Omnimaga
Calculator Community => Other Calc-Related Projects and Ideas => TI Z80 => Topic started by: Zera on March 26, 2010, 05:39:26 pm
-
I was thinking about an off-topic thread I started a short while back (http://ourl.ca/4107) about A.I. cheating in computerized board games. The game in question was Hnefatafl. A lot of people showed interest in playing the game; although there really aren't many (maybe one or two) computer renditions of it available. Actual boards are also hard to find in this day and age, given the historic context of the game.
I was curious if anyone might be interested in coding a fairly simple version for calulators using BASIC. (or possibly using xLib, Celtic, or even Axe) Since I know quite a lot about the game, I could put together the necessary assets, along with detailed descriptions of the rules and all available moves. As a basis for comparisson of coding difficulty: The game is not strategically as complex as Chess, but the logic in how the rules and moves would be coded could be a bit more difficult. Hnefatafl pieces have very basic movements, but there are so many different ways to capture a piece. This very often depends on the arrangement of several pieces relative to sections of the board itself. (with certain tiles of the board having special properties of their own) Secondly, there are two different objectives of play, depending on which side is played: One side is tasked with simply escaping the board, while the other is tasked with preventing escape, or capturing / cornering all opposing pieces. It's very different from games like Checkers or Chess, which provide you with similar pieces and equal forces. In Hnefatafl, the defending army is always outnumbered. Regardless of which objective is played, completely wiping out the opposing side will still result in victory either way.
I could also propose some strategies that the A.I. would use. Defining a flexible computer logic for this might not be possible. It would probably be necessary to define very deterministic strategies and moves the A.I. would deploy when it examines the placement of opponent pieces on the board. Randomized moves most certainly would not work. The A.I. would just be crushed.
It would be necessary to provide a text file with the game detailing these rules. I think trying to cram all the necessary text assets into the actual game would kill a lot of memory. I could draw-up a suitable document, along with images to show examples of movements and captures.
One other thing: Would it be possible to use any grays? This would make the board and piece designs vastly easier to draw. The board is not your standard checkered ordeal. I have to make several tiles look unique, as well as distinguishing both forces, and the king's piece. An 11x11 board is most certainly going to require that each tile be 4x4 px. (unless uncommon resolutions are allowed; like 5x5, 6x6, etc. - that would definitely help) In monochrome, there won't be eough leniency to distinguish everything.
Working concept: (still in progress)
(http://img408.imageshack.us/img408/8261/image1y.png)
-
unfortunately, it is impossible to use grays as background in BASIC. Notice how in your Tromino concept I got rid of all grays behind sprites and scores. That's how limited BASIC and libs is. Sprite masking is just too slow to support decent grayscale , not to mention hard to work with with gray background x.x. The only possible gray is either:
1) if only two sprite maximum that can move around has grayscale (or just one if it's 4 level gs)
2) if only the screen areas that never gets updated has grayscale
It looks interesting, though
-
Can't you make it monochromatic, and make 1 player white with black outline, and the other filled-in black?
-
Ya, that is kinda what I was thinking. I don't see a real need for grayscale. Just sayin'.
-
Oh, and may be I would like to try to code this. I would need to learn how to play first though.
If the board is 11x11, then you might be able to make the sprites 5x5, since the vertical screen will only be 55 pixels.
-
OFF TOPIC
When few people are crazy, we call it psychopath.
When many people are crazy, we call it religion.
Agreed.
-
This is a revised idea using monochrome and 4x4 tiles:
(http://img38.imageshack.us/img38/820/image1ib.png)
Off-topic @ above: Please don't hi-jack project threads into moral, philosophical or religious debates. It will most definitely derail the discussion, and I will have to go through the needless hassle of starting a new thread just to draw focus back to the original subject.
-
OFF TOPIC
When few people are crazy, we call it psychopath.
When many people are crazy, we call it religion.
Agreed.
yeah, really off topic.
but i think grayscale is needed, because note the darker-grey kings in the corners
-
Zera, I'm willing to give this a try. Let me know how I can contact you to get all the information I need.
Also, I haven't messed with AI much, so I'll probably need some help (at least on concepts) there.
Anyone else can give this program a shot too, just in case I can't finish it. ;P
-
I like 4x4 version but I would prefer a 8x8 monochrome one I think
Also we need to avoid extremly off-topic posts indeed x.x. I usually just create a new topic for it in Randomness or put it as side comment to an off-topic post in reply to the person concerned
-
It is tempting but I just don't think I'm experienced enough to try this. But for the problem Gale said all you would need to do is make sure its a different kind of sprite I think. That should make it so the "darker-grey kings" are there.
-
I like 4x4 version but I would prefer a 8x8 monochrome one I think
Only problem with that is the board it 11x11 which means that we'd have to make the board scrollable. (which is doable but can be quite annoying) =(
-
An 8x8 version would be great, but the entire board wouldn't fit on a single screen. I'm not sure that would work too well. The player really needs the ability to ascetain the entire board at once.
The 4x4 idea I have so far seems okay. You can discern all the different pieces. The only trouble now is creating the unique board tiles. That's difficult in monochrome - even if I used 5x5 tiles.
-
Is that last screenshot 4x4? It doesn't look it but I could be wrong, I don't know.
-
Are tile resolutions like 5x5 feasible? I've always believed they had to be multiples of 4 or 8.
Is that last screenshot 4x4? It doesn't look it but I could be wrong, I don't know.
The pieces are. There is some border along the outside of the grid, as well as 4x8 text tiles.
-
Yeah, those sizes are feasible. Not as easy, but still doable.
Made even easier because the board doesn't move or change too much. Mostly just adding pieces.
-
i guess 4v4 or 5v5 would be better then
-
Oh, they look 3x3 but I guess I'm miscounting. Thanks :)
-
Oh, they look 3x3 but I guess I'm miscounting. Thanks :)
You're seeing correct. There's a blank row at the top and right side to separate the pieces
-
What do you think about this? (using 5x5)
(http://img337.imageshack.us/img337/5578/image1d.png)
-
nice!
-
Biggest trouble is representing the king's throne. This is the only board tile that needs to be emphasized in some way. It appears in the center, as well as each corner. The tile is important in determining both captures, and victory.
-
I'm really willing to give this program a try!
Will you either attach the rules on post or PM me them? Or if you don't have time to write all of them, I can always wiki it.
Edit:btw, the 5x5 tile looks really nice!
-
Here (http://www.leikmot.net/englisch/ehnefatafl.html) is a great page detailing specific rules - captures, victory conditions, etc. It also has diagrams.
I should note that the game is very historic, and as such, there is not an official recollection of the rules. What we know about the game is mostly extrapolation. The above site says Hnefatafl is played on a 13x13 board, but this is actually debated. Most versions I've seen use 11x11. (including the one I'm planning) Some versions don't include the more advanced capture rules; like shieldwall. I'm trying to make a version that is as thorough as possible.
I don't have all the assets completed - just what you've seen in the thread so far. AP is currently looking into coding it, so I don't want to bring anyone else in on the project unless they consult with him, as well. It might be awkward if multiple people took a stab at it, and coded several versions at once. :P
Current shots:
(http://img714.imageshack.us/img714/382/image61h.png) (http://img194.imageshack.us/img194/7924/image1ms.png)
-
I love these ^^
-
Using those 2 pictures and my limited knowledge of Hnefatafl, the progress so far is attached.
It's just the title screen, picking attacker or defender, and showing who's turn it is based on your choice. (With the GUI so far.)
Posting this because I couldn't sleep and I wanted to show everyone that I'm definitely going through with this project.
(Zera, let me know what you think so far and if there's any changes you want made.)
EDIT: Core engine's mostly done. Just missing rules on capturing king, the shield-wall rule, and winning in general.
Of course, I haven't even started on the AI yet. I'll throw up another screenshot later.
-
Wow nice! Glad you are working on it, {AP}!
I better start reading the rules to play on calc!
Question tho. Will it be available to play against human, not just AI? It would be fine on calc, I guess...
Oh, and how about AI vs. AI for people who wants to get a quick glimpse of some of the strategies that AI use?
I know I'm asking too much at this point, but I want to know if it would be feasible.
-
Great job, AP! Looks very accurate so far. I like the layout you use for the player status info.
Question tho. Will it be available to play against human, not just AI? It would be fine on calc, I guess...
Oh, and how about AI vs. AI for people who wants to get a quick glimpse of some of the strategies that AI use?
I know I'm asking too much at this point, but I want to know if it would be feasible.
I suppose that is up to AP. I don't know what the technical limitations of pure BASIC would be, in that respect.
I originally only intended for A.I. versus human.
-
Wow very nice indeed. Nice job so far, I hope you finish this :)
-
Alrighty, I've almost completely finished the engine which means I'm closing in on coding the AI.
(Progress on Engine: Everything but shieldwall rule and the custom options (which we hope to make into the final version))
This is where I need a bit of help. If anyone has had some experience coding an AI and/or can write up a bit of pseudo-code to get me started it'd help greatly. If you need, here's the notes on behaviors I took from Zera, but I don't need you to write up the whole code, I'm just looking for an efficient way to fit an AI of this caliber into RAM.
General and specific A.I. strategies:
-------
Black
-------
- Initial movement: Black at 6B will move to 6C, 6J will move to 6I, 10F to 9F
and 2F to 3F. (if not obstructed, or wandering into a capture) This is a common
opening move for the black side.
- Will attempt to surround the throne from all sides whenever possible. (but
will not carelessly position itself to become captured)
- If the king moves toward any corner boundary, all remaining black pieces will
attempt to obstruct his path, and pre-emptively obstruct his path to other
corners as well.
-------
White
-------
- Initial movement: Pieces at 7D, 7G, 5D and 5G are among the first white
pieces to move, as their movement does not set other white pieces up to be
captured. Almost any other movements would result in this. Until these pieces
are captured, or black pieces are not prepared to capture any white pieces up
against the throne, other white pieces will try to avoid moving around.
- Pieces directly bordering the king's throne will remain stationary as long as
possible, to avoid giving black the opportunity to surround the king. If the
outer white pieces are moved away and the bordering pieces could be captured up
against the king's throne, then they will move themselves out of the way.
- The king will generally not attempt to move unless a number of white and / or
black pieces have been captured, or there is a very expedient route for his
escape that isn't likely to be obstructed. (i.e., a corner of the board is
almost completely empty)
-------------
Other Notes
-------------
Well, it may just be necessary for the A.I. to only plan moves that protect its essential assets,
and go after essential opponent assets. Otherwise, the A.I. would just move toward the most
expedient route of capturing opponent pieces.
For the black side, the four corners would be considered assets that need to be guarded, and the
king's throne would be considered an asset that needs to be surrounded. For the white site, the king
would consider the four corners an asset that simply needs to be reached, and the white pieces
would consider the center throne an asset that needs to be guarded.
When there are no available moves to pursue these assets, or they are already protected / seized,
then the A.I. would just choose the fastest route to capture opponent pieces.
That is, for all intents and purposes, good enough to have a decent A.I. It does not provide any kind
of real strategy or thinking to the A.I., however. The A.I. would not be robust enough to plot ways to
trick opponents, or make them arrange their pieces so they can pursue their assets more easily.
-
As a side note, we are planning to include a configuration option, since the game rules aren't always agreed upon by various sources. These will allow you to pick custom rules; such as:
- King is required to make at least two moves to escape (the king can't leave the board in a single turn)
- Shieldwall captures on / off (the board boundary can be used to corner and capture large numbers of opponent pieces)
- Winner is determined after a single round, or players must switch sides after the first round, and the winner is determined by who won the most rounds or captured the most pieces
-
Nice to hear some progress. Btw should I make a subforum for this?
-
I'm not sure. The project is actually amazingly close to done. (Big obstacle is the AI really.)
Aren't the subforums just for projects being worked on?
-
Aah ok then, and yeah they are.
(I wish you good luck with AI)
-
A sub-forum wouldn't really be necessary, since there isn't a whole lot of detail surrounding the game. All those details can be hashed-out in a single thread. (plus, the project is pretty small in scope, anyway)
I didn't expect it to progress this quickly, though. AP is doing a great job. :)
-
yeah true, it's amazing at the speed he progressed. I went to work then he alerady had most stuff done o.O
-
Alrighty, we're still trying to figure out a way to handle the AI, but until then, I made a 2-player version. (Pass-the-calc style, not linked.)
There's no readme yet since this is mostly a test, but here it is.
Run prgmA to start it. (This will take up pics 1-3 and matrix [A] as well as various variables (to be expected) and L1 and L2)
Hit enter to select the piece you want to move and the arrows pick where to move the piece, then enter again to finalize the move.
It'll say whose turn it is at the bottom right. To cancel a move, just hit enter on the spot the piece you're trying to move is in.
That's really all there is to the controls. To actually learn to play the game, you'll have to use this site (http://www.leikmot.net/englisch/ehnefatafl.html) for now. We'll write up a readme later.
(Note, the screenshot is no where near what an actual game would look like. I just made white win as fast as possible and showed black capturing a piece. Just to show you that the engine is pretty complete.)
-
mhmm I like the pass the calc style for such game because you don't need two calcs to play it ^^, very nice and quite fast so far ^^
-
Thanks. =)
I just reuploaded on that last post. Zera pointed out various bugs. (All just not completely following the rules. All captures but sheildwalling now supported. Also, fixed a small graphical bug on the win screen where the right side of the box was not drawn in.)
-
Excellent job, AP. This looks great so far. :)
We're having difficulty hashing out the details of how the A.I. would function, but we've come up with something that might produce a somewhat feasible A.I. Not sure if it will actually make its way into the game, though. Guess we'll have to wait and see.
-
Yeah that's one of the challenge of BASIC: speed. So the AI has to be optimized like crazy to keep it fast enough to be playable x.x. It might be doable, tho
-
That looks awesome! I'll have to look into it when I have time! Awesome job {AP}! ;D
-
We also have a readme that will detail the game's rules, along with illustrations. (since I doubt many people are familiar with the game) It's completed, so I'm not sure which version he'll include that in.
-
Nice, but please make the readme either in HTML or in PDF. HTML would be better for compatibility, since every computer comes with at least IE, but is harder to create. PDF requires Adobe Reader, but it's free. For doc files you need MS Word Viewer and prepare for compatibility issues if you view the files on a different version
-
I already got an HTML readme all set-up.
Think I should have the images with the readme or host them online?
If it's with the readme, then it works even without being connected to the internet but it's huge.
If the images are hosted online, the file is a LOT smaller but they have to have internet to see it.
I'm thinking the latter is better (since they need internet to get the file anyway...) but I have had time without the internet before.
I don't know... what do you guys think?
-
It shouldn't be a problem to include the images in the HTML, since I assume the whole package is going to be archived together. The filesize won't be anything to gawk at. Honestly - I think any HDD can spare 200 KB or so. :P
I would mostly worry about the hosting going down at some point. If people are downloading the game a couple of years from now and they can't pull-up any images from the serve, then it would kind of suck for them.
I'm not really opposed to the idea, though. Whatever you think is best.
-
You have a point.
With the file it is!
I'll release one more version here later (probably on the 2nd because I won't be online tomorrow) and then, if I figure I can't manage the AI, I'll upload a version to Ticalc.org.
-
we can have images integrated in HTML files? Is that supported by all browsers and how does it work? Are they in HEX/bin or something? Or do you just mean include them in a folder? I think it might be best to do the former in this case, since some ppl have no internet at home.
I'll upload in download section when you posted new version on ticalc ^^
-
It shouldn't be a problem to include the images in the HTML, since I assume the whole package is going to be archived together. The filesize won't be anything to gawk at. Honestly - I think any HDD can spare 200 KB or so. :P
I would mostly worry about the hosting going down at some point. If people are downloading the game a couple of years from now and they can't pull-up any images from the serve, then it would kind of suck for them.
This is why I'd include them. :)
This is a great game! Awesome job! *Now if only I understood the rules... ;D
-
Or do you just mean include them in a folder?
Yes, that's how we're doing it. It's impossible to do any other way without being connected to the internet. (as you thought) ;P
Also, I'm going to release the source of the code commented separately in case someone wants to add to the project since my motivation has died out recently. It'll contain what EVERY variable is used for and what's going on at parts of the code. It's written in pure basic so it shouldn't be too hard to pick up and change around after the comments are included. I was really hoping to get the AI going, but it's proving more difficult than I can handle.
I'll have the downloads ready later today. I'll be a bit busy now.
-
If anyone does decide to pick-up on the code and wants to implement some sort of A.I., they can contact me for the details we've hashed-out so far. I had a few ideas, but I'm really not sure if any of them would be too feasible.
Either way, it's cool to have a strictly two-player version, at the least. :)
-
I'll try it^^ I want to try to implement AI
-
I hope you don't give up AP but nice work so far on this! I hope your motivation returns at one point
-
Here's the final version I'll be doing.
Let me know of any bugs since I haven't quite tested out this last one but I'm pretty sure that it still works just fine.
Readme included and all.
The source with comments is separate. It's just a txt file and it's not 'copy-pastable' to any TI-editor I know of. Just use the files in the zip of the release when editing and refer to the txt file for any questions on what the code does or what the variables used are for.
If you have any further questions about the code itself, contact me however you wish. If you have any questions about the AI, contact Zera.
-
Ok cool downloaded it now ^^
-
For me, somehow source+comments are cracked. I see bunch of chinese characters... don't know why -_-;;
oh well, I'll study the program itself
-
Looks great. :)
I like the changes you made to the readme since I last checked it out. The BG color is definitely more fitting.
I'm wondering if this will also be hosted in Omnimaga's downloads section. Might make it easier to find, for those who don't regularly browse the forums.
-
I'm gonna look at the code, and try to implement the AI...
If it doesn't work, oh well. 2 player is also fine with me^^
I just have to teach my friends how to play it. I really like the readme too. It looks professional
-
If you can't see the source+comments, try this site:
http://removedfromgame.com/source+comments.htm
-
okay i got it. Thanks^^
-
Looks great. :)
I like the changes you made to the readme since I last checked it out. The BG color is definitely more fitting.
I'm wondering if this will also be hosted in Omnimaga's downloads section. Might make it easier to find, for those who don't regularly browse the forums.
yeah I thought about it, it will be much easier to find. I don't know yet for the AI version, it would most likely need to include the multiplayer one, contain code from it or be by an Omni staff, but since jsj is working on it it will most likely land in there. I might upload the multiplayer version earlier, though.
-
Okay, seriously, I'm working on it. I intend to put in the menu for the custom rules, make A.I., save games, etc.
Also, I changed the board design a bit, and the king piece. Here's the screenshot of the board.
Tell me if you like it. If many like the older one, then it's not hard to change it back^^
Also, there's the screenshot of the menu.
Edit: lol I just found the spelling error on the Hnefatafle rules... there's not supposed to be extra 'e' at the end. Just changed it^^
-
I love how this looks like, nice job guys so far :)
-
I originally included dots along the board grid to help the player discern each tile, but AP left them out for some reason. I assume there was probably a reason for that. Not really sure. Either way, I do like that you added them back in. (ableit, in a slightly different fashion than what I originally produced)
I can't say I'm too fond of the new king design, though. In games where the pieces weren't represented as actual figurines, I've noticed the king's piece usually had a cross to differentiate him from the other pieces. I tried to follow in that method, but it's difficult to produce something very detailed with such a small resolution. On the other hand, your version does seem easier to discern than mine.
I attached some of the working ideas I had for an A.I. script, if this could help in any way. I was thinking one way to handle the A.I. is to give it a list of various moves. It would cycle through this list, from priority of first entry to last, until it came across a move that could be executed; then it would execute it. If it couldn't execute any of the moves on the list, then it would select a "random" piece and move it to a "random" location. (within the confines of the game's rules, of course) The moves list I created details what I think would be the most essential moves a player could make in order to win the game. It would create a very deterministic A.I., and the A.I. would make a few dumb moves or set itself up for a capture. It would be *some* form of A.I., though.
-
I made the king like that, because when I first played it, it looked more like black's piece, not white's. So I took out two pixels... and thanks for the AI text! It will help me immeasurably!
After researching, I intend to include these custom rules:
Castle- Affects the nature of the central square
King Capture- Affects how the king is captured, either with 2 pieces or 4 pieces
King Strength- Either Armed or Weaponless
Shieldwall
# of rounds
-
I haven't heard of the first two rules. The "armed king" rule, to my knowledge, is just a house rule someone else created which doesn't fall into common use.
Just me, personally - I would only include the shieldwall on / off rule, the king movement rule, (king needs to make two moves in order to reach a burg) and the round type. (either a single round, or require players to switch sides and play a second round) The others would just be too uncommon. If you're taking over the code, though, what you include is your prerogative.
If you want to do some testing on the A.I., I can help with that, since I'm very familiar with the game. The notes I included earlier are just some essential moves I've thought of. If we get into testing, I might be able to examine the A.I.'s behavior and define some additional movement rules. It's kind of hard to create rules that would allow the A.I. to actively capture opponent pieces, but some of the rules I already included will capture according to a few common exploits and oversights that most human players would set themselves up for.
-
oh okay^^ but I don't get the king movement rule, what do you mean by making two moves in order to reach a burg? isn't the king automatically making two moves to get there? like, going all the way to the edge, and then corner?
-
This rule means that the king must first position himself next to the burg, as opposed to moving directly into it in a single turn - that is, as he is moving toward the burg, he must stop immediately next to it and end his turn. On white's next turn, he can move into it and end the game.
This rule tries to balance the strategy between attackers and defenders, as it's fairly simple for the king to find a safe, unobstructed route of escape.
As for the movement you detailed above - the king's movement is like that of any other piece: He can move any number of spaces in any straight direction. He does not need to leave the board in any specific number of moves. (that is, he doesn't have to shoot straight for the boundary of the board, and then a burg) The king can, in fact, be chased all around the board and back while he's attempting to escape.
By the way - if you want to get in touch with me about any specific rules or details, you can always hit me up for a PM conversation. I'm also on AIM, Yahoo, Google Talk and IRC.
EDIT: Upon further research, I would give more consideration to the "unarmed king" rule, as well. Apparently, there are a sizable number of sources that know about this rule. It even appears in some computerized versions of the game.
-
Oh okay! Those won't be too hard to implement^^
The rules that I wanted implement was from the java version of this game. So I figured it was variant in many sources, but apparently not... And I'll implement the unarmed king rule. They are not that hard to do.
I think I'll finish this either by tomorrow or tuesday lastest. Wait... I have to make the AI and test it, so it might go upto wednesday.
-
Good luck! I hope you don't have too much speed problems :)
-
Okay, I got the "King's Escape" (meaning the king has to take 2 turns to get to burg) and the weaponless king implemented. When I finish the king's capture along with shieldwall and the AI, it is pretty much done.
Also, I implemented the saving feature. Coming along very nicely, way better than I expected :)
-
glad to hear it's going nicely :), keep up the good work :)
-
Good work, indeed. I didn't expect you to put this much detail into it. :)
-
So this is pretty off topic (but related to the game) but how do you even pronounce it? I skimmed and didn't see anything so sorry if I missed it :P
-
It's pronounced like Hentai-Fatal I think
-
Hmmm, ok. Thanks.
-
I was kidding :P
Because for some reasons the first time I saw this topic title I misread it as hentai-fatal x.x.
-
Something like neh-puh-topple
I suppose you can substitute the P sounds for V sounds, as well. P, V and F are all rather ambiguously sounded in some languages.
-
Haha well it made sense to me so I didn't question it :P
But thanks Zera for a correct pronunciation :P
-
I'm uploading what I have so far... I still didn't implement the shieldwall, and A.I., and the surrounding capture (that looks like GO capture).
What I did accomplish from {AP}'s game are:
- different board design
- choice of who goes first
- # of rounds to play (1~5)
- Win condition (either highest wins or highest captures)
- 'King needs to make 2 moves for escape' option
- Saving/loading
- Gameplay log (although it only stores last 2 moves)
After I'm done with the three things that needs to be done, the game should be complete...
btw, there may be some bugs, especially dealing with who wins, etc. I tried to fix all the bugs that I found, but there might still be some that I did not see. So if you find one, please tell me. what the settings were during the time of finding bug would really help too.
And, I'm too lazy to make an animated screenshot right now :P If anyone wants to do it, go ahead. ^^
-
great update! Question, though: by "choice of who goes first" you mean it's random, right? Not letting the person choose?
-
no, it's choosing whether the attacker or defender goes first. It's the first setting in configuration page
-
A few concerns:
- The burgs and king's throne should look alike. The tiles have the same properties. It's more of a technicality, though. It doesn't hurt to have the king's throne use a different design, but it is kind of awkward to someone who plays the game regularly.
- The number of rounds and winning conditions are kind of odd. Normally, the winner of the game is determined through two rounds of play. At the second round, each player switches sides. Whoever won both rounds wins the game. If it's tied, then whoever had the most captures wins the game. These are not separate conditions.
- Who goes first should not be randomly determined, or adjustable. Whoever plays the attacking side makes the first move. When players switch sides in the next round, this means both of them get to go first at least once during the entire game.
It probably doesn't hurt the gameplay balance to have these rules adjusted any which way, though.
EDIT: By the way - you may want to edit the title screen to include your name in the initial credits, too. :)
-
Oh okay^^ Will change those.
-
Today, I spent the entire morning researching on the AI for the board games. I found some interesting articles, and it seems like Minimax Algorithm (http://en.wikipedia.org/wiki/Minimax) is good for these kind of games. I don't know if it will be fast/efficient enough with BASIC, and I still haven't read all of it yet. I hope it works and if it does, it will definitely lead to a very good AI, not too hard, not too easy.
-
Nice I hope it runs fast enough for your needs.
-
Approximately how long should the AI think?
I want it to be playable in regular TI83+, so in that speed, would like, 1 minute enough? or is it too long?
-
If the A.I. is plotting a move for over a minute, that's definitely too long. The player is going to think the game has become unresponsive. I'm thinking no more than 6 seconds or so - if that's feasible.
-
hmmm... yeah a minute would be too long -_-;;
Well, I haven't really coded any AI part yet, except black's first move, and I don't exactly know how fast it's going to be. Hopefully, less than 10 seconds. If it goes over, then I guess I have to make the AI less intelligent, or make a progress bar for AI's thinking to show that the game is not unresponsive.
-
10 seconds may be ok. If there are difficulty settings, have the AI think for less time on lower difficulty settings
-
pronunciation is [hn'єfatafl] I think, you can practically read it as if it were phonetic notation. I'm not an old-norse expert, but I've heard some modern Norwegian. (got a norwegian-speaking friend)
-
Finally... I got some progress on AI. The attacker can *sometimes* eat white piece. There are still a lot of debugging to do, because there are times when attacker doesn't move at all (which is illegal btw).
The thing is that, it is kind of slow :( It takes around 10 seconds for AI to think and move, but I only implemented like 1/4th of the whole, functioning AI. I guess debugging and working with it might show me some optimization... This AI program takes about 2000 bytes.
Hopefully, I can get functional yet not so slow AI by the end of the week :P/me goes back to work
-
Good job. :) I was kind of wondering what happened since the last update. :P
10 seconds isn't too bad. Hopefully, optimization will tone it down a few seconds, though.
-
Nice to see progress. I hope there's a workaround for the slow AI. It doesn't seem too bad, though. Do you think there could be ways to do it in a faster way even if it sacrified file size? In TI-BASIC, sometimes it's better to optimize for speed rather than size. In a space shooter, if you need to display 12 ships every frame, never do:
For(A,1,3:For(B,1,4:Output(A,B,"M
End:End
Do this instead:
Output(1,1,"M
Output(1,2,"M
Output(1,3,"M
Output(1,4,"M
Output(2,1,"M
Output(2,2,"M
Output(2,3,"M
Output(2,4,"M
Output(3,1,"M
Output(3,2,"M
Output(3,3,"M
Output(3,4,"M
You'll waste about 70 bytes by doing this, but the speed gain will be considerable. You'll almost get half a frame back because For( loops slow things down so much.
-
Hmmm. I do use For group, but if I didn't, it would be extra 1000 bytes at least X.X
I might try it tho, if I can optimize them. However, first things first, I'm gonna try to implement all aspects of AI, including not making stupid choices like moving into position where you will be eaten, defender AI, making different choices based on the custom rules, etc. Now that I got the basic attacker engine, I think I will be able to do these more easier.
-
Okay, I implemented an attacker AI. It's still not complete, but it is doable, although it takes some time for AI to think (around 20 seconds or so, depending on legal moves that it can make)
There are probably tons of bugs, like what you see at the end of the screen shot (it's 4 minutes long -_-;; most of the screen shot is AI thinking :P )
Shieldwall not yet implemented, victory condition not yet implemented (well, it does check for victory, but I haven't yet wrote the code to tell you that you won or lost, and ends the game), Defender AI not yet implemented, etc. There are bunch...
The Attacker AI is mediocre at best, because it doesn't search the whole board for the good position. It only searches few areas, so sometimes it misses a real obvious moves (like surrounding the king to win).
Oh, btw, when you try to quit the game, it will give label error, because currently the program is split into many subprograms to make my life easier coding. So if you want to quit the game, just press 'on'.
Otherwise, I think it's pretty good. If there's any kind of bug that you find, it would help me a lot if you can tell me^^
Since the file names seem to be already used, download the game
in TI group file (http://www.mediafire.com/?gww0m0nd4dj)
in zip file (http://www.mediafire.com/?gmluy3k0mbo)
-
There's a bit of a quirk in the rules: When black moved to 4F, the white piece against the throne wasn't captured. It should have been.
Any piece positioned against the throne (regardless of whether the king is occupying it or not) is captured when an opposing piece positions itself to the opposite side.
-
This looks very nice. If it wasn't for the slow AI speed and the game beginning sprite loading, it could almost be mistaken as an ASM game.
I will try it when I get some time :)
-
There's a bit of a quirk in the rules: When black moved to 4F, the white piece against the throne wasn't captured. It should have been.
Any piece positioned against the throne (regardless of whether the king is occupying it or not) is captured when an opposing piece positions itself to the opposite side.
Yeah, I know about the rule. I was aware of this, and I have to fix the capturing. I fixed some more bugs, and looking for speed optimization. Hopefully the AI can get down to around 10 seconds
-
Since the file names seem to be already used, download the game
I usually postfix my file names with the date so I don't hit that problem.
-
I just add 2, 3, 4 and so on depending of the version. Just click your file, rename it, then choose it to upload. Not too hard.
-
I did that, but somehow, it didn't let me upload. It was like, Hnefatafl zip file 1
but still, i couldn't, and I don't think I can change the name of the 8xg file, else doesn't it get corrupted?
-
strange, then you need to reformat your computer x.x cuz it's a problem on your end.
-
I haven't been working on this from last update. I had bunch of tests from school, including Calculus, English, Biology, etc. So I was busy writing math programs for Calculus (downloading programs from ticalc and modifying them to fit into my use) and studying English and Bio. I will resume working on this starting from today.
-
I hope you have success in your exams. Can't wait to see updates on this, factory and TLM when you finished exams :)
-
oh yeah, the factory. lol I totally forgot about it :P I hope to finish this within a week, and start working on TLM again.
I think I did pretty well on the exams. I don't have anything major now, except AP tests given on May.
-
aaaah ok ^^, glad to hear ^^
As for Factory I was worried x.x
-
Went ahead and updated the first post so everyone knows the new download is on page 7.
Was the king's throne issue addressed, or is that going to be in another release? The other capturing mechanics seem accurate, though.
-
Btw is the game supposed to work when both players are set to A.I? Will the two AI's try to beat each others or will it error?
-
The game is supposed to work when both players are set to AI, and I want them to play differently. And the king's throne will be in another release (I already fixed it, but I want the defender's AI to be working)
-
Aaah ok I'll give it a try then
-
Wait... if you set the AI as defender, it won't do anything. I haven't coded for defender yet. Right now, I'm figuring out the way to optimize the attacker's AI.
-
Oooh ok, thanks for the info
-
YES! I finally found a way to make the AI faster!! I made a small look-up table that gets created every time AI goes. The position and its value that the calculator already found is stored into two lists, one for position and one for value, and if the calculator later tries to search for the value of the position that it already found, then it just gets the value from the list and move on. Using this, the AI speed got cut in half, resulting in ~10 seconds for the AI movement. I have to run tests for any bugs or things that might be present, and exactly how faster it runs, but I think it's good enough. And the amazing thing is that it only added ~50 bytes to the program (although it requires two ~30 dimension lists during the run-time)
-
That's awesome. Sounds like A.I. time will be significantly cut short.
Some rounds could go on for a pretty long while, depending on how skilled the A.I. is. It might have been tedious to wait some 20 seconds for it to execute its move each time.
-
Wow nice optimization. I agree with Zera. Sometimes I played chess with my bro and we would think for 20 seconds each sometimes, causing a game to last a few hours x.x
-
:'( Sorry I lied. It did make the AI think shorter, but only on certain times when there are overlapping position thing. I was excited that it took 10 seconds to take the 2nd and 3rd move, but as more possible moves opened up, the longer AI took to make a move... AI spends 8~70 seconds depending on available moves possible. And note that I'm running the program with TI-84+SE, so it would take twice as long on TI-83+... :(
The AI itself is not too stupid, sometimes it makes obvious mistakes, but it also surprises me with the move that I would normally miss if I didn't look carefully. But it would be easy to destroy AI if you are intermediate player playing white.
I tried to make another version of AI, where the list contains all the possible moves and their values, and the positions and values fluctuate, but it took too big of a memory (two 1000 bytes lists), speed gain wasn't too big, and the AI got more stupid, so I got rid of that idea.
What I am going to implement is the alpha-beta pruning (although idk if that's going to lower the speed significantly) and some more minimax algorithm. Also I might try to get rid of the For( group (it will increase the size of the program a lot, but I think for AI the speed is the utmost importance). Hopefully, I can get something to work.
-
That might not be a problem, if the longer moves are the ones that occur less frequently. If those types of moves aren't generally important to the A.I.'s skill level, then maybe they could be removed, or supplemented with another strategy.
-
Well, it seems like that the AI takes longer and longer as the game goes on, because there's more available moves possible when the pieces are more spread out and the board gets more complex. So the average speed seems to be around 30~40 seconds :(
-
ouch sorry to hear, I wish you good luck
Hopefully a solution can be found x.x
-
yeah... I learned through this that programming AI can be a real headache.
-
I myself never got too far into AI. It seemed so complicated for me and I feared my code would alwyas be slow, especially if I made a game with real-time AI