A.I. RNG hates me


#1

I know this has been brought up a lot and I have seen a lot of people say that it is not true after running tests but I am 100% positive that the A.I. cheats.

You can especially tell with troops like Black Mangles and Dracos. I try using Black Manlges all the time but always remember why I don’t stick with it, I NEVER devour an enemy. Yet, I am sure plenty can attest that when facing it in PVP, it works every other time, usually on it’s first cast. I like using Dracos with mana generators for it’s trait but I only actually destroy enemies 1 out of every 8 casts, which is better than my percentage with BM, but nothing close to the A.I. which seems to happen almost every time. I don’t face many of them but when I do, I can count on getting destroyed.

Then to Kraken, I get devoured by Kraken on it’s first cast 9/10 times. Sure I also get a decent amount of devours with him, about 25 - 40% of the time but I can expect to be devoured with every first cast when facing in PVP or GW.

I am not bringing this up so that this can be fixed (would be nice) but just wanted to vent and see if anyone still insists that this is not skewed in the A.I.(s) favor.


#2

I will respectfully disagree and comment no further


#3

Have you actually tallied the numbers with pen and paper? The other day I was convinced that Wisps hit the 3rd slot more often than anything else. Then I ran some tests against a 4x Wisp defence team and actually wrote down how often each slot was targeted and couldn’t see any bias (it still looked like there was some bias for the first 20 or 30 casts, but after that it started levelling out). Write down the results of 100 or 200 Black Manacles/Dracos casts, both from you and the AI, and see if the percentages are still way off. And don’t forget about devour immunities when counting the Black Manacles. Eyeballing probabilities is very inaccurate, especially since you’re more likely to remember the annoying instances than the pleasant ones.


#4

I disagree, and would provide my own statistics, but I’ve been known to substitute gut feel for statistics 46-48% of the time, so I will refrain.


#6

I could see writing it down if it was close, but it’s as close as 90% to 10%, not close. I must be the unlucky one the A.I. takes it out on


#7

I agree with you to an extent. I used the infamous beast team for a week in PvP and got three devours. I don’t know how many casts but I tend to get between 200-300 trophies. So a fair amount of battles and a fair amount of casts.
However someone else will have experienced the complete opposite and will devour most of the time.

I don’t use troops that have a chance to do things any more. Because for me, they don’t have a chance. But others, as I mentioned, will have an entirely different experience.


#8

I had the same feeling about the wisps. It felt like 75%-90% on the third slot, when it should have been 25%. Your intuition about these probabilities can be way off.


#9

I don’t think that this fits with what I am saying. The card does not state “20% chance to Charm the third enemy”, that was just a feeling of yours while I understand you say it is just a feeling of mine that their percentages are off, but over a year and a half, these destroys and devours have never evened out.


#10

The Wisp charms a random enemy. I would assume that that means 25% chance for each slot. Regardless of what is supposed to mean, it felt like 75-90% and turned out to be 25% after testing, so I don’t see the difference to your case.


#11

Yeah that’s true


#12

Here’s my problem with these cards and why I understand your frustration completely @Robbg1986:

  1. Regardless of the % be it 10 or 50, it is a dice roll EVERY TIME. This means that the computer IS capable of succeeding 4 times in a row on a 10% proc, while I AM capable of failing 4 times in a row on a 50% proc.

  2. Over time, the numbers balance out and there is no TRUE favoritism/ cheat/ etc., but generally speaking we as players are not interested in the spread over 1000 matches we are just concerned with the 6 casts during this battle, ie the NOW.

  3. Despite 2 and because of 1, I just try to avoid any card that ONLY does something on a % chance. It keeps me sane, satisfied, and really pissed at Mab when she freezes me out of control of the board! :rage:

I hope that was helpful.

tl;dr- I feel the OPs pain, but understand it is recall bias. :wink:


#13

After a bit of delving into salt,seed,RNG…

Recall bias is a very dismissive term and I don’t believe it is applicable to all cases.


#14

I know people throw around words like “all the times” or “not recall bias”… etc.
Still… honestly tho, if your win-lose record in the battle log isn’t something like 20% win or less, then I really think the problem is just recall bias.


#15

Echoing the suggestion to start tracking and counting instances of the spell. If something is up, it’ll show up in your numbers. If you wind up seeing roughly the right probability, you can rest easy knowing that either (a) things are exactly as they should be, or (b) the Universe knew you were testing and skewed the RNG.

On second thought, maybe that second explanation isn’t cause to rest easy.

Edit: I am going to toot my own horn, here.


This is what I am talking about. In the time since that thread went up, my tribute rate has remained constant – and below expectation. But because I am tracking everything, I know it’s low and don’t have to rely on infamously shaky human memory.


#16

The system is known to go on streaks, if you feel the game is really against you, it might actually be (although not as much as you think it is). Either play 10 minutes of explore, soul farm or just simply take a break and let it get unstuck on rolling values against your favour. Then hop back in and it should work kind of fairly.


#17

Does playing other forms of the game somehow reset your rolls? First I’ve heard of it.


#18

Pretty sure it’s a misguided application of the Gambler’s Fallacy. By getting the “bad rolls” out of the way in Explore mode, you get the “good rolls” when you return to PVP/Guild Wars.


#19

My Kraken is full. He eats troops on the regular…


#20

Why should Wisps not hit the third spot more often than others? It’s a good spot to hit. I find that AI uses Wisps quite smartly, alas.

…and anyway RNGsus hates everybody.


#21

This is a very good point. To the extent the AI is at all intelligent, it should be prioritizing Charming the 2nd and 3rd spots.