Just wanting to know what i should be able to expect this to do when it gets a little more "polished", and what i should expect it to never be able to do (not for a long time at least)?
Honestly? I feel the hardest thing for a computer to do is to determine that board X is better than board Y. For instance, me having a 7/7 out, and, you having nothing out, is, obviously better for me. But, if you have a 7/7 out that gives +1 to all mobs, and, I have a 9/9 that doesn't, who has it better? That's hard to say. That's just a very basic one, there are much more advanced things, such as RND, etc... I feel it'll probably get to a point where it can maybe win ~ 50% of the time at rank 18 or so, and, that's where it'll probably die down. The *only* way I see to advance from that would be to make an amazing algorithm to actually be able to state "This board state is better than this other board state" (Which, will be insanely hard on it's own), then, to loop through all possible moves you can possibly make this turn, and, calculate which board is the absolute best for me. >I did the math here for how long it'd take, but, re-reading it, it was super boring. It ended up taking ~ 20 seconds, assuming you had a setup that could process 1 board per MS per CPU cpre (I.E. in 2MS, it has to generate two sequential boards (Following the rules of Hearthstone to get there), rate them, and, return the highest scoring board), and, that you had six cores that had absolutely no overhead. This also assumed that you were playing in two destined categories "Hand" mode and "Board" mode, where, you can't combo anything from your hand with your board, and, visa-versa, it, finally, also assumed that there were ten cards on the board, five yours, five the opponents, and, you had five cards in your hand.
Expect basic logic and a 50% win rate. Good for leveling your characters and earning a trickle of gold. Not good for ranked. Works pretty well with minions. Not so well for spells, buffs, etc.
wouldnt cost 20€ if it would be capable of doing more just look at poker bots and what a good poker bot costs my expectation from HB: buy it to let it grind gold on a 50% win ratio on unranked, i dont except it to do more
The team said there shooting for a 60-85% win rate. I would love it to be 85% that crazy. but will take 60 and above
I'm already getting that with HRengine using the warlock rush deck they posted on their forums. I left it on over night and it went 51-52. I'm expecting this bot to be very capable if it matches the quality of the other bots on this site. 60-85% win rate seems quite ambitious though. Hopefully they can get it to average 7 wins in arena, even if I have to make the deck manually.
We're working on the "basics" right now. We're working on some cool stuff to make the bot far, far more intelligent. (If things work as we expect, we may need to even tone it down a bit) We'll see how things go. For now, we're trying to work out all the little quirks that we need to handle for general botting, before we can really dive deep into the AI. At the same time, we're slowly working on a public API for people to use. We're just trying to figure out how to expose it in a way that's "easy" to use.
funny thing is I had 8/0 for the first time today. 8 games, 8 wins (casual) but then the 9th game desynched or froze or something and I had to kill the bot. I am liking it a lot though.
The buddy team, "WHEN COMMITTED" to a project can pull of some fantastic bots.. The exile buddy at a core is brilliant, can be customized to do just about anything. Deamon buddy was great when they where commited and honorbuddy is still imo number one for wow bots.. With that being said.. If the wrong devs get on this project and the good ones get pulled of to work on other bots, then there wont be much hope for it.. I currently own this bot, But im using HRengine atm as its logic destroys this bot. What i would like to to is this bots CORE with HRengines logic, becuase i know one day im going to get banned using HRengine due to the way it works.
As Apoc is on it i am fully confident they will push this way beyond awesome. What most guys don't realize, and that basically makes me sad and angry the same time, that this is still B E T A. Reading comments it really feels like people totally overlook this fact. No one (buddyteam) ever stated, that the current available range of functionality is anywhere near the limit, NO ONE. For the people without any clue, wait till we have some logic behind this and then come and complain again how bad the bot is. You are arround for almost two years, based on experience you should know better... Totally wrong aproach. Great leaders in history never won wars, based on copying others strategies. They defined them themselves.
I was speaking to it's current capabilities, not the long term goals. relax dude. take a look at my posts you can see I am a big supporter of this bot and have been spending a lot of time trying to assist in making it better.
