Where is the future of this bot going?

[emerysong]

i am constantly seeing misplay posts(ai problems) or people having ideas about "bettering" this bot and im seeing very minimal responses from developers or the community

my questions are these

1: where is the future of this bot going? will it take years for this to change for the better?
2: why isnt there a community edited version of this bot ( i dont see any NON dev people being able to work or submitt thier work on this that much like in honorbuddy or demonbuddy ...hearthbuddy almost seems like a secret)
3: are people activly working on this bot or is this bot on the bottom of the totem pole? (i feel that it is with the very low responses from people, devs , and the community activity on these fourms)
4: whats the future of the AI? i see many people talking about other bots AI working alot better
5: why cant we have other emotes mid game? seems like if you can greet at the start it shouldnt be to hard to have it say other things



whole reason i bought this bot was to get gold chars of each class (which im very close to getting now after 1 month,,,, i have played since day 1 manually) and it has done that very well atleast


I understand these questions are more so my personal thoughts and pref. but to have a more in depth awnser to some / all and other questions would be gratly helpful in people understanding where this bot is headed
this is by no means a bash on devs i apprecate your guys hard work on all the buddy products i just think if you arent activly working on this why not let the community work on it

i havent researched the fourms that much to see if there is currently stuff in the very near future coming out ( big changes to the AI and such.) and if i have missed anything or hurt anyones feelings im sorry in advance this is ment only for informative purposes

 

[Tony]

i see no reports for misplays from you,so if you have any issues use support issues section providing your log files

those bugs getting reported,getting fixed as well

 

[emerysong]

as i stated in the first sentence of my post (i hope you read and didnt just auto respond) i said im SEEING alot of misplay posts. not that im the one doing the postings...

and giving a canned response about giving logs doesnt help further my questions asked.

the middle section of my first post says most of what i needed this bot for, im just asking these GENERAL questions so the community can be better informed

nothing about my questions requires any logs or section hopping because my questions asked are general questions about the bot and these forums have so little traffic they almost need to be in the front section to get the most views/opinions

and please keep in mind the last line of my original post

Quote: "i havent researched the fourms that much to see if there is currently stuff in the very near future coming out ( big changes to the AI and such.) and if i have missed anything or hurt anyones feelings im sorry in advance this is ment only for informative purposes " end quote

 

[Tony]

1)it changes everyday


2)thats something the community must answer


3)yes they do


4)getting improved every day according our tests and user reports


5)cause we dont think that will help in any way

there


tho all the above are already answered and stated on forums

 

[Percy]

You actually could modify the behavior of the bot yourself if you know some C# but it’s true that requested features like midgame emotes, a better report function for misplays, improvement of the mouse movement algorithm, etc. get rejected by the team. It should be possible to program Plugins for the midgame emotes and a report function but you have to find someone who writes them. The team, at least a part of it, seems to have a clear image how the bot should evolve: Improve the design (current beta), fix misplays, keep the bot stable but don’t add new features or improve the play speed as that could risk the stability. But there are also Team members who improve the bot by writing new Routines for deck types and send quick fixes for misplays before they get rolled out with major updates. Lazy cat is a good example for a community member who writes Code for the Project. He also had planned a program for midgame emotes but dropped it because of unknown reasons.

This is my opinion and results out of experience with the community and Bot.

 

[DaSoul]

There should be something like https://www.thebuddyforum.com/hearthbuddy-forum/232286-hsbtoolbox.html directly integrated into Hearthbuddy that tracks/logs what the bot plays.

After a game, the tracked/logged data should be prepared in an easy way, for example a gui, that lists the moves that have been done with radion buttons for "correct move" "wrong move" and a text field for comments.

This report should be automatically send to the hearthbuddy devs, they analyze it and make necessary changes.


Or, another idea:

A tool that logs the users manual play style and learns from it. E.g. moves, card combos and mulligan. Another tool collects all this player data on a server and finds similar moves.

Over time, the analysis tool learns different moves and play styles and knows, which cards to combine.

After x times of the same move, it will be integrated into the bot logic.

Initial development effort would be high, but the learning curve/improvement rate would be exponentionally higher than now.

 

[ThyFate]

2) About 2 years ago, personally chose not to do anything custom for this bot because of how unfriendly the API was, there were a lot of restrictions based on "we don't think it's necessary". (But that was 2 years ago, so things might have changed).

There should be something like https://www.thebuddyforum.com/hearthbuddy-forum/232286-hsbtoolbox.html directly integrated into Hearthbuddy that tracks/logs what the bot plays.
After a game, the tracked/logged data should be prepared in an easy way, for example a gui, that lists the moves that have been done with radion buttons for "correct move" "wrong move" and a text field for comments.
This report should be automatically send to the hearthbuddy devs, they analyze it and make necessary changes.
Or, another idea:
A tool that logs the users manual play style and learns from it. E.g. moves, card combos and mulligan. Another tool collects all this player data on a server and finds similar moves.
Over time, the analysis tool learns different moves and play styles and knows, which cards to combine.

After x times of the same move, it will be integrated into the bot logic.
Initial development effort would be high, but the learning curve/improvement rate would be exponentionally higher than now.

Cards like Piloted Shredder, Servant of Yogg Saaron, Yogg himself, Ragnaros , knife juggler and other high variance cards will ruin that learning curve very quickly. Complexity of handling exceptions related to those variances is a lot higher than the reward this will yield.

 

[DaSoul]

2) About 2 years ago, personally chose not to do anything custom for this bot because of how unfriendly the API was, there were a lot of restrictions based on "we don't think it's necessary". (But that was 2 years ago, so things might have changed).


Cards like Piloted Shredder, Servant of Yogg Saaron, Yogg himself, Ragnaros , knife juggler and other high variance cards will ruin that learning curve very quickly. Complexity of handling exceptions related to those variances is a lot higher than the reward this will yield.

But that's just a bunch of cards.

I mean just instead of putting in every possible move by hand. Directly learning from human players would increase the AI a lot.

 

[ThyFate]

But that's just a bunch of cards.

I mean just instead of putting in every possible move by hand. Directly learning from human players would increase the AI a lot.

Assuming players know how to play optimally. (which, in my experience, most of botters do not)

 

[DaSoul]

Assuming players know how to play optimally. (which, in my experience, most of botters do not)

Sure not 100% perfect, but still way better than current logic.

 

[Lazy cat]

I agree with Phelon. Buying a bot we should get a working tool that does its job.
And Hearthbuddy provides users the easiest way (of all bots that I've seen): install the bot and pressed the Start button - what could be easier? (for 95% of users this is enough).
The remaining 5% - it's enthusiasts who explore the bot, change settings, watch replays.

Thanks to enthusiasts who like to watch the game (please note, it is "LIKE" and not "MUST"), as well as with new tools that make this search simpler and faster, it became possible to quickly find mistakes and send them to support. This allowed us to gets not just a "faceless" updates, but specific solutions to our problems.

So what about the future you shouldn't worry. Better to worry about your own nerves and let the bot to do all the dirty work for you

 

[ThyFate]

And Hearthbuddy provides users the easiest way (of all bots that I've seen): install the bot and pressed the Start button - what could be easier? (for 95% of users this is enough).

Your research is outdated, I am 75% sure that all 4 bots are click start and go do your thing.

 

[Lazy cat]

click start and go do your thing.

Yes, all the programs work this way (but in this case I meant something more).
Everyone is free to make their own choice and have his own opinion.

 

[ThyFate]

My 2 cents from point of view of a bot that is not mentioned. In order to improve AI both, buddy and ranger devs need to realize that Silverfish AI is outdated and structurally bad for all new cards post TGT. Minor patches to "improve" it, is a terrible philosophy as it's the same as fixing broken glass with a ducktape. Sure the glass holds now, but the glass now looks like crap and has areas that you can't see through. There is no reason for a bot not to be able to hit legend in Hearthstone since it only requires 51%+ winrate and robotic stamina (hello bots). I am not trying to say it is easy, but with Silverfish AI (which is garbage, I will say that) it will be hard, or even impossible (hence lack of legends on both above mentioned bots).

 

[sepefeet]

My 2 cents from point of view of a bot that is not mentioned. In order to improve AI both, buddy and ranger devs need to realize that Silverfish AI is outdated and structurally bad for all new cards post TGT. Minor patches to "improve" it, is a terrible philosophy as it's the same as fixing broken glass with a ducktape. Sure the glass holds now, but the glass now looks like crap and has areas that you can't see through. There is no reason for a bot not to be able to hit legend in Hearthstone since it only requires 51%+ winrate and robotic stamina (hello bots). I am not trying to say it is easy, but with Silverfish AI (which is garbage, I will say that) it will be hard, or even impossible (hence lack of legends on both above mentioned bots).

You might want to learn your history about SB, it uses a branch of Silverfish too.

 

[ThyFate]

You might want to learn your history about SB, it uses a branch of Silverfish too.

I've been with it for 2 years now. I know it's history, the good, the worst and the grey. It has no traces of silverfish for at least 2 years now (since it was hаcked and posted on nulled at least). I know that because I did my research on every bot and their limits and I currently have my own profile extension of wirms default AI. If you prove me wrong with evidence, I will apologize, admit that I am wrong and leave with shame.

 

[sepefeet]

SF has the ability to be ran in a client/server manor although it has been neglected for a long time... I don't care to waste time finding what little evidence there may be in his client just to win a pointless argument.

 

[SIL3N7]

I've been with it for 2 years now. I know it's history, the good, the worst and the grey. It has no traces of silverfish for at least 2 years now (since it was hаcked and posted on nulled at least). I know that because I did my research on every bot and their limits and I currently have my own profile extension of wirms default AI. If you prove me wrong with evidence, I will apologize, admit that I am wrong and leave with shame.

Bot maker last post was within 2 years and his last update looks like it was 6 months ago was part of his testing team

 

[tozededao]

Well community isn't going to help if devs don't show interest on the bot either. This bot was never the best HS bot on the market and devs apparently are fine with it. I remember the old days where devs were actually worried about being the best bot on the market and actually had better goals and QA.

 

[CalifRHCP]

Well community isn't going to help if devs don't show interest on the bot either. This bot was never the best HS bot on the market and devs apparently are fine with it. I remember the old days where devs were actually worried about being the best bot on the market and actually had better goals and QA.

There isn't much interest in HS bots because you can't really make any money off Hearthstone, so a lot of the bot base aren't interested.