either. Conversely, Karpov more than once managed his time badly and blundered in time pressure around move 35. I'm glad for a time control which doesn't allow players to shoot themselves in the foot in this way, and insures they will have at least a minimum amount of time per move.

As for the notation, I also object to bringing in the compass directions. I tend to be a minimalist where these things are concerned, and if you don't need more than the coordinates of the squares, I don't see why you should introduce them. I personally would prefer e2e3 type notation to the current system.

On the other hand, chess notation isn't purely cartesian either. I'm not sure why this is. Perhaps people (non-mathematicians?) have found it easier to replay the games with the system that has actually become standard than it is to replay the games with coordinates alone.

One of the great things about running your own server is that you can do things however it makes the most sense to you. If it were my server, I would force everyone to use coordinate notation only. And believe it or not, there are still people who use only descriptive English notation for chess. In the future, when Arimaa has become widely popular (and I expect it will surpass Twixt, Trax, etc. as well as 99% of chess variations), there will probably multiple servers with different ways of doing things, and then popular opinion will decide.

Title: Re: Choice of server, amongst others...
Post by omar on Oct 14^th, 2004, 3:08pm

Wow three deep questions in one post. Well this reply might get rather long so bear with me.

1 Why is the server in HTML form?

I guess what you really mean is why is the gameroom interface an HTML page. Well I don't plan on keeping it that way forever, eventually I want to redo the gameroom in Flash. I did it in HTML so that it would be quick and flexible since I didn't know what features I wanted to have in the gameroom. The backend server is just CGI scripts and a database. Writting a full blown backend server which you can telnet into would have been very time consuming and would right away require the gameroom interface to be done in Java or Flash, since telnets are not possible with just plain HTML. Or maybe some other interface which the users would have to download and run seperately from a browser window.

In the long run something like the FICS server with a flash or java based front end would be the right way to go. At one point I tried to look for the FICS server code, but did not find it. Do you know if it's still available.

2. Why is the notation not standard algebraic like chess?

I did consider using standard algebraic notation, but I chose this notation because I thought it conveyed more information to humans and still used the same number of characters. With the current notation we can easily tell which piece moved and in what direction. For computers it's also good because the current notation can easily be converted to algebraic notation.

One nice feature about this notation over algebraic notation is that it can easily be played in reverse direction to undo the steps. With this notation when a piece is trapped we can tell which piece was removed from the board. Otherwise a program would have to keep track of that seperately to allow going backwards throught the steps.

3 Why are the time controls so complicated?

It won't seem so complicated once you get used to it :-)

Actually the Arimaa time controls are very similar to the start_time/increment system. The start_time is like the starting reserve time and the increment is like the time per move with the main difference being that the time is given before the start of the move rather than after.

Aside from this the Arimaa time control provides a few more optional parameters. One to specify what percent of the unsed move time is added to the reserve. Another to limit how much total time one can accumulate in reserve and another for stopping the game after a fixed amount of time. Though these are optional parameters, we usually tend to use them in tournament games because they help control the duration of the game. But if the duration of the game is not a factory the time controls can be specified with just two parameters.

Omar

Title: Re: Choice of server, amongst others...
Post by MrBrain on Oct 15^th, 2004, 10:22am

I too do not like Arimaa time controls. It forces the players to play at a specific rate, rather than giving them the flexibility to use time where it is most needed. This decreases game quality. Sure, it may make it more fun to watch (I would disagree with this), but what's more important? The game itself or the spectators. I think the answer should be clear. Quality of the game -- the ability of the players playing the best game possible -- must take precidence.

But I've found that for me, another big detriment about the way the time controls are set up is -- it gives one no chance to use the facilities without losing a significant portion of reserve time, which in Arimaa time controls is a very important thing. It's no fun to try to "hold it in" when you're running out of reserve time. :-/

Title: Re: Choice of server, amongst others...
Post by CryingClown on Apr 14^th, 2010, 3:02pm

Yeah I too would like to play at a timecontrol I prefer. And when I click someone's name it opens it in another big window. Change needed.

Title: Re: Choice of server, amongst others...
Post by ocmiente on Apr 15^th, 2010, 10:05am

Writing as someone who never played chess except for casually, or against a computer, I like the time controls, and the notation - though I'd be OK with the algebraic notation too.

Regarding time controls, I was thinking about chess clocks the other day. Don't know much about them, and I was wondering, are there chess clocks that support the Arimaa time system, or if they all do?

I'm pleased that Fritzlein thinks the algebraic notation would be better. I love he 'Beginning Arimaa' book he wrote, and have read through parts of it multipe times, but the notation in that book was more difficult, in my opinion, than either the algebraic notation or the current Arimaa notation used on the server. Any chance for a second, or special, edition using a different notation?

Title: Re: Choice of server, amongst others...
Post by Adanac on Apr 15^th, 2010, 10:42am

on 04/15/10 at 10:05:02, ocmiente wrote:

I'm pleased that Fritzlein thinks the algebraic notation would be better. I love he 'Beginning Arimaa' book he wrote, and have read through parts of it multipe times, but the notation in that book was more difficult, in my opinion, than either the algebraic notation or the current Arimaa notation used on the server. Any chance for a second, or special, edition using a different notation?

I've always thought that symbols for pieces & directions would be the most intuitive but that's obviously not too practical in most cases. For example, a gold elephant pulling a silver camel would be very intuitive with:

{picture of gold elephant} {picture of silver camel} {left arrow} {down arrow}

I believe that is more quickly understood than:

Eh3w mh4s Eg3s mh3w
or
Eh3ws mh4sw
or
EM {left arrow} {down arrow}
or
EM<V

(though it doesn't appeal to me aethetically, I have to admit that the last example is fairly intuitive and doesn't require any special symbols)

If symbols are unavailable I personally prefer the algebraic notation Eh3w mh4s Eg3s mh3w or Eh3ws mh4sw but that might just be due to my chess background. Non-chess players might prefer the latter option.

Title: Re: Choice of server, amongst others...
Post by ocmiente on Apr 15^th, 2010, 11:01am

I think using the coordinates, rather than a picture or character for the piece moved is easier because using the picture or character requires extra work to translate the notation into what happened.

So, if the notation says E->, or something like that, OK... that's simple enough, but I then have to go through the extra mental step of finding the elephant on the board, and if I'm looking at the board from Silver's perspective, reverse the direction. a5b5, on the other hand, takes no extra thought. It's even more work when you have a move like D^. I move a dog... which dog... it must be the only dog that can move north... etc. Too much to think about... making the notation a little more verbose helps a great deal in making it easy and quick to translate the move into action, for me.

I guess what I'm trying to say is that I prefer a notation that is easier to understand over one that is concise.

Title: Re: Choice of server, amongst others...
Post by Fritzlein on Apr 15^th, 2010, 12:10pm

on 04/15/10 at 10:05:02, ocmiente wrote:

Any chance for a second, or special, edition using a different notation?

I'll consider a second edition when the first edition sells 1000 copies. In other words, no. :P

Nevertheless, given another chance, I wouldn't use the same notation. Nor would I use the game room notation, nor straight algebraic notation, because they are too verbose. Yes, pure algebraic notation is easy to understand, perhaps easier than anything else that has been suggested, but that doesn't make it easy to read. In particular there is nothing to indicate the logical groupings of pushes and pulls, and no way to indicate that a piece has moved more than once. EDv> does require you to locate the elephant, but only once, after which locating the dog is easy, but the more verbose notations require you to determine four different times which piece is moving.

If I wrote another book, I would keep experimenting. Furthermore, I would encourage other people to try using their pet ideas in chat and in the forum even if they don't write books. It is easy to argue abstractly about the merits of the notation you prefer, but it is only when you try to use it that its flaws become apparent. I have learned that the hard way. ::) If you want your notation to be adopted, using it yourself will do more to that end than trying to convince everyone else to use it.

Title: Re: Choice of server, amongst others...
Post by omar on Apr 17^th, 2010, 1:50pm

on 04/15/10 at 12:10:47, Fritzlein wrote:

If I wrote another book, I would keep experimenting. Furthermore, I would encourage other people to try using their pet ideas in chat and in the forum even if they don't write books. It is easy to argue abstractly about the merits of the notation you prefer, but it is only when you try to use it that its flaws become apparent. I have learned that the hard way. ::) If you want your notation to be adopted, using it yourself will do more to that end than trying to convince everyone else to use it.

There is a forum thread where we discussed using a shorter notation. Might have been about a year ago. During that time I tried recording games using various different notations and timing how long it took. We need to do more experiments like that and share our results.