Warhammer: Vermintide 2

Warhammer: Vermintide 2

View Stats:
Valaya's Brew Does Not Feel Like 50%
"Using your Career Skill has a 50% chance of granting you the effect of the potion you are currently carrying."

I feel like this must have some sort of hidden mechanic, like an internal cooldown or something, because it does not feel like it triggers half the time you use your ult, at all.

I will be compiling list of recorded runs of CW runs with this boon here.
Video Exhibit A - 9 procs out of 40, credit to Xaphnir
https://youtu.be/ZSBhH89brk0

https://steamcommunity.com/app/552500/discussions/0/3133918821804237624/?ctp=7#c3133919455870478891


PRIMER for the topic (READ BEFORE RESPOND)
How does Valaya's Brew work?

Based on looking at the lua source script, it is essentially modelling a DISCRETE RANDOM VARIABLE to determine when the skill should work.

Why does it use a model for discrete random variable?

Because it's calling math.random(1, 2) [1], and "math.random() generates pseudo-random numbers uniformly distributed" [2]. It's discrete because the parameters supplied to this function is math.random(1,2) and math.random(lower, upper) generates integer numbers between lower and upper (both inclusive). So it is generating pseudo-random numbers between (1,2) based on a DISCRETE UNIFORM DISTRIBUTION of 1s and 2s.

1- https://github.com/Aussiemon/Vermintide-2-Source-Code/blob/def7078ff2697f75ad4731fd981ea9c275cd9769/scripts/settings/dlcs/morris/morris_buff_settings.lua#L1856
2 - http://lua-users.org/wiki/MathLibraryTutorial

What is a uniform distribution and what is a discrete uniform distribution?

I am going to provide a massively toned down ELI5 version of this because 90% of the people that have posted in this thread won't understand the technical jargon.

A Uniform distribution is a Probability Density Function that governs how likely it is for a particular event of a RANDOM VARIABLE to occur. A famous example of this is the "Bell-Curve" for a continuous distribution.
Following along, a Discrete Uniform Distribution is a probability density function that governs how likely it is for a particular event of a DISCRETE RANDOM VARIABLE to occur. In this case, all you need to know is that the probability density function says it is equally likely for 1 or 2 to be generated according to this distribution.

What does this mean?

This means the code is generating numbers based on a discrete random variable, which can only contain 2 values {1, 2}. This is a special type of discrete RV, which is called the Bernoulli random variable. Discrete RV can contain any number of possible outcomes {1,2,3,...} but Bernoulli can contain exactly 2. In addition, since math.random generates "PSEUDO RANDOM" numbers based on a distribution, we can safely assume the BINOMIAL DISTRIBUTION is used here, which is special type of Bernoulli distribution if each trial is independent.

This is important because with the Binomial Distribution, which models a random event with exactly 2 possible outcomes, where the chance of "success" is p, we can calculate the probability of getting exactly x successes out of N "tries", denoted as P(X=x), or fewer than x success P(X<=x), and there are many many online calculators that do exactly that.
ie https://stattrek.com/online-calculator/binomial.aspx

Take the example above by Xaphnir where Valaya's Brew (50% proc chance) procced 9 times out of 40 attempts.

N = 40, p = 50%, then P(X=9) = 0.00024869121, or 0.0248%
Last edited by OK BOOMER; May 12 @ 10:42am
< >
Showing 1-15 of 142 comments
morsdood Apr 25 @ 5:18am 
welcome to RNG
Originally posted by morsdood:
welcome to RNG
If you flip a coin 1000 times and get 150 tails, are you gonna say it's "rng" or that it's a rigged coin.
50% chance doesn't equate that it will activate half of the uses...
Originally posted by WreckerALeX:
50% chance doesn't equate that it will activate half of the uses...
Yes it kind of does...

Flipping a coin has a 50% chance of landing on a tail. it may not land on a tail every other coinflip, but over a sample size large enough, say 10000 coinflips, it should land on a tail ROUGHLY 5000 times.
I didnt do any numbers, but i gotta agree, it doesnt feel like its 50%.
Originally posted by OK BOOMER:
Originally posted by WreckerALeX:
50% chance doesn't equate that it will activate half of the uses...
Yes it kind of does...

Flipping a coin has a 50% chance of landing on a tail. it may not land on a tail every other coinflip, but over a sample size large enough, say 10000 coinflips, it should land on a tail ROUGHLY 5000 times.
No, it mean that there is 50% chance that you will get tail every time.
You can flip a coin 1 million times and only get tail 1 million times. The chance are super low but thats RNG for you.

I would suggest you search online on what RNG really is, since 50% activation is 50%, not every 2 ult you do will activate it.
Record a bunch of observations and see if it's close to 50%. Law of large numbers state it should approach the expected result. Me personally from my limited sampling, it doesn't get close to 50% but need more samples.
Originally posted by Rider of Retarded Lion Prince:
Originally posted by OK BOOMER:
Yes it kind of does...

Flipping a coin has a 50% chance of landing on a tail. it may not land on a tail every other coinflip, but over a sample size large enough, say 10000 coinflips, it should land on a tail ROUGHLY 5000 times.
No, it mean that there is 50% chance that you will get tail every time.
You can flip a coin 1 million times and only get tail 1 million times. The chance are super low but thats RNG for you.

I would suggest you search online on what RNG really is, since 50% activation is 50%, not every 2 ult you do will activate it.
It seems you don't know what probability theory is, because if you did, you wouldn't tell someone to search what RNG is, lol.

I can tell you the probability of getting only tails on 1 million coin flips with a FAIR coin. it's 0.5 to the power of 1 million, OR roughly
1.01003×10^-301030

THATS more than 300000 ZEROS after the decimal, that's more zeros than there are in a GOOGLE. That's not a "super low chance and RNG for you", that's mathematically impossible.
Originally posted by Starship Pooper:
I didnt do any numbers, but i gotta agree, it doesnt feel like its 50%.


Originally posted by tap and die:
Record a bunch of observations and see if it's close to 50%. Law of large numbers state it should approach the expected result. Me personally from my limited sampling, it doesn't get close to 50% but need more samples.

Base on my personal observations. It's roughly 10% activation chance, but a larger sample would be appreciated

(or some fatshark dev could just confirm if there is a hidden mechanic)
Rack Apr 25 @ 6:21am 
It also doesnt work with some poitions
I've had my fair share of "rng is rng" in other games, but never a discrepancy this large. I'm also theorizing there is a hidden mechanic we don't know, or Fatshark straight up misinterpreted or mistyped the odds...... in very much Fatshark fashion.
Eudicots Apr 25 @ 7:25am 
I feel like it triggers roughly half the time used it with bombs and strike. And by probability a single instance and a large pool can have different chances even if its just considered unlikely not impossible. You'll have to have some sort of data to back your claim.
Originally posted by Eudicots:
I feel like it triggers roughly half the time used it with bombs and strike. And by probability a single instance and a large pool can have different chances even if its just considered unlikely not impossible. You'll have to have some sort of data to back your claim.
I only have my personal experience, so I don't want to be jumping to any conclusions here. That's why I said it "Feels" like it's not 50% and not it's "definitely" not 50%.

Also, what class were you playing? A class with a long ult CD like IB or short ult CD like handmaiden. I've been mostly getting this boon on handmaiden which has a really short CD, like around 15 seconds. And it does not trigger half the time me. Like I said, it triggers roughly once in 10 ults.
Eudicots Apr 25 @ 7:47am 
Originally posted by OK BOOMER:
Originally posted by Eudicots:
I feel like it triggers roughly half the time used it with bombs and strike. And by probability a single instance and a large pool can have different chances even if its just considered unlikely not impossible. You'll have to have some sort of data to back your claim.
I only have my personal experience, so I don't want to be jumping to any conclusions here. That's why I said it "Feels" like it's not 50% and not it's "definitely" not 50%.

Also, what class were you playing? A class with a long ult CD like IB or short ult CD like handmaiden. I've been mostly getting this boon on handmaiden which has a really short CD, like around 15 seconds. And it does not trigger half the time me. Like I said, it triggers roughly once in 10 ults.
I was playing with WHC and GK so the cool downs were longer than a Handmaiden. Haven't tried every class of portion combination yet.
I have noticed that this effect does feel nothing like 50% at all.
I used my ult like 10 times in a row and n o t h i n g. I thought it was bugged so I just switched around different potions.
Maybe it depends on the ult you have? If you got a short cooldown its less likely and if you got an long cooldown its nearly always maybe?
I played battlewizard with double ult and it can infact trigger with both ults(happened twice in a row quite a long time after not working at all).
But thats just what I can see, if there was a reliable way to get X boon's I could replicate it and just try it out(is there a mod for this by chance?)
< >
Showing 1-15 of 142 comments
Per page: 15 30 50