• Advertisement


  • Content count

  • Joined

  • Last visited

Community Reputation

102 Neutral

About Plusekwal

  • Rank
  1. 3d Animation program

    Does anyone know a 3d animation program that is easy to use? It's very hard to find 3d animated models on the web, so I thought about animating on my own, but I dont want to spend weeks learning how to. Thanks in advance.
  2. C++ + OpenGL - where do I begin

    Hi! I want to make games using c++ and opengl. However I dont know where to start. -which and what files should I download? -which IDE would you recommend me? -how do I start using OpenGL in C++? -what is a typical base for a game in c++? -what files does opengl have and where should i put them relative to the folder of my project? are there dlls and such? thanks in advance p.s. currently i have only visual studio 2008 - is it good for making games?
  3. games for android

    I want to make games for android, where can i get started? Which language and IDE would you recommend me?
  4. This is the first serious game I have finished, so I want some feedback on it. trailer: [media]http://vimeo.com/33864437[/media] it is all about capturing the enemy pawns while creating your ones which makes kupChie(the name of the cube on which the game takes place) to rotate You can play in singleplayer, multiplayer or in the campaign which is simply beating the computer on increasingly difficult levels Screenshot [img]http://kupchie.comxa.com/screenshots/shot.JPG[/img] [url="http://www.box.com/s/iyu8d8bodo0hrvd1bvm7"]DOWNLOAD LINK[/url] it runs fine on my computer with the following specs: Windows XP 1.7 GHz processor 1 GB of RAM if you download the game, please do not skip the tutorial and tell if there's anything unclear how to finish it or in it
  5. Socket creation

    What might be a reason an application to be failing to create any sockets under windows 7 and in the same time to create them and run fine under windows xp?
  6. the error was [quote] directive "netsh add allowedprogram {app}\game.exe AppName ENABLE" has no value[/quote] I'm trying to add an exception in the port forwarding table using code. The protocol I use is TCP.
  7. The multiplayer of my game for some computers works fine, however others are unable to connect with other players since some routers are blocking game packets. I know I need to add an exception in the port forwarding table of the router. What I don't know is how to do it. I tried putting [code]netsh add allowedprogram {app}\game.exe AppName ENABLE[/code] in the installation file of Inno 5.3.1, but it gave me an error.
  8. Pausing the game in multiplayer

    Thanks for the replies. I think I will add both the both the automatic player thing Tom Sloper suggested and the pause - it will be without vote, but the player will have to give a reason in order to pause. The pause will be limited to 3 mins and there'll be a small countdown before the game is unpaused if cancelled earlier. There'll be also one pause per player. During the pause the screen won't go black or gray - infact the only difference will be that players wont be able to make moves and the time will be stopped and it'll write PAUSE with the time remaining somewhere. What do you think about this? Also, I think it's important to add that there won't be any teams or stats or even profiles - [u]everyone can play with different nickname every time s/he joins the multiplayer room[/u]. Also there won't be any competitions (or atleast there wont be built in system for this) - so I think there isnt really any reason to delay the already lost game.
  9. I'm developing a puzzle game where players take turns to make their moves. The game features a multiplayer mode over the net, and I've there put a time limit of 45 seconds per turn so games dont last forever. My question is: is it a good idea to enable the players to pause the game when they need to do so? Imagine you're playing the game and you are being called on the phone or you need to go to the toilet or something else which prevents you from playing for a while - it would be annoying to lose a turn or even the game just because of that. The pause would be limited to 5 mins and one player would be able to pause the game only two times so it doesnt get abused, but still that is likely to annoy the other players and attract internet trolls. The chat will still be available when the game is paused. So what do you think, do the advantages of implementing a pause make up for the possible drawbacks?
  10. [quote name='alvaro' timestamp='1311010800' post='4836898'] How do you lose pieces? Also, after a new piece is spawned, the coloring of the squares involved goes away? [/quote] You just move onto enemy pieces to destroy them. And the colour doesnt go away, so this is a kind of limit for the number of possible pieces. However, the enemy player can always change the squares in your colour into his colour.
  11. [quote name='IADaveMark' timestamp='1310922617' post='4836418'] [quote name='Plusekwal' timestamp='1310884380' post='4836264'] Analysing the potential moves by even only one turn after the AI will take way too much computing time. It looks like I'll have to use MCTS [/quote] [img]http://public.gamedev.net/public/style_emoticons/default/blink.gif[/img] From the sound of what little you have told us about your game, your state space can't possibly be THAT big. Either that or your are running the game on a platform with the computing power of an abacus. [/quote] Well, three moves per turn can have many possiblities and as far as I understood from that website, minimax needs at least three to four turns to compute. Also, there's some pretty complicated math involved when creating new pieces. [quote name='Emergent' timestamp='1310950298' post='4836559'] [quote name='Plusekwal' timestamp='1310884380' post='4836264'] I was asking if minimax can use only the results from one turn ahead. [/quote] Oh, ok, gotcha. [quote name='IADaveMark' timestamp='1310922617' post='4836418'] From the sound of what little you have told us about your game, your state space can't possibly be THAT big. Either that or your are running the game on a platform with the computing power of an abacus. [/quote] I'm not sure about that. The board is small, but each turn is 3 moves long. Naively, he'd have to search 6 plys just to see a full turn ahead -- not so much "min-max" as "min-min-min-max-max-max." Naively, this gives a min-max branching factor of (4*10)^3 , which is huge.* So it's not trivial. That said, many of these turn sequences result, after 3 moves, in identical gamestates, so I don't think the "true" branching factor needs to be so high. It could be kept down e.g. by hashing visited states, or perhaps by thinking carefully about what states are reachable. *(I'm assuming 10 pieces, each of which can move in one of 4 directions.) [/quote] I was actually considering maximum three moves per piece to be considered - one for eliminating maximum number of enemy pieces, one for colouring(see below) and one for escaping/defending own pieces. [quote name='alvaro' timestamp='1310954897' post='4836596'] The enormous branching factor that results from compound moves is a serious obstacle to the success of minimax in certain games, as I mentioned in my first response. So MCTS seems promising. You can consider each separate piece moving as a proper move, since MCTS doesn't require turns to alternate, the way minimax does. If the nature of the game is such that random moves will progress towards the end of the game, you might be able to set something up without too much effort. If not, you either need to implement a biased playout policy that will make progress towards the end of the game, or you may have to combine an evaluation function with MCTS, so after a few moves from each side you'll use the evaluation function as reward. Please, let us know if you are making any progress. Also, it would be cool if you could post a full description of the rules. [/quote] I'm still wondering if switching to MTCS will be worth it. I dont know if MCTS will be applieble to this game. Also, computing time might not be that big after all. As for the rules: -the board is 9x6 squares big -the game ends when there are pieces of only one player left -every player starts with one piece only -one piece can do five actions - moving in the four directions and colouring the square which it is currently on in the colour of the player -upon colouring if the new coloured square makes a row, a column or a diagonal of three colourd squares in the colour of the player, new pieces are spawned at the empty places [quote name='Emergent' timestamp='1310955092' post='4836600'] [quote name='ApochPiQ' timestamp='1310950886' post='4836567'] Per piece, you compute a list of potentially reachable spaces on the board (do I move this guy 1, 2, or 3 times?). Then you prune conflicts (i.e. two pieces land on the same square). Then you prune the number of valid combinations using a simple allowance metric (i.e. only 3 moves are allowed per turn). This gives you a finite, bounded set of moves which is not substantially larger than the set of potential moves in an average chess game, if my napkin mathematics does not betray me. [/quote] Yeah, that was my thought too, and it might work, but there are some situations that confuse me. E.g., [code] XXXX XXXX XXXX XXXX XX X XX X XX X XX1X X 2X -> X2 X -> X21X -> X2 X XX1X XX1X XX X XX X XXXX XXXX XXXX XXXX [/code] (1, 2, and X denote pieces; spaces denote empty space; this is a three-move sequence where the order of 1 and 2's moves matter.) [Aside: The code tags do weird things, adding and removing spaces; getting ASCII art to look right is a pain...] I'm not saying that there isn't an efficient way to resolve this so that you keep the attractive properties of just considering what squares are reachable (in which case you just get something like 3k moves), but how to do it doesn't jump out at me. [/quote]As for this situation, piece 2 can move upwards an piece 1 upwards and then leftwards - there is no difference. Besides, due to the nature of the game the board will never get that crowded.
  12. [quote] I'm sorry, but I don't understand your question. I can't even parse it.[/quote] I was asking if minimax can use only the results from one turn ahead. [quote] It sounds like you just don't understand the idea of minimax yet. Just think of it as propagating numbers up a tree to begin with. They start at the leaf nodes, and work their way to the top. Once you have all the numbers, [i]then[/i] think about the actions that they imply. Play with the Java applet at the bottom of [url="http://www.ocf.berkeley.edu/%7Eyosenl/extras/alphabeta/alphabeta.html"]this page[/url] for a bit; you'll see how it works. [/quote] [quote name='IADaveMark' timestamp='1310850440' post='4836140'] [quote name='Plusekwal' timestamp='1310809089' post='4835934'] Anyway, I got a question about the minimax formula: is it a good idea to based on the consequences from the AI's moves only? [/quote] Minmax, by definition, needs to iterate through the potential game spaces. Since the other player is likely changing the game state along the way, you pretty much can't derive a realistic result from Minmax unless you are taking the other player's moves into account. [/quote] Analysing the potential moves by even only one turn after the AI will take way too much computing time. It looks like I'll have to use MCTS
  13. [quote name='alvaro' timestamp='1310648668' post='4835232'] Maybe you should tell us a little more about your game: How big is the board? How many pieces does a player typically have? What are the winning conditions? The type of games that allow multiple moves per player are often hard for computers (see [url="http://en.wikipedia.org/wiki/Arimaa"]Arimaa[/url], whose move mechanics are similar to yours). There are roughly speaking four approaches to making board-game AI implementations (I am improvising here, so don't be surprised if I missed something): * A set of hand-crafted rules: It's what you have now and it can be made fun for some games, but it typically results in brittle AI that makes stupid moves in many situations. * Minimax: It worked great for games like chess and checkers. * MCTS (Monte Carlo Tree Search): It's all the rage in computer go these days, and it can even be used for non-deterministic games and games with more than two players! * ML (Machine Learning): It has worked well for some games like backgammon, but it might require learning a lot about a difficult field, and it's unclear how generally applicable it is. If you can write a decent evaluation function for your game (a function that takes a position and returns a number that gives you an idea of who's ahead), try minimax. If this is not possible, you might be able to use MCTS instead. Neither of these things is very easy to learn, but there is a lot of good info on the web about them. [/quote] 1. The board is 9*6 squares big. 2. All players start with 1 piece and can hardly have more than 15 to 18, simply because the board is not big enough. I'd say a typical number of pieces in midgame is 10. 3. A player wins when there are no more enemy pieces on the board. Anyway, I got a question about the minimax formula: is it a good idea to based on the consequences from the AI's moves only?
  14. I'm making a board game and have reached to the point where I need to program the AI. And I did it - it was easier and took less time than I have expected. The problem is that the AI is trivial and too easy to beat - I tried to do everything I could think of, algorithms work as expected, but it's still quite easy. The rules of the board game are rather simple: there is only one type of unit or piece or pawn or whatever. Pieces can only move to the 4 neighbour squares and can also create other pieces under certain conditions. Players take turns to move 3 times any of their pieces(i.e. 3 pieces once, or one piece three times or 1 piece once and another one twice) The AI first compares the number of its pieces against the total number of all other pieces. Then, depending on the ratio, it uses four different algorithms to react to the situation. These algorithms include checks for undefended enemy pieces, creation of new pieces, defending of its own pieces, etc. I can post the exact algorithm if necessery. So how do I make the AI more difficult?
  15. [quote name='rip-off' timestamp='1308596119' post='4825603'] How frequently were you polling for messages? [/quote]As soon as the game recieved some data from the server, it requested another one. I dont know if the chat getting updated every 5-6 seconds would be nice for the players. [quote]PHP Is *not* a good language for "long poll" type situations. (Google for "long poll" and "comet" to see this mechanism described in more detail)[/quote] That's why I'm looking to implement a push-type chat. [quote]If your choice is "PHP or nothing" then you either have to put in an IRC plug-in on your web site, or live with the terrible performance you'll get out of PHP.[/quote] Are you saying I can't make a good irc client built in the game with PHP? I'm fine with whatever technology satisfies my needs.
  • Advertisement