I think it'd be more interesting to build a deck for the A.I to use, sort of like Rafaam, and practice against that. That way Blizz doesn't necessarily have to work on teaching the bots each and every deck's play style. The downside is that the bots would probably suck at playing their decks.
To make a bot that can effectively play anything outside of an aggro deck would take a HUGE amount of work. I've written some computer AI for simple games and it can be time consuming for even simple games. I can't even imagine the amount of time and man power it would take to create a passable bot for anything outside of an aggro deck.
It doesn't only work for simple aggro but for any deck that has a simple set of rules to follow without particularly caring (or in the case of AI knowing or properly guessing or anticipating) the opponent's reactions and future moves. Take for example three classic decks with an explosive and usually overwhelming midrange: In the case of secret paladin, as explained so many times from players and experts (for example Firebat), the main thing to do is simply curving out with the game's most overpowered cards. In the case of zoo and mage mechs, simply playing as many minions as possible while applying constant pressure with trades and face (plus an extra rule to keep board presence within a certain range to limit the AoE loss). In the case of combo druid, playing sticky minions and removal until you can one-turn-kill your opponent. I haven't gone into the coding myself, but there are details of botting online if you look.
Really, you can't take a loss in casual? Are you sure you could stomach losing to AI?
Playing versus the AI, as explained many times, is easier from a psychological standpoint because you don't feel embarrassed by misplays. I've witnessed it many times myself. Just google "ladder anxiety" - it also applies to casual and many other games besides HS.
To make a bot that can effectively play anything outside of an aggro deck would take a HUGE amount of work. I've written some computer AI for simple games and it can be time consuming for even simple games. I can't even imagine the amount of time and man power it would take to create a passable bot for anything outside of an aggro deck.
It doesn't only work for simple aggro but for any deck that has a simple set of rules to follow without particularly caring (or in the case of AI knowing or properly guessing or anticipating) the opponent's reactions and future moves. Take for example three classic decks with an explosive and usually overwhelming midrange: In the case of secret paladin, as explained so many times from players and experts (for example Firebat), the main thing to do is simply curving out with the game's most overpowered cards. In the case of zoo and mage mechs, simply playing as many minions as possible while applying constant pressure with trades and face (plus an extra rule to keep board presence within a certain range to limit the AoE loss). In the case of combo druid, playing sticky minions and removal until you can one-turn-kill your opponent. I haven't gone into the coding myself, but there are details of botting online if you look.
You do realize that, with the exception of combo druid, all the decks you mentioned could arguably be considered aggro decks right? I've never seen a bot pilot a non-aggro style deck effectively, and the reason why is because the amount of coding it would take to get a bot to play on a level anywhere close to a human for a non aggro-style deck would be tremendous.
The AI and crappy tutorial is probably why so many players are so bad at this game. And I do put the fault with Blizzard on this one, because it's actually pretty ridiculous how little this game teaches players about the game.
To make a bot that can effectively play anything outside of an aggro deck would take a HUGE amount of work. I've written some computer AI for simple games and it can be time consuming for even simple games. I can't even imagine the amount of time and man power it would take to create a passable bot for anything outside of an aggro deck.
It doesn't only work for simple aggro but for any deck that has a simple set of rules to follow without particularly caring (or in the case of AI knowing or properly guessing or anticipating) the opponent's reactions and future moves. Take for example three classic decks with an explosive and usually overwhelming midrange: In the case of secret paladin, as explained so many times from players and experts (for example Firebat), the main thing to do is simply curving out with the game's most overpowered cards. In the case of zoo and mage mechs, simply playing as many minions as possible while applying constant pressure with trades and face (plus an extra rule to keep board presence within a certain range to limit the AoE loss). In the case of combo druid, playing sticky minions and removal until you can one-turn-kill your opponent. I haven't gone into the coding myself, but there are details of botting online if you look.
You do realize that, with the exception of combo druid, all the decks you mentioned could arguably be considered aggro decks right? I've never seen a bot pilot a non-aggro style deck effectively, and the reason why is because the amount of coding it would take to get a bot to play on a level anywhere close to a human for a non aggro-style deck would be tremendous.
Beyond a certain point it's just semantics, so it doesn't really matter if we call them "aggro" or "midrange", although if you have followed HS you would know there is a clear distinction between them - and these actually are midrange and not aggro. In any case, I am telling you again that as far as I know - I have never used any bots, never will, and have never done particular research on them - these are only some of the decks that people have used to succesfully bot up in ladder up to legend, due to the aforementioned reasons.
You do realize that, with the exception of combo druid, all the decks you mentioned could arguably be considered aggro decks right? I've never seen a bot pilot a non-aggro style deck effectively, and the reason why is because the amount of coding it would take to get a bot to play on a level anywhere close to a human for a non aggro-style deck would be tremendous.
Beyond a certain point it's just pointless semantics, so it doesn't really matter if we call them "aggro" or "midrange", although if you have followed HS you would know there is a clear distinction between them - and these actually are midrange and not aggro. In any case, I am telling you again that as far as I know - I have never used any bots, never will, and have never done particular research on them - these are only some of the decks that people have used to succesfully bot up in ladder up to legend, due to the aforementioned reasons.
The secret paladins and mech mages that bots have been able to play effectively are their aggro variants. Not all secret paladins and mech mages are mid-range (and yes, I know the difference). Bots aren't the greatest at deciding when to trade, so they are typically set to go face and only trade if they have to, which means they can only run aggro style decks effectively. I've seen the code for a Hearthstone bot so I know a little bit about it.
I haven't gone into the coding myself, but there are details of botting online if you look.
I would suggest you to write a flowchart. Not one of these oversimplified one you see with memes, but a relatively realistic one that you can play the game and make "decisions" with. See how good you can play using your chart, how large your flowchart become, and how often do you get stuck because you haven't properly written a decision point for it.
Rollback Post to RevisionRollBack
People who refuses to play aggro out of principle are even worse than people who play exclusively aggro.
One should seek to become a complete player and play all archetypes, including ones that he despises for whatever irrational reasons.
I haven't gone into the coding myself, but there are details of botting online if you look.
I would suggest you to write a flowchart. Not one of these oversimplified one you see with memes, but a relatively realistic one that you can play the game and make "decisions" with. See how good you can play using your chart, how large your flowchart become, and how often do you get stuck because you haven't properly written a decision point for it.
I have been coding since I got my first computer (an Acorn Electron in 1984), and I have done doctoral research on neural network programming, so I know very well what is involved in this. Besides that, memes have virtually nothing to do with what I think on this or any other matter.
The AI and crappy tutorial is probably why so many players are so bad at this game. And I do put the fault with Blizzard on this one, because it's actually pretty ridiculous how little this game teaches players about the game.
I have been coding since I got my first computer (an Acorn Electron in 1984), and I have done doctoral research on neural network programming, so I know very well what is involved in this. Besides that, memes have virtually nothing to do with what I think on this or any other matter.
I don't think anyone is arguing it would be difficult to program, just time consuming. The resources would just be better spent elsewhere
I have been coding since I got my first computer (an Acorn Electron in 1984), and I have done doctoral research on neural network programming, so I know very well what is involved in this. Besides that, memes have virtually nothing to do with what I think on this or any other matter.
I don't think anyone is arguing it would be difficult to program, just time consuming. The resources would just be better spent elsewhere
I don't think anyone needs an AI that can reach legend and play with no misplays at all. So you wanted a flowchart: if you can trade a minion into more valuable one, do it.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
I think it'd be more interesting to build a deck for the A.I to use, sort of like Rafaam, and practice against that. That way Blizz doesn't necessarily have to work on teaching the bots each and every deck's play style. The downside is that the bots would probably suck at playing their decks.
I hope OP isn't one of those qqers whocry about meta decks in casual....
People who refuses to play aggro out of principle are even worse than people who play exclusively aggro.
One should seek to become a complete player and play all archetypes, including ones that he despises for whatever irrational reasons.