|
Hi. I am Travis. I played Brood war a very long time ago under a few usernames and the game has helped shape who I am today.
I haven't played in well over 10 years, because life took over. I am now a CS student, and in my free time (the little bit that I have) I am working on a brood war bot.
Here is what I am looking for:
A protoss player, english speaking, who is a very high level player - I don't know what that means these days for a foreign player - like B+ or something on Iccup? (I don't even know if Iccup is still the server of choice or anything).
Basically I am looking for an english speaking protoss player who is capable of competitng with top koreans. You know, like maybe take a game off of them every once in a while.
The purpose would be to discuss strategy - build orders that would be practical for my bot, and how my bot should react to given information. I do have decent understanding of the game - I was very good back in the day - but I haven't kept up with it and I don't know the maps and also I know that the game has evolved.
The consultant would not be paid, and would just need to be someone who was often online just able to chat or reply to messages occasionally or something, because this is a *very* long term project. What they would receive is pride and credit for contributing strategy, and the joy of seeing their contribution eventually come to life.
And before people reply - I know that liquipedia exists. I am looking for much more information than is available there, though.
Thanks!
edit: oh, and for those that are curious. The timeline for seeing a working bot (something competing in http://sscaitournament.com/) is probably by the end of the summer. Mostly since I will have way more free time in the summer. But to actually make a bot top level I anticipate it may take much longer.
|
I would be happy to assist, but I'm playing terran. (probably would fulfill your skill requirements)
Can help a bit on TvP I guess. :-)
|
I'll definitely take you up on that, I haven't followed bw much but I do know of you.
PVP is the first matchup I am focusing on though because it's the most simple. So PvT will come later
P.S. is nw.blackman *the* legendary blackman?
|
On March 28 2017 21:57 travis wrote:Hi. I am Travis. I played Brood war a very long time ago under a few usernames and the game has helped shape who I am today. I haven't played in well over 10 years, because life took over. I am now a CS student, and in my free time (the little bit that I have) I am working on a brood war bot. Here is what I am looking for: A protoss player, english speaking, who is a very high level player - I don't know what that means these days for a foreign player - like B+ or something on Iccup? (I don't even know if Iccup is still the server of choice or anything). Basically I am looking for an english speaking protoss player who is capable of competitng with top koreans. You know, like maybe take a game off of them every once in a while. The purpose would be to discuss strategy - build orders that would be practical for my bot, and how my bot should react to given information. I do have decent understanding of the game - I was very good back in the day - but I haven't kept up with it and I don't know the maps and also I know that the game has evolved. The consultant would not be paid, and would just need to be someone who was often online just able to chat or reply to messages occasionally or something, because this is a *very* long term project. What they would receive is pride and credit for contributing strategy, and the joy of seeing their contribution eventually come to life. And before people reply - I know that liquipedia exists. I am looking for much more information than is available there, though. Thanks! edit: oh, and for those that are curious. The timeline for seeing a working bot (something competing in http://sscaitournament.com/) is probably by the end of the summer. Mostly since I will have way more free time in the summer. But to actually make a bot top level I anticipate it may take much longer.
if you can get your hands on him, Draw is the person to talk to. He plays at the requested high level and is a native English speaker.
|
Hello fellow Bot programmer,
I know certain tricks/strategies that most bots cannot handle (even in 2016 the worker rush worked versus certain bots). So when you have a bot, let me know. I will run my own test bots against it to see if it can deal with any kind of rush.
|
On March 28 2017 23:36 travis wrote:I'll definitely take you up on that, I haven't followed bw much but I do know of you. PVP is the first matchup I am focusing on though because it's the most simple. So PvT will come later P.S. is nw.blackman *the* legendary blackman?
You're welcome to contact me if you have a specific question.
NW.blackman is Julia.
|
Human play and bot play are completely different beasts, and this is more true the higher the level of play is. I don't see how nuances in ways human games develop matter. You will just be up against some bot that has tricks to disrupt the algorithms of other bots, and your bot will lose.
|
I'm a very, very high-level player. You can ask me anything.
|
Australia4514 Posts
I almost did the same thing a few years ago. Most bots that i saw did very poor economy management. Just super simple things like mining from the closest mineral patch as first priority and things.
I would focus your bot on; A) rushing. B) utilizing range units as much as possible.
Perfect micro with range is super powerful, and impossible for humans to compete against.
Once you have that, you would only need to improve strategy once you start facing other bots which do the same.
|
range is why Terran is OP lol
|
Interesting project.
I can´t see a reason why a bot shouldn't be able to dominate this game.
If the programmer can realize the huge potential advantages of a bot in the areas of micro and multitasking, and pair it with good enough decision making... GG.
|
On April 04 2017 04:22 wimpwimpwimp wrote: Interesting project.
I can´t see a reason why a bot shouldn't be able to dominate this game.
If the programmer can realize the huge potential advantages of a bot in the areas of micro and multitasking, and pair it with good enough decision making... GG.
The reason that bots aren't very good is because there is too much decisionmaking, it's very very complicated
things that come naturally to humans (cutting off enemy units, retreat or scout paths, knowing where to defend drops, etc etc) can be hugely complicated problems for bots
making say, a GO bot or even a poker bot is way easier because the games are turn based with a much more limited decision scope
|
Lancerx and dewalt are two russian active players that are A- . They both speak english.
|
On April 04 2017 04:22 wimpwimpwimp wrote: Interesting project.
I can´t see a reason why a bot shouldn't be able to dominate this game.
If the programmer can realize the huge potential advantages of a bot in the areas of micro and multitasking, and pair it with good enough decision making... GG.
Because simply, a human being has a soul and therefore has creative thinking and instinct. It can change its decision at whatever time for whatever reason, making it completely unpredicatable. A machine cannot. The only way that there ever is going to be an AI as powerful as the top human players, would be if there where hybrid macn/machine beings, combining machine macro, micro and multitasking with human creative skill. I hope this is a good reason that you can see now.
|
I bet 10$ that bots will start winning humans in next 10 years (i mean progamers)
|
On April 04 2017 04:26 travis wrote:Show nested quote +On April 04 2017 04:22 wimpwimpwimp wrote: Interesting project.
I can´t see a reason why a bot shouldn't be able to dominate this game.
If the programmer can realize the huge potential advantages of a bot in the areas of micro and multitasking, and pair it with good enough decision making... GG. The reason that bots aren't very good is because there is too much decisionmaking, it's very very complicated things that come naturally to humans (cutting off enemy units, retreat or scout paths, knowing where to defend drops, etc etc) can be hugely complicated problems for bots making say, a GO bot or even a poker bot is way easier because the games are turn based with a much more limited decision scope
I understand that there will be a lot of situations like the ones you mentioned above that the bot will need to be able to handle. And that programming the "sub routines" to properly handle each and one of them may be a challenge and represent quite a lot of work. And also that making sure the execution of these sub routines are handled in a way so that they are not interfering with each other and are prioritized in a sensible way may be tricky. But certainly not impossible.
On April 04 2017 05:26 pebble444 wrote:Show nested quote +On April 04 2017 04:22 wimpwimpwimp wrote: Interesting project.
I can´t see a reason why a bot shouldn't be able to dominate this game.
If the programmer can realize the huge potential advantages of a bot in the areas of micro and multitasking, and pair it with good enough decision making... GG. Because simply, a human being has a soul and therefore has creative thinking and instinct. It can change its decision at whatever time for whatever reason, making it completely unpredicatable. A machine cannot. The only way that there ever is going to be an AI as powerful as the top human players, would be if there where hybrid macn/machine beings, combining machine macro, micro and multitasking with human creative skill. I hope this is a good reason that you can see now.
The relevant outputs of your "soul", "creative thinking" and "instinct" for a computer game scenario where winning is the only objective are still individual mouse clicks and keyboard presses manipulating the game variables in a fairly limited amount of ways. Furthermore a bot is already a man machine hybrid in the sense that there will be humans designing the rules behind all of its decision making.
As for unpredictability there aren't really that much variation in high level brood war games when you view them from the perspective of strategy. Sure the situation of two players facing off against each other in a game of brood war can produce an infinite amount of tactical scenarios, and probably making the bot able to make sense of these and if needed adjust its strategy according to what's going on is another problem facing the bot creator. But again this is only a challenge, not impossible.
|
TLADT24917 Posts
On April 04 2017 06:25 kogeT wrote: I bet 10$ that bots will start winning humans in next 10 years (i mean progamers) you think that TL posters will remember that bet in 10 years?
|
On April 04 2017 06:50 BigFan wrote:Show nested quote +On April 04 2017 06:25 kogeT wrote: I bet 10$ that bots will start winning humans in next 10 years (i mean progamers) you think that TL posters will remember that bet in 10 years?
That you weren't here 10 years ago doesn't mean some of us weren't. But I have to conceide you the point. Considering the state of SC2, Blizzard and TL, I also don't think TL will still exist in 10 years.
As to be more on point, since people are now debating the flaws of bots right now, and how they aren't able to count their number of dragoons/marines/tanks and make a decision about if they will come out on top of a fight before they engage, it is pure folly to ask a A- player to help you for free for months.
|
These are not problems, these are concepts. Concepts a machine doesn' t understand. I have played vs the best bots around, and they are easy to beat, even for me, while i am a d+ player. Sure, not very easy, but once i understand who that bot was concieved, the actions it takes, i could beat it more and more easily. A human being is extremly more complex than any machine i have ever seen or heard of.
Think of these: the best cameras and capturing devices in cinema, machines that costs thousands and thousands of euros, are only aiming to replicate the human eye, which is capable of capturing an enormous amount of information. In the case of bots playing starcraft, you assume that bots can reproduce an aspect that has never, ever been created in a machine: imagination, instinct, creativity. You are simply wrong in assuming this is easy, or even possible by only hardware and technology.
I would bet 100 euros that even 10 years from now, a completely mechanical bot, would be able to beat in a bo7 any top programer in the top 10
All this being said, i am not against bots, and believe there is a lot of meaning and education, and fun to those who are making bots. I am just pointing out their limitations, as a computer program.
|
Bonyth at least writes decent English, FyRe_DragOn is canadian.
|
|
|
|