Uke isn't AI. He's more like a punch bag.
There is something in Training where you face Uke who fights back but he is coded to do a certain move so that's not really A.I.
This is pretty much what multiplayer is but I am not sure how coding Uke to have some sort of intelligence for every mod will work.
I still support this though.
There is something in Training where you face Uke who fights back but he is coded to do a certain move so that's not really A.I.
This is pretty much what multiplayer is but I am not sure how coding Uke to have some sort of intelligence for every mod will work.
I still support this though.
blue
pink
pink
I mean, everyone keeps saying it's hard af, but in all honesty.
People are making bots in League of legends, and that's more highly advanced than toribash.
If you look at it this way, making a script that can copy down moves from online play, openers from different mods, or anything in that general aspect. Have it take down who frame, what joints, what position, and have it lined up with what that person did next. Or have it set to have uke learn from what you just did last game. If you did one move, it puts that move into it's memory bank, letting it use it at random next fight.
The only thing you'd have to do is something that takes in fact how close ukes joints are from touching the ground to be able to 'save' ones self while trying to continue the fight.
And from seeing what awesome things people are creating, this really isn't that hard of a step forward. I mean, we have something that can record joints and use them at will when we want them. Why not modify that into something that can record and put it into a code used for uke's moveset?
I don't think saying how hard it will be, but if someone actually sits down and tries to do it, it could be done quite easy. I mean, we've had/i've seen people that use multiplayer bots. That's no different than getting uke to fight against you. It's actually easier. You only have to worry about 1 side, instead of both blue/red.
Pretty sure if someone said, "I'd pay 100,000tc for someone to make this" it'd be done in no time. It's just no ones inspired to do something that requires a little bit of effort for no payout.
People are making bots in League of legends, and that's more highly advanced than toribash.
If you look at it this way, making a script that can copy down moves from online play, openers from different mods, or anything in that general aspect. Have it take down who frame, what joints, what position, and have it lined up with what that person did next. Or have it set to have uke learn from what you just did last game. If you did one move, it puts that move into it's memory bank, letting it use it at random next fight.
The only thing you'd have to do is something that takes in fact how close ukes joints are from touching the ground to be able to 'save' ones self while trying to continue the fight.
And from seeing what awesome things people are creating, this really isn't that hard of a step forward. I mean, we have something that can record joints and use them at will when we want them. Why not modify that into something that can record and put it into a code used for uke's moveset?
I don't think saying how hard it will be, but if someone actually sits down and tries to do it, it could be done quite easy. I mean, we've had/i've seen people that use multiplayer bots. That's no different than getting uke to fight against you. It's actually easier. You only have to worry about 1 side, instead of both blue/red.
Pretty sure if someone said, "I'd pay 100,000tc for someone to make this" it'd be done in no time. It's just no ones inspired to do something that requires a little bit of effort for no payout.
Originally Posted by Hacks
I mean, everyone keeps saying it's hard af, but in all honesty.
People are making bots in League of legends, and that's more highly advanced than toribash.
If you look at it this way, making a script that can copy down moves from online play, openers from different mods, or anything in that general aspect. Have it take down who frame, what joints, what position, and have it lined up with what that person did next. Or have it set to have uke learn from what you just did last game. If you did one move, it puts that move into it's memory bank, letting it use it at random next fight.
The only thing you'd have to do is something that takes in fact how close ukes joints are from touching the ground to be able to 'save' ones self while trying to continue the fight.
And from seeing what awesome things people are creating, this really isn't that hard of a step forward. I mean, we have something that can record joints and use them at will when we want them. Why not modify that into something that can record and put it into a code used for uke's moveset?
I don't think saying how hard it will be, but if someone actually sits down and tries to do it, it could be done quite easy. I mean, we've had/i've seen people that use multiplayer bots. That's no different than getting uke to fight against you. It's actually easier. You only have to worry about 1 side, instead of both blue/red.
Pretty sure if someone said, "I'd pay 100,000tc for someone to make this" it'd be done in no time. It's just no ones inspired to do something that requires a little bit of effort for no payout.
no it's not harder
being able to make a bot that moves according to another entities position and performs an action once said entity is within range is not that hard
Making a bot that reacts to ENDLESS outcomes and predicts when they'll hit an object and reacting to how your opponent moves is much harder
even if A.I is implemented it won't be topnotch, a white belt probably wouldn't have a problem beating it
