This is why the bot cannot get a constant win rate of +50%

[hangeland]

Hello guys and welcome to this thread explaining why it is not possible for the current state of hearthbuddy to accomplish a win rate of 50% or above and is constant. There is two factors that make the bot impossible to achieve this constant win rate.
The first is the fact that whenever three games in a row have been won, one gets additional stars.
The second is the fact that the bot will reach a rank at some point, where it cannot get below, due to stupid moves (no offence to the creators and so) and tougher opponents.
With a good deck and setup, it would surely be possible to peak rank 17 or maybe lower - this does not change anything. The bot is running for, say, a week and remain a win rate of +/-50% and keep bouncing from rank 20, 19 and 18. For this to happen, the bot should only get very very very few win streaks in order to maintain this position between rank 20-18 and keep the +/-50% win rate. Too many win streaks would give less than 50% win rate, because we would have to loose 2 games each time we won 1 game to maintain rank position.
Looking at the odds of getting a win streak will make it seem unlikely that we would not get more than a few win streaks in a week, when the win rate is 50%. Therefore, assume that the bot reaches rank 18, where the bot seems to have a real struggle to win the games. Now, let's say that we had a winstreak before in the previous matches and need 2 stars to get promoted to rank 17. The match the bot will play is won by the bot either by the opponent drawing bad cards, plays it bad or something like that. Now, where the bot is promoted to rank 17 it is highly unlikely for it to win any games and therefore it will surely loose a lot. The first match is lost and it is promoted to rank 18, now it will not win any games at this level due to the players also being to hard so it will loose another 3-4 stars (don't exactly remember how many stars is needed at rank 18). But because we only won 3 games to get to rank 17 with 1 star, but it took us 4-5 losses to get back to rank 19. This would result in a win rate below 50% each time the bot had a win streak as we would climb the ladder more and more and eventually meet the wall at some point.

I hope this makes sense to you, if not please ask to specific parts of the text such that I can elaborate further.

Best regards.

 

[DaSoul]

So if my hearthbuddy shows ~50% winrate after 16 hours, it's a lie?

 

[hangeland]

So if my hearthbuddy shows ~50% winrate after 16 hours, it's a lie?

No, it is not a lie. It is possible to have ~50% win rate after 16 hours. This is for the people stating that they have a deck and a setup yielding +50% winrate over a longer period of time, which is not possible atm.

 

[chrisngoftw]

No, it is not a lie. It is possible to have ~50% win rate after 16 hours. This is for the people stating that they have a deck and a setup yielding +50% winrate over a longer period of time, which is not possible atm.

Of course, it is possible - but only up to a certain rank.
To be more precise, accurate titles would say something like "~50% win rate up to rank 15" (for example)

People are using the bot to grind golden portraits, levels, and gold.
They shouldn't expect to climb the ladder with it.

 

[DarkOS]

correct. and honestly, if someone is able to put together a deck, let the bot run for 16 hours and maintain a +50% i would be surprised. i dont claim to have all the best card and deck configs, but i have watched how the bot responds and plays as i do work (letting run in a vm on a second monitor - its strangely relaxing...) it makes weird decisions that are costly. I have never seen my bot go better than level 17 at ANY time of day or day of the week.

Again, i dont claim to know the best deck configurations but i would be interested to see what deck you put together to achieve a +50% win rate over 16 hours (which by the way means you would continue to rank up levels because you are winning more than losing).

 

[hangeland]

Of course, it is possible - but only up to a certain rank.
To be more precise, accurate titles would say something like "~50% win rate up to rank 15" (for example)

People are using the bot to grind golden portraits, levels, and gold.
They shouldn't expect to climb the ladder with it.

You did not understand what the thread says - without being flippant. The statement of yours saying "Of course, it is possible - but only up to a certain rank." No it is not, read what I wrote above and try to understand, why it is not possible.

 

[hangeland]

correct. and honestly, if someone is able to put together a deck, let the bot run for 16 hours and maintain a +50% i would be surprised. i dont claim to have all the best card and deck configs, but i have watched how the bot responds and plays as i do work (letting run in a vm on a second monitor - its strangely relaxing...) it makes weird decisions that are costly. I have never seen my bot go better than level 17 at ANY time of day or day of the week.

Again, i dont claim to know the best deck configurations but i would be interested to see what deck you put together to achieve a +50% win rate over 16 hours (which by the way means you would continue to rank up levels because you are winning more than losing).

You got the message and yeah the bot makes the weirdest decisions. + you added something that i forgot to make clear in the thread: "which by the way means you would continue to rank up levels because you are winning more than losing".
However, the bot is a better alternative than not playing

 

[Mario27]

So if my hearthbuddy shows ~50% winrate after 16 hours, it's a lie?

you have alot of people that concede in the beginning to to let our bot crash with the R6025 error they know how to do it but i think i wil fix all those error soon with a regcleaner to repair all dlls from the runtime Libary sometimes they concede and it doesend crash but alot of times it does

 

[DaSoul]

you have alot of people that concede in the beginning to to let our bot crash with the R6025 error they know how to do it but i think i wil fix all those error soon with a regcleaner to repair all dlls from the runtime Libary sometimes they concede and it doesend crash but alot of times it does

sry to say it that rude, but that is bullshit

 

[dinfar]

To be honest you have some huge flaws in your logic, but we'll try to sort that out.

You can have a 50%+ win rate due to:
- Not all people are playing ranked. Some people just play normal to do daily quests. Result: You can't use the lattersystem in your argument.
- Some people play ranked by them selves (while the bot play sometimes). Result: You can't use the lattersystem in your argument.
- Some people concede until they reach the highest (read: baddest) level, in order to get matched with newbies. Heck, even normal players do it: Why the (constructed) reward system is broken : hearthstone Result: You can't use the lattersystem in your argument.

I get your point about the 50%, but it simply isn't true. You could ask for people to standardize their runs before they run the bot.

 

[Endus]

I get a consistent 50% win rate with Murloc decks (not even using the API that's currently being developed). Usually I'll climb up to rank 16 or so and then plummet back down to 20. It usually evens out in the end.

Paladin Murloc

Warrior Murloc

 

[hangeland]

To be honest you have some huge flaws in your logic, but we'll try to sort that out.

You can have a 50%+ win rate due to:
- Not all people are playing ranked. Some people just play normal to do daily quests. Result: You can't use the lattersystem in your argument.
- Some people play ranked by them selves (while the bot play sometimes). Result: You can't use the lattersystem in your argument.
- Some people concede until they reach the highest (read: baddest) level, in order to get matched with newbies. Heck, even normal players do it: Why the (constructed) reward system is broken : hearthstone Result: You can't use the lattersystem in your argument.

I get your point about the 50%, but it simply isn't true. You could ask for people to standardize their runs before they run the bot.

First of all, thanks for the reply. It is well structured and easy to understand. However, I do only agree in one argument, which is:
"- Not all people are playing ranked. Some people just play normal to do daily quests. Result: You can't use the lattersystem in your argument." This is true, so I forgot to say that this only counts for ranked.
"- Some people play ranked by them selves (while the bot play sometimes). Result: You can't use the lattersystem in your argument." I am not sure, whether I understand this correctly, so a case could be that one stops the bot and takes over the control? But this would lead to a manipulated win rate, since the bot did not play all the way, hence the win rate is a combination of the bot and the player, which is merely useless.
"- Some people concede until they reach the highest (read: baddest) level, in order to get matched with newbies. Heck, even normal players do it: Why the (constructed) reward system is broken : hearthstone Result: You can't use the lattersystem in your argument." The bot would not run continuously and this would also lead to a manipulated win rate. One should never interrupt the bot, when recording a win rate, because this would influence the win rate and in order for other users to achieve the same win rate, the same interaction should be made approximately, which is not recorded.

 

[eagleeye1]

I am running one version of Kripp's Paladin deck with the bot. I have played over 250 games with it. My win percentage every 40-50 games is around 45-46%. While not 50% or higher, I am very pleased with this deck and the bot's results.

I suspect if the bot made less dumb decisions, like not playing the hero power when Redemption was up, or not missing lethal sometimes, or prioritizing Starving Buzzards or other minions that can be Argus'd on later turns, the win rate would be much higher than 46%.

I suspect that some decks could eventually be played as well as any human player too. We have chess programs that can beat humans - this is no different. With enough skilled logic, anything is possible.

 

[Endus]

Someone asked me to post the decks I mentioned earlier. Here's the Paladin one. As you can see it has quite a few Legendaries in it which I'm sure attribute to my win rate.

I already deleted the Warrior deck (after I finished the portrait) but it was basically the same, just substituting the class cards for the Warrior weapons/charge minions.

 

[JTPace]

The bot queues up to play games for 8 hours straight, and over those 8 hours every single enemy it faces disconnects on turn one, and every game is a win, 100% win rate over 8 hours.

It isn't likely, but it is possible.

If you want to play the hypothetical game I can too, I think you need to re-word your argument.