Think i know why RNG is inconsistent in PoE
" Define true random, please. For me, true random would be that you can draw multiple numbers, they follow a distribution, and they are iid (independent and identically distributed). Do you have a different idea of what "true random" means? Using the CMB allows you to produce such numbers. Random number generators typically used (and available in most programming languages, and probably used by GGG) are deterministic pseudo-random-number generators, and don't produce iid numbers. But they are "sufficiently random" for a lot of use cases. Remove Horticrafting station storage limit.
|
|
|
If they "follow a distribution" and produce identically distributed results then you don't have a random process. Again... there is no such thing as a random process so trying to compare your "definition" to something that doesn't exist is going to be incredibly frustrating. There should really avoid the term completely. Use "unpredictable"... it's much less likely to provoke nonsense like this thread.
Last edited by Shagsbeard#3964 on Feb 4, 2021, 12:45:22 PM
|
|
" 1 is pretty exactly 1 more than 0. If you mean, it is infinitely many times 0, that is incorrect, too. 1/0 isn't infinite. It simply isn't defined. " Seems like your pseudo-random-number generator wasn't the best, then. If the random number generator would produce truly IID numbers, then there can't be a strategy that improves the odds. Maybe you just got lucky. Or unlucky, depends on the point of view, because you now believe you can influence the odds, which you can't, at least not if GGG is using an unbiased, high-quality pseudo-random-number generator. Remove Horticrafting station storage limit.
|
|
" Not sure what your mathematical background is. Do you know the term "independent and identically distributed", also called IID, and its meaning? If not, I am afraid we are talking past each other here. Remove Horticrafting station storage limit.
|
|
" What i meant, that if for example you had one ticked, chance to win lottery was 1:1000000 and if you had 2 tickets your chance to win was 2:1000000 (or 1:500000), you chance to win improved twice. Now if you had 0 tickets and 1 ticket, its no chance at all vs very small chance. Going from 0 to 1 isnt improving your odds by 1, 2, 3 or million times, your chances became infinitively larger. Correct me if i am wrong. " You're right. Rechecked the calculations and found a mistake. In corrected calculations change was only 1%, this can be attributed to sample size not being large enough. My mistake was, that i was too hasty and didnt double check. Edit: Observed intersting thing, in a sequence of randomly generated numbers between 1 and 7, the largest streak of unlucky numbers (assuming that lucky number is 1) was 48. If a chance of successfully corrupting a gem was roughtly 1 in 7 (getting +1 level), it is possible to get a streak of 48 failed corruptions. amazing. "The only thing necessary for the triumph of evil is for good men to do nothing" - Edmund Burke Last edited by Toshis8#1464 on Feb 4, 2021, 4:34:25 PM
|
|
|
RNG is alway pseudo-random and can be tweak by devs. ( not to be confused with thing like how evasion work in POE compare to dodge, evasion is not RNG ... it's cycle-based )
Pretty sure Unique drop are not "truely ramdom". For exemple, in the pool of items that you can find in a specific zone/ilvl ( if restriction exist ) you will find 10 copy of the same bad unique you always find every league before finding a "good unique" that you find 1 time every 5 league ... because devs tweak the generator like they want. If The number 1 is bad item and number 2 is the good one you influence your generator by adding more 1 than 2 in your number pool, so to speak. Last edited by Deonbekende#4380 on Feb 5, 2021, 2:03:49 AM
|
|
|
My rng has been dogshit this league.
So was performance. I'm not done with the league yet. So no money for dev's. No good rng + no good performance = no bucks. Simple formula, right? "Parade your victories, hide your defeats. Mortals are so insecure."
Poe 0.2/10 Returning to poe in 3.27: ATROCIOUS game performance, 5/10 league and I apprently missed the loot back in the last league. The more things change the more they stay the same.. |
|
|
If we took 0 and 1, and ran it through random number generator, we would get something like this:
0 0 1 1 0 0 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 1 1 1 0 1 0 0 0 0 0 0 1 1 1 1 0 0 0 0 1 0 1 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 1 1 1 0 1 1 1 1 0 1 1 0 1 0 0 1 1 1 0 1 0 0 1 0 0 0 0 1... The hypothesis is if game developers were to apply some additional conditions to change number distribution, like: If first number is 1, then next number has increased chance to be 1 again by x%. If 1 repeated 8 times in a row, 0 gets increased chance by x%. And we would get something like this: 11111111000000001111111100000000... With small variations here and there. Another example: if number 1 appeared atleast 5 times in the last 10 rolls, in next 20 rolls 0 has increased chance by x%. You get the idea. Would game developer use algorithm to modify rng and what their reasons could be? edit: I dont know how it actaully is, but this could be done to prevent people from having very long streaks of bad luck "The only thing necessary for the triumph of evil is for good men to do nothing" - Edmund Burke Last edited by Toshis8#1464 on Feb 5, 2021, 2:38:09 AM
|
|
" Colloquially, you are right. That is how people describe it, "an event that can happen is infinitely more likely to happen than one that cannot". Mathematically, you are wrong. If you compare two events, event E1 with a probability p1 to happen, and event E2 with a probability p2 to happen, then E1 is (p1/p2) times more likely to happen than E2. For example, a randomly chosen day of the year is 2.5 times more likely to be a weekday than a weekend (p1=5/7, p2=2/7, p1/p2=(5/7)/(2/7)=5/2=2.5). However, if p1=1 and p2=0, then p1/p2=1/0, which isn't infinity. It is not defined. 1/0 has no result.
Spoiler
You could argue that it is infinity, because 1/x tends to infinity if you start with x=0.0001 and make x smaller and smaller, getting ever closer to zero. However, you could just as well argue that it must be -infinity, as 1/x tends to -infinity if you start with x=-0.0001, and make x larger, getting ever closer to zero from the negative side.
Spoiler
Side note: Infinity isn't even a number, it is actually a pretty complicated concept. It is also not very intuitive. For example, you would probably agree that there must be more natural numbers (1,2,3,4,5,...) than prime numbers (2,3,5,7,11,...), because, I mean, that is obvious, no? Wrong. Both are infinite. There are exactly as many natural numbers as there are prime numbers, because you can map them to each other. Like, to every natural number n, you can map exactly one prime number (the n-th prime, so natural number 5 would map to prime number 11, for example).
On the other hand, if sets of things contain infinitely many elements, that doesn't necessarily mean they contain "the same number of" elements. For example, there are infinitely many natural numbers. There are also infinitely many real numbers (real numbers are things like 1.3, 2.85903418904, 5.0, pi, sqrt(2) and so on) between 0 and 1. However, there are a lot more real numbers between 0 and 1 - infinitely more - than there are natural numbers. And yes, I know, it is weird. " Cool. " Yeah, RNG can be quite shitty, and humans are really bad at understanding randomness. Like, our gut feeling towards how random numbers should work is often very wrong. If you are interested, there are formulae that you can find that describe how often such really unlucky streaks are expected to occur. They aren't as rare as you might think. Remove Horticrafting station storage limit.
|
|
" That could of course be. We know they do something like that for evasion (and maybe dodge?). They massage their numbers to be "less random" because it makes gameplay smoother. I would doubt that they do it for gem corruptions, but of course I could be wrong. You could, in principle, find out by corrupting a very large number (100s to 1000s) of gems and then applying statistical tests to investigate their randomness. Remove Horticrafting station storage limit.
|
|
