Forums Tyrant

Auto Farm Deck, Mission 89...

172 posts

Flag Post

So the HD4 parts in this post doesn’t matter anymore, because Eva + II x 10 rapes mission 89 upside down on auto.

My Results-
Sample space: 100
Losses: 2

Friend’s Results-
Sample space: 100
Losses: 4

Theoretical losing scenarios:
(a)
Early Mystic Obelisk → Heartripper. Instant kill to up to 3 IIs.
Loss rate (From both samples) as result of scenario: 0
(Super unlikely, heartripper nearly always drops before mystic is up, or commander is dead because the units did not tank enough)

(b)
Ridiculous regen streaks as a result of diesel and carrion retriever
Loss rate as a result of scenario: 4 (one for me, 3 for my friend)
(Depends on how unlucky you are, really. It takes 2 flurries for diesel to kill commander, so if you really are that down in luck…)

©
Really, really unlucky irradiate drop
Loss rate as a result of scenario: 1
Irradiate hits the enfeebled II, weakens another to zero damage that is in front of an enemy unit (not sure which card it was)
I had one loss as a result of this, but extremely unlikely?

(d)
Loss out of some other thing I did not see while I was autoing: 1

(e)
Heartripper first turn has a chance of killing the enfeebled II, but usually drops from poison and giving you field control again.
Loss out of this scenario: 0

Combination of some of the above scenarios can be a source of your defeats as well. Ridiculous, or even slight regen streaks + a sad irradiate/heartripper drop CAN possibly screw up your game, but the results do show that this deck is by far the best auto farm deck so far.

THE FOLLOWING IS NOW POINTLESS AND OBSCURE. DO NOT BOTHER READING ON!


THE CHALLENGE

  • Build a HD4 auto gold/exp farm deck that has a success rate above 90%, if possible, 95%
  • That does not take unique, commonly used deck cards, to reduce the need of having to take out cards from your deck all the time.

    This means, none of the decks in either of these two threads (Search function):
    http://www.kongregate.com/forums/65-tyrant/topics/155530-reputation-and-gold-farm-endgame?page=1
    http://www.kongregate.com/forums/65-tyrant/topics/158674-best-deck-for-farming-homeland-defenders-4?page=1

    My current decks, prototypes, and previously “failed” decks:


    Best:(THIS IS THE RECOMMENDED DECK FOR FARMING HD4 ATM)


    74 / 80

    Analysis:
    Loses under following conditions:
  • Plague Duster drops against Sand Crawler or Colossus
  • Fury Walker lead procs flurry multiple times with warehouse support
  • Early bulldozer drop that dodges activations

    Reason for drop:
    Severely depends on the daily drop algorithm of some sort… (people have a problem with this statement, so whatever) It suddenly failed 3 times in a row, which meant something was wrong. Friend who is testing it at the same time as me gave exact same feedback on the same day.

    Old Prototypes:

  • Tried to give hunter a purpose
  • Same weakness as plague duster, had to drop.
  • First test of Ryoko instead of Dalia

    3/4 (small sample, dropped the moment I got a loss)



    Prototype to check if moar nightstalkers is better

    24 / 27
  • Lost 2 fights in a row to plague vs colossi starts. I’m tired of the dusters, I’m going to find less luck reliant alternatives.




    19/23
  • fails to early havoc + warehouse

    I will require y’all to help test and give feedback on these decks ;)
    I understand not everybody has so many of my favourite awesome uncommons, but I will try and find alternatives.
 
Flag Post

Reserved

Not sure why image links are failing, and formatting failed without using br tags.

 
Flag Post

what about something like this?

Thadious
5 devastator
5 irradiated infantry
(just uncommons and commons.)

current results:
W:39 L:8 (it seems like the main problem is not the counter, but the ‘heal all’ ability of the commander. when i lose i just cant kill the units of the AI fast enough.)

I play raids now, so i cant playtest it too much. but i just won 2 games with it. maybe just luck :D I’ll edit this post with more results later.

 
Flag Post

Devastators have the same problem with plague, they will suicide themselves on any sand crawlers/collossi, so it won’t work well.

They’d have to rely on warehouses to cancel their attack :P

 
Flag Post

Malort, 7 plague dusters (lol gold packs), 3 dominated hatchlings. Only got this many dusters/DHs recently. 95%ish on auto I think, occasional AI stupidity against CWs can mess it up. I don’t really mark wins/losses but it feels very solid. Fewer dusters/DHs work too, the extras are just insurance.

(11W/0L so far, I’ll just mark wins/losses until I get to 50 or get bored, and update here)

 
Flag Post

Sheesh what a waste of time, energy and gold. Those 80 tests of yours have taken quite a shitload of time to do, and 80 permutations of a deck versus 80 permutations of another deck (both generated randomly) gets you a pretty random win %.

As you said "…suddenly failed 3 times in a row, " What a frikken surprise that the result from a random take from some billions of permutations is random. Don’t worry though, there is no “daily drop algorithms” – do you seriously think they would even bother to create something like that? Random is random (seemingly or otherwise), sometimes it’s bite-you-in-the-ass-random and sometimes lets-your-shitty-deck-win-10-times-in-a-row-random.

With the number of permutations, getting a reliable statistics of your win% will take the rest of your life, good luck with that (even a crappier deck can get lucky – or rather more lucky than average – 10 or even 100 times in a row).

 
Flag Post

To fix images, simply replace the -’s with asterisks or some other symbol.

 
Flag Post

Nice thread… I think 74/80 on auto is pretty decent, and I do find it hard to imagine a 100% deck, since there just is too much which can randomly go wrong…

Also, ignore piemee – yes, you won’t be able to get a fix on an exact win rate, but it seems like the question is very useful.

Having not looked at the recently, my only question would be whether you would consider a devastator, which against raiders is basically the functional equivalent of tiamat

 
Flag Post

algorithm for card drops isnt fully random. and nope, im not just saying that without proof.

 
Flag Post
Originally posted by kaboomba:

algorithms for card drops isnt fully random. you’ll realise it changes day by day, if you try talking to people who have saved up gold to drop on a large number of packs at once.

please stfu if you dont know what ure talking about.

Oh poor kanooba… Considering you wouldn’t even understand the source code if you saw it, you should really be really, really quiet about speaking of them. Idiots see patterns where there is only coincidence / series in RNG – which are to be expected once enough numbers are pulled out. Just because the morons who happen to hit on these series group up hailing some “zomg! daily drop algorithm” doesn’t mean it’s there. I bet people who have been hit by lightning could group up and come up with something just as silly – “Zomg we got hit by lightning the same day, it’s a frikken conspiracy11oneoneeleven there’s an algorithm in the clouds 8)”.

Leaving that aside, drawing a line from pack contents to Auto’s card order is even more silly, he might as well say that the suns position was in the wrong place – thus the AI played in a bad order.

Feel free to prove your little conspiracy theory on the algorithm, until you do, “please stfu if you dont know what ure talking about.”

 
Flag Post

well, i’ll admit that drawing a link between them is silly. they may be completely unrelated.

by the way, try talking to anybody who has bought more than 30-40 packs at the same time. please stfu if you dont know what ure talking about.

and edit was done before you replied, anyway. oh well. keep painting every argument as a conspiracy theory. you seem to like doing that.

 
Flag Post
Originally posted by kaboomba:

well, i’ll admit that drawing a link between them is silly. they may be completely unrelated.

by the way, try talking to anybody who has bought more than 30-40 packs at the same time. please stfu if you dont know what ure talking about.

and edit was done before you replied, anyway. oh well. keep painting every argument as a conspiracy theory. you seem to like doing that.

Yes I do, they are much more fun to bash at then >8)

But yes, if you get a sequential series of numbers from a binary RNG, you will more likely end up hitting a series of some sorts since the whole thing is nothing but a seed-based pre-determined series that distributes evenly only after getting enough samples, which hangs somewhere around 3-6 times the amount of different samples one can get for most – including the flash – RNG’s. It’s no mystery that pulling 30-40 random groups sequentially from an RNG ends up giving you a bad distribution, it’s a curse of binary seed based RNG’s, not some intentional mean algorithm meant to bash people.

I’ll believe the algorithm theory once someone buys 1000 packs in a single day and the distribution is not relatively even. Until then I’ll keep throwing a 20-sided dice 10 (or 20, even 30) times and whining how the dice is rigged since it many-a-times gives me the same number multiple times during those 10 throws instead of the ’ mysteriously expected even division of numbers’.

 
Flag Post
Originally posted by kaboomba:

well, i’ll admit that drawing a link between them is silly. they may be completely unrelated.

by the way, try talking to anybody who has bought more than 30-40 packs at the same time. please stfu if you dont know what ure talking about.

and edit was done before you replied, anyway. oh well. keep painting every argument as a conspiracy theory. you seem to like doing that.

Seriously, this kind of thinking just highlights an utter lack of understanding of coding (and, for that matter, coders. I don’t know any who would bother to code a big conspiracy-theory-algorithm rather than just type rand(111) or similar).

As has been said, RNGs are generally flawed but please try to understand that just because you think you see a pattern and your buddy thinks he see a pattern, doesn’t mean there is one. If you want to educate yourself, go read up on RNGs, wikipedia would be a good place for you to start.

 
Flag Post

Actually, i dont think the particular RNG of card drops is meant to bash people. Neither do i think its a negative conspiracy. Neither do i think coders have nothing to do but to bash people. Neither do i have an “utter lack of understanding of coding and coders”. Neither do I think my evidence is completely conclusive. I guess thats your entire argument.

Assume much? Maybe if you werent so busy trying to be hostile u’d learn something. Just saying.

I actually think its entirely probable theres a random function for card drops generated by the game every x hours or so, as a hidden dimension for the game, for people find out and have fun with it. Despite what you apparently think about coders, (i dont know, that they’re lazy and cant be bothered?) hidden mechanics are actually common in many games.

 
Flag Post

I think its entirely possible theres a random function for card drops generated by the game every x hours or so, for people find out and have fun with it.

Despite the fact that implementing it this way would serve no purpose (which coders loathe), be unecessarily complex (which coders loathe), and waste time coding (which coders, and their managers, really loathe).

Doesn’t seem very logical, does it.

Maybe if you werent so busy trying to be hostile u’d learn something. Just saying.

You have nothing to teach me. Just saying.

 
Flag Post

Actually, hidden mechanics serve a purpose in many games. They generate interest, and sometimes, a community that examines different mechanics and data that lasts a long time. It is not necessary that such mechanics are hidden deeply. Many times, it is simply that certain mechanics are not disclosed.

Its actually entirely logical that coders want to encourage such things. Not frequently, of course. Much of coding does involve streamlining.

A very simple example would be elitistjerks and World of Warcraft. As of a few years ago, anyway. but pretty sure the community is still going strong.

And, about the teaching part, believe whatever u want, pompous ass. lol. Believing the sky is purple doesnt make it so.

 
Flag Post
Originally posted by piemee:

Sheesh what a waste of time, energy and gold. Those 80 tests of yours have taken quite a shitload of time to do, and 80 permutations of a deck versus 80 permutations of another deck (both generated randomly) gets you a pretty random win %.

As you said "…suddenly failed 3 times in a row, " What a frikken surprise that the result from a random take from some billions of permutations is random. Don’t worry though, there is no “daily drop algorithms” – do you seriously think they would even bother to create something like that? Random is random (seemingly or otherwise), sometimes it’s bite-you-in-the-ass-random and sometimes lets-your-shitty-deck-win-10-times-in-a-row-random.

With the number of permutations, getting a reliable statistics of your win% will take the rest of your life, good luck with that (even a crappier deck can get lucky – or rather more lucky than average – 10 or even 100 times in a row).

sigh

Why are you looking at billions of permutations, when the sample size listed here is 80? And you certainly don’t have to work thru all 1.3E13 combinations to find a statistically relevant win percentage.

With 80 test runs for a particular deck, I wouldn’t call it “a pretty random win %”. While it is likely not the theoretical maximum win percentage, there is a certain level of certainty achieved. I haven’t taken the time (and won’t) to calculate the level of certainty in this case, but I am willing to guarantee that the test results are greater than 50% certain – which is what random results would be.

It would be good to keep in mind that you can normally answer a question without doing every possible trial.

 
Flag Post
Originally posted by kaboomba:

Actually, hidden mechanics serve a purpose in many games. They generate interest, and sometimes, a community that examines different mechanics and data that lasts a long time. It is not necessary that such mechanics are hidden deeply. Many times, it is simply that certain mechanics are not disclosed.

Its actually entirely logical that coders want to encourage such things. Not frequently, of course. Much of coding does involve streamlining.

A very simple example would be elitistjerks and World of Warcraft. As of a few years ago, anyway. but pretty sure the community is still going strong.

And, about the teaching part, believe whatever u want, pompous ass. lol. Believing the sky is purple doesnt make it so.

I am very familiar with elitistjerks and the difference between what you are suggesting and what they put so much effort into is that they have the following:

Actual facts on the mechanics, divulged by developers.

In-depth statistical analysis of a relevant sample.

And finally, a policy of not posting ANYTHING without hard evidence. If you aired your type of baseless opinion and tried to present it as fact on EJ, as you have on these forums, you would likely be infracted for it, and definitely ridiculed.

Call me pompous if you like, but out of us there is one who knows something about coding games and statistical analysis with scientific rigour, and as your comments continue to show, it’s not you.

 
Flag Post

actually i got 2 players to buy 30-40 packs at the same time. and they roughly got 7-9 emps each. The fact that its a small sample size, explains why i didnt post a thread about it, presenting it as a fact to the community. thats why it was mentioned in passing. This is what is known as a ‘preliminary conclusion’. And is in fact, the first step in how statistical tests are designed. And this is also why this wasnt posted in an elitistjerks forum. Last i heard, tyrant wasnt so hardcore.

Its true i didnt seem to state it in this fashion. Well, i must draw the parallel with the grammarnazi – picking small errors and creating entire posts of insulting and contemptuous language in which everything is ridiculed, many assumptions are made, and otherwise incomprehensible flaming is indulged.

Due to the difficulty of accumulating so much gold, sample size is obviously low. I have some evidence, and put it forth. Together with the reasonable unlikelihood of gathering more samples. You have none whatsoever.

All you have is a horrendous personality, and inability to admit the possible validity of contrary inferences drawn. Oh, and a reflexive urge to resort to name-calling, straw man attacks, and basically invalid argumentation. If you want to educate yourself, go read up on argumentation. wikipedia would be a good place for you to start.

 
Flag Post

I hope you guys are discussing something good with all that tldr, otherwise you’re just wasting each other’s time. The deck listed above is 11 wins 0 losses so far, and I’ll be a bit busy with raids today/tomorrow so I won’t be doing more testing, but I think it’s something people should aim for. Plague dusters and malort are definitely key, and the more dusters, the merrier. Some kind of massive DH rush has potential as well imo, though a cannon walker or perhaps posideon for siege should probably be added as well, to get rid of the annoying warehouses.

 
Flag Post
Originally posted by kaboomba:

actually i got 2 players to buy 30-40 packs at the same time. and they roughly got 7-9 emps each. The fact that its a small sample size, explains why i didnt post a thread about it, presenting it as a fact to the community. thats why it was mentioned in passing.

Due to the difficulty of accumulating so much gold, sample size is obviously low. I have some evidence, and put it forth. Together with the reasonable unlikelihood of gathering more samples. You have none whatsoever.

All you have is a horrendous personality, and inability to admit the possible validity of contrary inferences drawn.

Your ‘evidence’, if it even exists, does not constitute evidence by any credible standard. You yourself admit that the sample size is too small, therefore you too have no evidence whatsoever. The difference is that my position is neutral; I have made no claims so I require no evidence. You on the other hand have failed to establish any rational basis for your position.

You claim that you did not present this as fact, but you can scroll up to see your claim that you have proof. If you have proof of something, then that thing is a fact, or what you have is not proof.

Carry on attacking my personality (as if you were in a position to make a judgement on that), it serves to highlight that you have no real argument to make in your favour. All I have done is call you out on this.

 
Flag Post

No. i have admitted that my argument should not have been stated as fact.

you have called me out on this, made many unjustified assumptions, repeatedly insulted and flamed and ridiculed me without provocation.

heres a tip, as to the concept you are displaying a sore lack of. its called arguing in good faith. if you were the recipient of any kind of higher education, you would know that any argumentation, hypothesis, or any piece of writing, takes place in good faith. This is so as to expedite analysis and make it more useful. Try googling it, the wiki definition is not a complete one as to its definition.

Not doing so, renders any attempt to take the moral high ground, laughable.

 
Flag Post

There’s a possiblity (no im not claiming this to be definitely true) that the random function used here, like in many other places is based on time elapsed since Date X. So.. at certain times of a day you should be able to get similar pack results. I remember a guy drawing 8 EMPs from 60(?) packs because he drew them all at the same time. (So I’m guessing he must have had a lot of repeats of other rares too, that aren’t as relevant/cool as EMP :P ) Otherwise nice discussion but little less flaming pl0x.

 
Flag Post
Originally posted by kaboomba:

No. i have admitted that my argument should not have been stated as fact.

you have called me out on this, made many unjustified assumptions, repeatedly insulted and flamed and ridiculed me without provocation.

heres a tip, as to the concept you are displaying a sore lack of. its called arguing in good faith. if you were the recipient of any kind of higher education, you would know that any argumentation, hypothesis, or any piece of writing, takes place in good faith. Try googling it, the wiki definition is not a complete one as to its definition.

Not doing so, renders any attempt to take the moral high ground, laughable.

Glad to see that you have now abandoned your argument entirely in favour of the personal criticism.

 
Flag Post

the moment u abandoned your assumptions, and straw man attacks, starting to actually attack the argument – i did admit that.

if u’d attempted less to flame, and more to read, u’d see that. a lack of imagination and perspective on your part is the true casualty of our conversation