Gaussian RNG issue in PoE (both games)

"

Timbo Zero#8289 wrote:

A map (or in poe2 parlance waystone) can be tier 1-15, a RNG generator can ‘roll’ to determine the tier you get. Without outlier filtration someone will get very bad sustain and someone else will get very good sustain.

Then why you even mentioned Gaussian? It has nothing really to do with it. Could just say that possible tier maps should be limited to, for example, +/- 2 of your current tier. Like we already agreed, the problem is with range of possible values and not with using Gaussian instead of, for example, uniform.

"

Timbo Zero#8289 wrote:

We should all get on average the same sustain, not (as has happen to me persistently in poe1) bad sustain every league to the point you have to buy them to keep playing.

It's not how it works :) It can't be persistently bad for a long time no matter which distribution GGG are using to draw that variable. Central Limit Theorem will make your average to be normally distributed with very low deviation. If you are consistently have troubles with map sustain and other players don't, it has nothing to do with luck, you are just doing something less efficiently.

Btw, in the long run doesn't matter at all what original distribution was — be it Gaussian, uniform, Poisson, binomial or literally anything else. Only 2 parameters matter in the long run — average value and deviation of original distribution.

"

Suchka_777#4336 wrote:

"

Timbo Zero#8289 wrote:

A map (or in poe2 parlance waystone) can be tier 1-15, a RNG generator can ‘roll’ to determine the tier you get. Without outlier filtration someone will get very bad sustain and someone else will get very good sustain.

Then why you even mentioned Gaussian? It has nothing really to do with it. Could just say that possible tier maps should be limited to, for example, +/- 2 of your current tier. Like we already agreed, the problem is with range of possible values and not with using Gaussian instead of, for example, uniform.

"

Timbo Zero#8289 wrote:

We should all get on average the same sustain, not (as has happen to me persistently in poe1) bad sustain every league to the point you have to buy them to keep playing.

It's not how it works :) It can't be persistently bad for a long time no matter which distribution GGG are using to draw that variable. Central Limit Theorem will make your average to be normally distributed with very low deviation. If you are consistently have troubles with map sustain and other players don't, it has nothing to do with luck, you are just doing something less efficiently.

Btw, in the long run doesn't matter at all what original distribution was — be it Gaussian, uniform, Poisson, binomial or literally anything else. Only 2 parameters matter in the long run — average value and deviation of original distribution.

Yet I have literally experienced map sustain that is so bad in a minimum of 75% of the many many leagues in poe1 that is so statistically impossible that it would suggest that the RNG seed in poe1 is account bound not character bound.
In poe2 (which is why I estimate, it’s too soon to know for sure) also does this. You can read and talk to many people who cannot sustain t15 waystones, and others that don’t believe them because they have such good sustain that they get confirmation bias that the low sustain players don’t know how to play the game.
Personally in poe2 so far I have broken my poe1 ‘curse’ and been able to sustain t15 waystones with no issue at all.
There is clearly outliers for ‘map’ sustain in both games and a rejection of out of limits RNG values would alleviate this for every player but still allow one player to do 10% better than average and another to do 10% worse than average. No filter and 27% of players are out of range but up to 13% for the better or the worse.

And yes you are correct that CLT with a real world die rolled an infinite number of times will produce an equal number of 1-6 results. Sadly in the computer world CLT doesn’t work because we can so far only create pseudo random numbers (by many of the ways you have named, but nearly everyone uses Gaussian because the code is so simple) and with out filtering the out of range results you get ‘unpleasant anomalies’.

"

Timbo Zero#8289 wrote:

And yes you are correct that CLT with a real world die rolled an infinite number of times will produce an equal number of 1-6 results.

This is not what CLT is about. CLT just says that average of multiple attempts will be drawn from normal distribution, with same average value and smaller deviation. And it woks totally fine with pseudo random numbers.

"

Timbo Zero#8289 wrote:

Yet I have literally experienced map sustain that is so bad in a minimum of 75% of the many many leagues in poe1 that is so statistically impossible that it would suggest that the RNG seed in poe1 is account bound not character bound.

I can't prove you that you are wrong, but for me it sounds like some kind of conspiracy theory, I wouldn't believe in it even slightly. Much simpler explanations is that different players are playing differently and those who are doing things less efficiently tend to blame statistically impossible bad luck.

"

Suchka_777#4336 wrote:

"

Timbo Zero#8289 wrote:

And yes you are correct that CLT with a real world die rolled an infinite number of times will produce an equal number of 1-6 results.

This is not what CLT is about. CLT just says that average of multiple attempts will be drawn from normal distribution, with same average value and smaller deviation.

"

Timbo Zero#8289 wrote:

Yet I have literally experienced map sustain that is so bad in a minimum of 75% of the many many leagues in poe1 that is so statistically impossible that it would suggest that the RNG seed in poe1 is account bound not character bound.

I can't prove you that you are wrong, but for me it sounds like some kind of conspiracy theory, I wouldn't believe in it even slightly. Much simpler explanations is that different players are playing differently and those who are doing things less efficiently tend to blame statistically impossible bad luck.

CLT is a concept we understand but to turn it to laymans’ terms would mean rolling a single die can only result in a 1,2,3,4,5,6 and an infinite number of rolls would create an equal number of each.

I agree from an outside perspective my claim of poe1 consistency in bad map sustain does sound like a conspiracy theory or player incompetence (not the word you used) but I can not prove this is true…
I have literally played the same way as a friend in the Phrecia event currently ongoing on poe1 and he has (again) had to give me maps because he is getting so many above the mean and I am (once again) getting so many below the mean. This effect has been happening since we met 3 years ago with one one league (Settlers) being the exception.

When 0.2.0 of PoE2 releases and we can all talk to each other I will (if you are in agreement) introduce you to him for verification. Unless you play on XBox in which case I can do it as soon as he wakes up.

"

Timbo Zero#8289 wrote:

CLT is a concept we understand but to turn it to laymans’ terms would mean rolling a single die can only result in a 1,2,3,4,5,6 and an infinite number of rolls would create an equal number of each.

Well yes, we can look from this angle too and say, that everyone basically gets same proportion of each tier map in the long run.

"

Timbo Zero#8289 wrote:

When 0.2.0 of PoE2 releases and we can all talk to each other I will (if you are in agreement) introduce you to him for verification. Unless you play on XBox in which case I can do it as soon as he wakes up.

But how I could verify it? I don't want to watch you both playing hundreds of maps to check it :)

And you will have different death rate, different MF (I'm not sure though if MF affects waystones), will probably have some difference in atlas strategy of juicing and clearing maps. Even with same death rate, maybe for example he goes to boss straight away when he see boss on the map and you are clearing some rares first, maybe it will make you to die more often before boss, while he is dying to those rares after doing boss. That probably wouldn't matter too much, but my point is that there are quite a lot of potential differences and it's hard to be sure that you are playing the same way.

"

Timbo Zero#8289 wrote:

"

Suchka_777#4336 wrote:

"

Soulscar1#4896 wrote:

The distribution simply implies the more you play (or more quantity you use) the more likely your "luck" is going to fall in the upper +ve deviation more often/sooner hence why even a little quantity is so strong.

What you are talking about? Expected value you get never change no matter how long you play. You got something completely wrong, which is common case when someone (OP) is trying to sound smart and say some stupid things in the foggiest way possible (otherwise, noone will believe that he is saying something smart).

Would you care to elaborate on why you think I am not smart enough to say what you say is smart sounding but stupid please ;)

I have used some exceptionally simple maths and basic C programming code to prove my point, and that I understand the issues. And left a challenge to GGG support to disprove my work.

You actually trolled yourself I offered a compliment.

High functioning Asperger’s can cause such things to occur frequently with me alas lol