Project Legend by OtisRight

[branmuffin]

Arena is where you build a deck out of cards they give you, then play against people and earn rewards. But you're always guaranteed a pack. It's 150G for an arena and 100G for a pack. So if you win a few games you're pretty likely to break even.

When I first started I bottled all my classes to 15/20 against AI, whichever is the level where you unlock the last basic card for the class. Because you don't have access to a bunch of cards in custom decks(what you'll use all the time after you unlock the cards) that you need until you get a class to that level

So yeah just focus on getting your classes leveled up against AI for now using the basic decks, then move up.
And make sure to watch the bot to learn to play, because this phase of basic decks is meant to teach new TCG players the basics. If you've never played a TCG before then it is probably more beneficial to play them yourself and start hitting after.

There's a good resource I like to use on Hearthpwn called Innkeeper, it tracks your card inventory and lets you sort the decks people post by how many cards you need for it. So you can look at only decks you have the cards for.

 

[Lita]

Albeit slightly off-topic, appreciate the constructive help you provided, branmuffin!
I shall do as you have pointed out and will start utilizing Innkeeper on Hearthpwn.

 

[glmillz]

\
@glmillz
Maybe bad luck, maybe it's your deck..hard to say. AfaIct you'll start facing the same metadecks with legendarys around rank 15- the only difference is people make less mistakes the higher you get.
\

I actually ended up going from rank 18-7 today no problem with an Aggro Jade Shaman deck using your aggro routine. Thanks much!

 

[elyanea]

mid range shaman > r16 to r9 > 27games, 23wins. will come back later for more feedbacks

 

[OtisRight]

@DoctorGGCUSA
Arena is currently not supported. I focus on standard ranked mode.

@Lita
As branmuffin said, there are many ways to get gold. To make profit in arena you need 7 wins on average. I wouldn't try to bot it, because almost none of those wild cards are implemented(or only partly). There are also some hidden quests: Hidden Hearthstone Quest List | BlizzPro's Hearthstone Try to focus on making one competitive deck(aggro decks are cheaper). Disenchant everything you don't need for it. It will be much easier to farm your daily 30 wins on ladder.

Also try to think like this: You buy a new game, what amout would you usually pay for it? Like 40 bucks, 50...maybe more? If you like that game, why not support the developers? Spending some money on packs would give you a big headstart. Of course you can grind arena until you get your cards but it'll take a loooong time, AND you need to be really good to make it worth the time.


@everyone
Thanks for the feedback guys.
Here is a poll like last month, please all vote: Highest Rank in December? - Straw Poll

Next update will be out on the 2nd of january. I wish you all a great start into the new year!

 

[bobdu40]

rank 4 on my second account, with pirate dragon warrior, plcontrol routine

 

[Iso]

To make profit in arena you need 7 wins on average.

Huh? 7-3 means free packs. 4-3 is a profit.

 

[glmillz]

Do you think that hearthbuddy will ever be able to support a Reno deck Otis? is it worth even diving into the code?

 

[OtisRight]

Pretty good results last month!
Highest Rank in December? - Results - Straw Poll
Over 55% hit at least rank 5, three people rank 1-3 and two hit legend. Awesome! Let's push this bot some more

@glmillz
Control/Reno/Combo decks are much harder to support then midrange/aggro. In control decks you have a lot of reactive cards and some value cards. You basically want to outvalue(or outcard) your oponnent. You need to save up your stuff for the best value and think many turns ahead to be sucessful. Unless somebody bots like 500+ games and points out every misplay, I'm not gonna bother. Sorry.

Update10.1!
Download latest update

Best decks to bot with right now(02.01.17):
(depending on the real oponnents and your rank ofc)

1. Pirate Warrior (the new one got me to rank 10 already!)
2. Mid Shaman (jade version recommended)
3. Dragon Warrior/Aggro Shaman
4. Dragon Priest
5. Jade Druid

changelog:

• added/updated some mulligan rules
• removed old decks and updated some existing one's Click
• never play Ship's Cannon after pirates have been played(force to play it first)
• increased penality to use Jade Lightning on small/not important minions
• don't play Naga Corsair after attacking with the hero(force to play it first)
• fixed a bug witch prevented the bot to play Thunder Bluff Valiant sometimes
• added a priority list for Netherspite Historian discover(similar to Sir Finley and Dark Peddler lists)
  (the priority numbers can be changed in BehaviorPLControl.cs )
• raised the numbers of current priorityMinions so the bot doesn't ignore them
• reworked penality for The Coin:
  • don't use on turn one if we have nothing playable next turn
    (doesn't apply to Totem Golem because it's op on turn 1)
• never attack face with the last weapon charge unless we have another weapon in hand - unless lethal
  (save it for upgrade stuff/pirate synergies/minion removal etc.)

 

[iLevi]

[*]never attack face with the last weapon charge unless we have another weapon in hand
(save it for upgrade stuff/pirate synergies/minion removal etc.)

I assume that it still hits face if that means getting one turn closer to lethal? Or if it means lethal?

 

[OtisRight]

I assume that it still hits face if that means getting one turn closer to lethal? Or if it means lethal?

How does it get you one turn closer? Does attacking this turn vs next turn change anything if you don't have another weapon anyway? Or do you mean that you may draw another weapon? Let's say you are in midgame and have a 2/15 chance to draw another weapon. So 13/15 times it is correct to hold the weapon. If it's lethal it's always allowed, afaik the ai will ignore all penalities when we have lethal, that's why I didn't mention it.

 

[Casutherland]

Pretty good results last month!
Highest Rank in December? - Results - Straw Poll
Over 55% hit at least rank 5, three people rank 1-3 and two hit legend. Awesome! Let's push this bot some more

@glmillz
Control/Reno/Combo decks are much harder to support then midrange/aggro. In control decks you have a lot of reactive cards and some value cards. You basically want to outvalue(or outcard) your oponnent. You need to save up your stuff for the best value and think many turns ahead to be sucessful. Unless somebody bots like 500+ games and points out every misplay, I'm not gonna bother. Sorry.

Update10.1!
Download latest update

Best decks to bot with right now(02.01.17):
(depending on the real oponnents and your rank ofc)

1. Pirate Warrior (the new one got me to rank 10 already!)
2. Mid Shaman (jade version recommended)
3. Dragon Warrior/Aggro Shaman
4. Dragon Priest
5. Jade Druid

changelog:

• added/updated some mulligan rules
• removed old decks and updated some existing one's Click
• never play Ship's Cannon after pirates have been played(force to play it first)
• increased penality to use Jade Lightning on small/not important minions
• don't play Naga Corsair after attacking with the hero(force to play it first)
• fixed a bug witch prevented the bot to play Thunder Bluff Valiant sometimes
• added a priority list for Netherspite Historian discover(similar to Sir Finley and Dark Peddler lists)
  (the priority numbers can be changed in BehaviorPLControl.cs )
• raised the numbers of current priorityMinions so the bot doesn't ignore them
• reworked penality for The Coin:
  • don't use on turn one if we have nothing playable next turn
    (doesn't apply to Totem Golem because it's op on turn 1)
• never attack face with the last weapon charge unless we have another weapon in hand - unless lethal
  (save it for upgrade stuff/pirate synergies/minion removal etc.)

Can you post your deck list?

 

[derdj]

[*]added a priority list for Netherspite Historian discover(similar to Sir Finley and Dark Peddler lists)
(the priority numbers can be changed in BehaviorPLControl.cs )

Thank you Looked into the code. We could make this process generic to decrease code duplication. But it's just a nice feature and kinda low priority

 

[sunocean]

Can you post your deck list?

he had done this, just click "Click" you can find his decks.

 

[sunocean]

happy new year for china, thx for your works.

 

[derdj]

bad behvaiour: Healing before drawing

Most of the time when the bot is kinda low hp he heals himself as the first action in turn with hero power, before drawing cards, as priest. In this case it costs the game because auf Corruptor battlecry lethal

 

[derdj]

On a side note:
I added a discover mechanic for Drakonid Operative to always pick the highest mana card available. Works like a charme and produces pretty good value for now. Testing this further...
Also I extended the PLControl-routine to count turns. So I am able to distinguish between turn numbers in the Netherspite Historian discover mechanics. I inserted a second Dictionary with priority on low-cost dragons to curve out better before turn 5 and kept the other list if it is turn5+. Testing this atm together with the drakonid discover. If this works we could tweak discoveries turn-based, which is pretty awesome

 

[derdj]

Playing Dragonfire Potion before attacking with non-dragon minion in own turn 9.

 

[OtisRight]

Thx for the logs, will check them soon!

On a side note:
I added a discover mechanic for Drakonid Operative to always pick the highest mana card available. Works like a charme and produces pretty good value for now. Testing this further...
Also I extended the PLControl-routine to count turns. So I am able to distinguish between turn numbers in the Netherspite Historian discover mechanics. I inserted a second Dictionary with priority on low-cost dragons to curve out better before turn 5 and kept the other list if it is turn5+. Testing this atm together with the drakonid discover. If this works we could tweak discoveries turn-based, which is pretty awesome

I thought about something similar like that but there are some problems:
The cards you get are always different, some high mana cards are much better then others(e.g.: twising nether vs ragnaros). Another problem is, the bot doesn't know what it's playing against. Maybe it's best to pick an armor smith(cost 2 mana) vs pirate warrior to get more health or a removal spell to kill that frothing etc.

For counting turns you can use some existing stuff like p.turnCounter or p.ownMaxMana. Would be nice if you can pm me your code

 

[sunocean]

hello buddy, i used the deck arrr. these is something not well. i have Grimy Gadgeteer ,korkron Elite and heroic strike in hand , the best is put Gadgeteer fist and it can buff Korkron . but it put korkron first. after that i change my language to English and record log for whole night . hope to help you.
View attachment Hearthbuddy 8904 2017-01-04 01.46.txt
View attachment log.zip