Mahidhar
Member
A lot of us(or even everyone) played Rokkit Ball, way before Fallen Empire: Legions was released. The element, that I ever felt in that game, to be really intriguing, would be, how the game's AI worked. Now, before you guys start to say "Wha...? I've seen better. Do you live under a rock?" stuff, let me remind you that Rokkit Ball, being a browser game, was a marvel, and it's AI's definitely something, to give credit to. I also really liked the way, human players would be replaced by AI players and vice versa, as soon as they leave/join the game, which prevented it to come to any sudden halts(which usually happens in Legions' servers).
There's also the way, these AI players responded to player's quick chat commands(whoever thought of that, is a genius in my book). You ask them to pass them the ball, and they do, even if they're right about to score.
I was thinking if we can apply the same to Legions, although I do understand, it will be a lot more complicated than in Rokkit Ball. Here are a few ways I thought of, which may give us a good start in integrating this feature.
Teaching AI, the different positions in the game: I believe, except capping, all other positions are fairly easy to take up, for the AI players. Still, I believe we can get them to use basic routes on the maps we got. Some of the problems we might be facing with the AI player roles are; constant updates to the game, like weapons, classes, armor, maps etc may prevent the AI players to play at a satisfactory level. But, this will only happen if the game stays in development for all eternity, and not get finalized at all.
Positioning the AI player in the game: The main reason, some/most/all of you would say that this isn't practical, is because of the chance(and a good one too) that the AI may not stay at /do what you expect them to do, and may feel useless. For example, take the front route on Zenith Cauldron(either side is fine). You, sometimes, spawn on the front bridge, but to start with a good speed, you fly to one of the towers, and jump from there. So, if a Cap-bot(AI Capper) spawns on the front bridge, he'll simply start from there, which results in him, traveling at a much slower pace. This is where the AI player positioning comes up. We could rectify the problem of AI positioning, by developing a certain command parameter. What I have in mind is, you type a command, then shoot a rocket somewhere in your base, like the top of a tower, and your AI player makes that a "rest" spot(just as the flag stand, is the "rest" spot for a HoF). That player could be your AI LD, ready to snipe or call out incoming enemies(explanation, later), block or damage them. Therefore, his performance will depend on how you place the bots on your base.
The positions of LO bots might also be done in a similar manner as above(maybe not the typing command, and shooting rocket at the spot). It might also be possible that, no positioning of LO-bots(lobots
) would be necessary, as playing LO mostly requires you to be at a higher spot, than your enemies, and distract them. Therefore, the Bots may simply be programmed to stay at a higher spot and shoot, give the top priority to killing the enemy capper, and/or e-grabbing. Different commands for different jobs.
Quick-chat as a communication link between Human and AI players: The above mentioned problems and hassles may be solved/reduced with this. From basic things like, getting the AI to pass the flag, to get them to respawn and change position, quick chat is the way to go. So, first of all, all bots in the game may be numbered like Bot1, Bot2 etc. Therefore, you'll have your quick-chat menu like this: Quick chat menu -> Commands(exists along with all the taunts, compliments, and other sub-menus) -> Bot1, Bot2 etc. Then you select the Bot you wanna give the command to, and all the commands show up. Bot 1 -> Respawn, Switch Position: HoF, Switch Position: LO etc. This'll be quite difficult, as there may be many commands to give, but the clutter can be reduced by creating sub-menus where needed. You might say; "This is a fast paced game, and we can't stop at a place, read all the commands and then, after a while, select one". True, but we all had that problem, when we first used quick chat. It's all a matter of getting used to.
A problem here that one might encounter is, to whose command will the AI respond to. There will indefinitely be A-holes, who wanna mess up your game, i.e. mess with your bots. A solution could be; only one person gets to command the bots. Although, this may become a problem, as controlling multiple bots may become an issue. Maybe integrating yet another command like; accept commands from player: XXXXX. The bots will then, respond to the other person, who has just been given the control. There might be situations where the Bot will get the same command from both the players, then what?(if Bot has the flag, and both players ask for a pass, for example) Simple, they'll work on a Admin>Mod basis. The second player given the command will be a second priority, for the Bots to accept commands from.
But all that may only happen in private servers. In public servers, the Bots may have their own minds, where commands turn into requests. An A-hole/noob player might ask for a flag pass, from half way across the map. Bot will deny. Bot is about to die(programmed to pass when near 50% health), Bot notices 3 players, where 2 of them ask for the flag. Player at 25m did not ask for the flag, player at 50 and 100 m asked for the flag. Bot calculates energy needed to throw to the farthest person, if energy not available, Bot switches to the player at 50 m, and so on. How the bot chooses between P1(human player) at 100m and P2 at 50m, to throw the flag, where P2 is an A-hole who just doesn't want P1 to get the flag: Bot senses enemy presence around both players, bot measures the height at which P1 and P2 stand, Bot measures health difference between P1 and P2(I've noticed how A-holes rarely have full health, when they pull stuff like that), Bot throws flag to the safest and the healthiest. Even if the A-hole did win, in a situation like this, it's probably for the best, as it would be P1's fault for not being at a good spot/free of enemies/healthy.
Bots may also decide stuff like that, based on the player's in-game accomplishments, like grabs, caps or assists. We can think of many ways, of how the bots would work , but that requires a lot of time to create and test.
That's pretty much all I had in mind. I understand if this is absolutely impossible to achieve, even by the slightest. Just wanted to dump this from my head, or my head would've exploded(I actually feel much relaxed now). Also, while typing this post, I forgot a few things I wanted to mention, but I may remember and post them as you guys comment, so don't flame me, if the post seems incomplete and vague.
There's also the way, these AI players responded to player's quick chat commands(whoever thought of that, is a genius in my book). You ask them to pass them the ball, and they do, even if they're right about to score.
I was thinking if we can apply the same to Legions, although I do understand, it will be a lot more complicated than in Rokkit Ball. Here are a few ways I thought of, which may give us a good start in integrating this feature.
Teaching AI, the different positions in the game: I believe, except capping, all other positions are fairly easy to take up, for the AI players. Still, I believe we can get them to use basic routes on the maps we got. Some of the problems we might be facing with the AI player roles are; constant updates to the game, like weapons, classes, armor, maps etc may prevent the AI players to play at a satisfactory level. But, this will only happen if the game stays in development for all eternity, and not get finalized at all.
Positioning the AI player in the game: The main reason, some/most/all of you would say that this isn't practical, is because of the chance(and a good one too) that the AI may not stay at /do what you expect them to do, and may feel useless. For example, take the front route on Zenith Cauldron(either side is fine). You, sometimes, spawn on the front bridge, but to start with a good speed, you fly to one of the towers, and jump from there. So, if a Cap-bot(AI Capper) spawns on the front bridge, he'll simply start from there, which results in him, traveling at a much slower pace. This is where the AI player positioning comes up. We could rectify the problem of AI positioning, by developing a certain command parameter. What I have in mind is, you type a command, then shoot a rocket somewhere in your base, like the top of a tower, and your AI player makes that a "rest" spot(just as the flag stand, is the "rest" spot for a HoF). That player could be your AI LD, ready to snipe or call out incoming enemies(explanation, later), block or damage them. Therefore, his performance will depend on how you place the bots on your base.
The positions of LO bots might also be done in a similar manner as above(maybe not the typing command, and shooting rocket at the spot). It might also be possible that, no positioning of LO-bots(lobots
) would be necessary, as playing LO mostly requires you to be at a higher spot, than your enemies, and distract them. Therefore, the Bots may simply be programmed to stay at a higher spot and shoot, give the top priority to killing the enemy capper, and/or e-grabbing. Different commands for different jobs.Quick-chat as a communication link between Human and AI players: The above mentioned problems and hassles may be solved/reduced with this. From basic things like, getting the AI to pass the flag, to get them to respawn and change position, quick chat is the way to go. So, first of all, all bots in the game may be numbered like Bot1, Bot2 etc. Therefore, you'll have your quick-chat menu like this: Quick chat menu -> Commands(exists along with all the taunts, compliments, and other sub-menus) -> Bot1, Bot2 etc. Then you select the Bot you wanna give the command to, and all the commands show up. Bot 1 -> Respawn, Switch Position: HoF, Switch Position: LO etc. This'll be quite difficult, as there may be many commands to give, but the clutter can be reduced by creating sub-menus where needed. You might say; "This is a fast paced game, and we can't stop at a place, read all the commands and then, after a while, select one". True, but we all had that problem, when we first used quick chat. It's all a matter of getting used to.
A problem here that one might encounter is, to whose command will the AI respond to. There will indefinitely be A-holes, who wanna mess up your game, i.e. mess with your bots. A solution could be; only one person gets to command the bots. Although, this may become a problem, as controlling multiple bots may become an issue. Maybe integrating yet another command like; accept commands from player: XXXXX. The bots will then, respond to the other person, who has just been given the control. There might be situations where the Bot will get the same command from both the players, then what?(if Bot has the flag, and both players ask for a pass, for example) Simple, they'll work on a Admin>Mod basis. The second player given the command will be a second priority, for the Bots to accept commands from.
But all that may only happen in private servers. In public servers, the Bots may have their own minds, where commands turn into requests. An A-hole/noob player might ask for a flag pass, from half way across the map. Bot will deny. Bot is about to die(programmed to pass when near 50% health), Bot notices 3 players, where 2 of them ask for the flag. Player at 25m did not ask for the flag, player at 50 and 100 m asked for the flag. Bot calculates energy needed to throw to the farthest person, if energy not available, Bot switches to the player at 50 m, and so on. How the bot chooses between P1(human player) at 100m and P2 at 50m, to throw the flag, where P2 is an A-hole who just doesn't want P1 to get the flag: Bot senses enemy presence around both players, bot measures the height at which P1 and P2 stand, Bot measures health difference between P1 and P2(I've noticed how A-holes rarely have full health, when they pull stuff like that), Bot throws flag to the safest and the healthiest. Even if the A-hole did win, in a situation like this, it's probably for the best, as it would be P1's fault for not being at a good spot/free of enemies/healthy.
Bots may also decide stuff like that, based on the player's in-game accomplishments, like grabs, caps or assists. We can think of many ways, of how the bots would work , but that requires a lot of time to create and test.
That's pretty much all I had in mind. I understand if this is absolutely impossible to achieve, even by the slightest. Just wanted to dump this from my head, or my head would've exploded(I actually feel much relaxed now). Also, while typing this post, I forgot a few things I wanted to mention, but I may remember and post them as you guys comment, so don't flame me, if the post seems incomplete and vague.
