Does anyone actually know how the RNG really works?

[Randomonium]

7 hours ago, Youmukon said:

Long story short, no it's not rigged. The way loots are generated is completely detached from things such as the class you are playing or the money you have invested to the game. Some RNG aspects aren't equiprobable however (i.e. the Daily Dash wheel behavior isn't the one of a standard dice, you have less chance to get a 6 than a 2 or 3, the game shows you that, since the wheel isn't split in 6 equal parts).

The thing is that since there is no transparency (ie showing you the roll), and psuedo RNG/lookup table type RNG used for most games propensity for streaks due to many different factors folks who do have the same problems with getting precisely the same wrong thing over and over and over and find out that nearly everyone they talk to had the same experience, you can understand how folks might come to the conclusion that it is indeed "rigged".  The same mechanisms used by the game to ensure it doesn't pay out too often should also be used as a failsafe for "unlucky" players as you put it.  Each time this unlucky person fails their chances should become a small amount better to reaching their goal.  You would see a whole lot less of these threads if you would apply some CYA to player experience and not just the game's bottom line.

 

While we are speaking of the games protections on too many payouts, could you pass along that the nerf that goes into effect on weekend payouts for epic dungeons needs some tweaking so it doesn't always payout 4 mans as 6 mans and 6 mans almost like a blue dungeon.  I can understand that because there are more players putting in more hours on weekends you don't want them getting too much and thereby encouraging farming, but it has become abundantly clear to those paying attention that running purples on weekends is practically useless due to the runaway RNG in place.

[zapzap77]

8 hours ago, Youmukon said:

Long story short, no it's not rigged. The way loots are generated is completely detached from things such as the class you are playing or the money you have invested to the game. Some RNG aspects aren't equiprobable however (i.e. the Daily Dash wheel behavior isn't the one of a standard dice, you have less chance to get a 6 than a 2 or 3, the game shows you that, since the wheel isn't split in 6 equal parts).

 

Thank you.

 

actually if the distributions on weapons that aren't yours were equal it would be evidence that it WAS rigged.

 

That said there is no real transparency on the wheel of fate.  it shows unequal sections but not what they correspond to.

 

So you are saying that when logic dictates that chances should be equal (class weapon chests/drops) they are, and when not there are ingame clues otherwise, right?

 

Good to know.  I only wondered because other games have done such trickery in the past, and that has been proven.  Also because it's trivially easy to make the RNG cheat, so naturally people are suspicious.

 

An employee saying on record it's fair RNG is good enough for me. :)

[zapzap77]

btw odds are abour 33% of getting screwed for 30+ at least once during your characters career are actually about 33%

 

this is made on n estimate of 20 weapon breakthroughs (valor stone one doesnt' count, not sure which other ones do not)

 

after working out the odds of getting screwed for 30 in a row, you then subtract that from 1, then raise the new number to the power of the number of weapon breakthroughs locked begind weapon chst, then subtract that from one. which gets areond 33% change of geting screwed at least once.

 

meaning a third of all platers. that's a lot.

 

and it gets worse with each weapon chest added.

 

think this clears up the arguments. 

 

PS: over 75% of players will have to open over 20 chests for  weapon at least once based on the same 20 figure.

PPS: if soul fighter needs a new weapon type, it will get MUCH worse.  over 86% of players who start a character after soul fighter release will get screwed for 30 on at least one of their weapon chests.  only way to fix that is to make the rng cheat by leaving one weapon out of the weapon chest that's dependent on your class (different for each class).

[Minity SN]

they are selling gambling mechanics to children.. I dont expect a company doing this to have fair RNGs :) atleast I would rig them myself if it means that I can afford my audi  A7 this way. just swap the perspectives and imagine you are the publisher/developer. everyone would do it

[Shadzar]

15 hours ago, Youmukon said:

In theory, roughly 1% of players will need 34~35 tries to get their class weapon out of a chest.

Thanks for explaining all that, and let me say that this right here is a MAJOR problem with RNG. You have set up to begin with that 1% of players are going to get a negative play experience. The game is purposefully designed to make a percentage of the populace not enjoy the game. This is just bad design.

 

Roughly 3% chance to get your class weapon while the others are seemingly "equiprobable" violates standard deviation. Since you are using (p=0.05), I guess you know probability formulae.

 

Assassin

Blade Master

Blade Dancer

Destroyer

FM

KFM

Summoner

Warlock

(Soul Fighter)

 

I included Soul Fighter because its weapons have begun to drop, or at least did 3 weeks ago when I last played because I sold the one I got from the Blackram Supply weapon chest on the market. So we know it is in the game and we all anticipate it coming. That is 9 classes

 

9*3=27%

 

Yup, that aint possible because each weapon chest gives a weapon, and that isnt 100%. So I would guess your @3% number is spot on for the class specific weapon, and the rest are equally probable spread on a linear distribution table for the remaining @97$ for those other 8 classes.

 

That means you deficient outlier of 3% can rear its head more often than the 1% of players you claim, to having a negative play experience. This is stunted by the fact SOME weapons can be bought and sold on the market, but the point of diminishing returns comes quicker as you level higher and things cannot be bought or sold on the market because they are "unable to trade". Can't find my other post but from memory:

 

Blackram NArrows

 

Golden Deva

Infernal

 

Hujikar

Profane

Brightstone

Storm

Siren

Dark

Silver Deva

Pirate

Warrior

 

 

That is "unable to trade" weapons by zones. Sure you can say that will change this week, but the problem is people are STILL stuck on that path. I can't speak for Silverfrost "unable to trade" that relies solely on RNG and that 3% chance of getting them as I am stuck on that path, unable to leave it unless I start over, abandon ALL the work I have done and move on to the Silverfrost path with a new weapon. Which makes that 3% drop rate even more drastic to players.

 

Sure some players can get lucky and beat probability and get that 3% on the first drop, but the fact that it takes ANY players 34-35 tries to beat probability (remember it is based on NOW, single instance cases of probability, you didnt say that the chance increases as you open more boxes so EVERY weapon chest is 3% independantly no matter how many you open and could mean someone with REALLY bad luck on the RNG could even take 100 tries to get their weapon) means that you are going to have irate players.

 

Purposefully having some players discriminated on by badly designed RNG is just bad design and bad management of a game. Only the designers of the RNG system are at fault for this and THEY take none of the fluff, but pass it on to CSR, forum moderators, etc. to have to deal with. The developer that came up with this RNG schema can jsut hide away free and clear of having to deal with the reprecussions of their actions and bad design.

[DomiSotto]

One of the reasons I did go to the Oathbreaker path was my 28 tries at Hujikar's weapon. Since then I was not dealt the unfortunate RNG, but since every roll is a separate dice roll, I can imagine just how hard the game could become if someone pulls poor odds. The only consolation is that it reduces the cost of feeding the weapon because you get a lot of elemental powders. Getting at least something out of it is silver lining.

 

@Shadzar: you will transfer to the OB path if you do not finish the Pirate one. It might come with a higher cost than you have intended, and that's unfortunate. but with the mushrooms it is not so bad.

 

 

[Shukran]

i dont want to go too much deep in statistic because i an rusty about it. but

 

100/8 = 12,5% chance toi get a specific class weapon.

 

u open 100 chests. how much is % to get weapon u need? 12,5%.

does it means that if you open 100 chests,  12 chests give you the weapon? no, since there is no correlation between chests, each chest has 12,5% chance to get the weapon. no formulae needed.

[Shadzar]

29 minutes ago, DomiSotto said:

@Shadzar: you will transfer to the OB path if you do not finish the Pirate one. It might come with a higher cost than you have intended, and that's unfortunate. but with the mushrooms it is not so bad.

Now, I wont. I am at the very screwed part: True Pirate Stage 5. Siren will move on at most stages to the new path, but I will have to complete True Pirate Stage 10 to be able to move on. It wont automatically switck me because there is no place to Switch as the switch points arent there but for 3 palces

 

Profane -> new path

Awakened Siren -> new path

Awakened Pirate -> new path

 

once you are in True Siren or True Pirate you are stuck completing them.

[Shadzar]

3 minutes ago, Shukran said:

i dont want to go too much deep in statistic because i an rusty about it. but

 

100/8 = 12,5% chance toi get a specific class weapon.

 

u open 100 chests. how much is % to get weapon u need? 12,5%.

does it means that if you open 100 chests,  12 chests give you the weapon? no, since there is no correlation between chests, each chest has 12,5% chance to get the weapon. no formulae needed.

Except it isnt linear distribution, it is made on purpose to make it HARDER to get your class specific weapon, and that chance was spelled out to 34-35 tries and 100/33= @3%. You can see through game the REST of the weapons are evenly distributed on a linear curve. which gives @11%~12% for each weapon a class cannot use.

[DomiSotto]

9 minutes ago, Shadzar said:

Now, I wont. I am at the very screwed part: True Pirate Stage 5. Siren will move on at most stages to the new path, but I will have to complete True Pirate Stage 10 to be able to move on. It wont automatically switck me because there is no place to Switch as the switch points arent there but for 3 palces

 

Profane -> new path

Awakened Siren -> new path

Awakened Pirate -> new path

 

once you are in True Siren or True Pirate you are stuck completing them.

Then you are getting the cheaper path, just like you wanted?

[Dlacik]

@Shadzar You are wrong. Youmukon said that 1% of players need 34~35 tries to get their weapon. That basicaly means the probability of getting your class weapon during 34 tries is 99%. If the chance to get your class weapon were only 3% you would have only about 64,4% chance to obtain a weapon for your class in 34 tries.
The 99% chance during 34 tries roughly match 11,1% chance to get your weapon class (evenly distributed probability among 9 classes).

[Shadzar]

Can you do math? 100/33 = ??

 

When you place 100 over a number that gives notation we call a percent. that means we get 3% chance of the class specific weapon. Probability states that if you have 100 people each open a box, 3 of them will them get the right thing while 97 will get the wrong thing.

 

You can easily find the distribution from the information given in this thread and through playing the game to determine this percentage.

 

NCSoft will have to show me more numbers to change my mind because math doesnt lie. you have a 97% chance to get a weapon you can't use from any weapon drop. Some will be lucky and get that 3% sooner, while others it will take longer, up to 34~35 tries.

[Shadzar]

8 minutes ago, DomiSotto said:

Then you are getting the cheaper path, just like you wanted?

Not really, I am just one that doesnt move on from an area to another until I have completed the current area. Since There are accessories left in Moonwater, I needed to continue with Moonwater paths until they were ALL done in order to progress so I was not looking back and backtracking for no reason and could focus on the new progression ahead of me. :)

 

Not about cheaper, just about my preferred playstyle @40 years of RPGs that differs from the forced meta in BnS.

[TheQuibbler]

26 minutes ago, Shadzar said:

Now, I wont. I am at the very screwed part: True Pirate Stage 5. Siren will move on at most stages to the new path, but I will have to complete True Pirate Stage 10 to be able to move on. It wont automatically switck me because there is no place to Switch as the switch points arent there but for 3 palces

 

Profane -> new path

Awakened Siren -> new path

Awakened Pirate -> new path

 

once you are in True Siren or True Pirate you are stuck completing them.

No, Wrong. 

Awakened Siren and Awakened pirate, has no choice. You can only move on to True Siren or True Pirate that those. 

 

True Profane > Oathbreaker

Awakened Siren > True Siren > Awakened Oath

Awakened Pirate > True Pirate > True Breeze

 

(I am not 100% sure about the naming of the oathbreaker parts, since I went for True Pirate.)

[Dlacik]

15 minutes ago, Shadzar said:

When you place 100 over a number that gives notation we call a percent. that means we get 3% chance of the class specific weapon. Probability states that if you have 100 people each open a box, 3 of them will them get the right thing while 97 will get the wrong thing.

This is where you are wrong. Noone said that when 100 ppl each open a box 3 of them will have right thing while 97 will have wrong thing. Youmukon said that if 100 ppl open 34~35 box, 99 of them will get at least one right weapon while 1 won`t get any right weapon. You can do your 100/33 if you want but it`s definitively not probability of getting right weapon what you get from that division.

[Shadzar]

No, she said out of ALL the player (100%) 1 % would have to try 34~35 times to get the weapon.

 

That means some people have only a 3% chance.

 

Now I wasnt trying to get into it, but it means the RNG isnt stable, but changes as a possibility for further permutation of the RNG and loot tables.

 

So just sticking with the simple, those 1% have only a 3% chance of getting the right weapon, that is the only real number we have to work with, so in a "fair" system that would mean ALL players have a 3% chance of getting the correct weapon, just some have better luck that others and are able to get within that #5 sooner than 34~35 tries.

[Dlacik]

@Shadzar Sry but that`s not how statistics or probability works. You can`t just ignore 99% of results and say the probability is 3% because 1% of results shows it.

 

I don`t know proper terms in english to detaily explain cause english isn`t my primary language so ask your math teacher about it or read something about it here https://en.wikipedia.org/wiki/Binomial_distribution

[Shadzar]

This is why i used plain english, not too many technical terms and stated that the 1% for all the math. that was a baseline. Maybe your child self can ask your math teacher in your native language since it has been decades since I had any teachers. so in proper fashion for us old fogies:

 

 

GET OFF MY LAWN! :p

[Arohk]

this is how RNG works, If your chance to get YOUR weapon is 1/8:

 

You open a box and get it first try

then

your next weapon will be 1/16 

if you also get that in first try

you are very likely to need 24 boxes on your third weapon and so on.

 

if your first weapon drops on the 20th box

you are more likely to get your next weapon from an early box.

 

 

 

=)

[Youmukon]

All trials are independent when it comes to weapon chests for example. The game doesn't look at previously attempted trials. Opening a class weapon chest is like rolling an 8 faced dice. Every new trial isn't related to the previous one. It still doesn't prevent averages and theoretical calculations to be made.

 

Saying that 1% of unlucky players will need 34~35 tries to get their weapon is just binomial distribution. You take the probability of failure (0.875), and raise it to the power of 34, you end up with ~1.07%. It means that chance of failure @34 trials are ~1.07%, which means that 1.07% of players are likely to fail obtaining their weapon after opening the chest 34 times. This is a purely theoretical value, it doesn't mean that it will always happen, nor that the game tracks it. But over a large sample of players, if we were to look at all chests opened, we should roughly find a similar failure rate (excluding blue keys).

1 hour ago, Dlacik said:

Youmukon said that if 100 ppl open 34~35 box, 99 of them will get at least one right weapon while 1 won`t get any right weapon.

Basically this :)

[Arohk]

still, rolling an 8 , 8 times in a row on an 8 sided dice is only a 0,09% chance.

 

sure it can happen to someone, and i will happen to that one blessed lucky player who gets ALL of his weapons in only a single box each.

[TheQuibbler]

1 hour ago, Arohk said:

this is how RNG works, If your chance to get YOUR weapon is 1/8:

 

You open a box and get it first try

then

your next weapon will be 1/16 

if you also get that in first try

you are very likely to need 24 boxes on your third weapon and so on.

 

if your first weapon drops on the 20th box

you are more likely to get your next weapon from an early box.

 

 

 

=)

No? This is not how it works.

Your first chest drops your weapon (1/8th chance) on the first try.

Then the next chests will still have 1/8th chance of your weapon. If you get it on the first try again.

Your next chest will now have, 1/8th chance to drop your weapon.

[Shadzar]

3 hours ago, Youmukon said:

All trials are independent when it comes to weapon chests for example. The game doesn't look at previously attempted trials. Opening a class weapon chest is like rolling an 8 faced dice. Every new trial isn't related to the previous one. It still doesn't prevent averages and theoretical calculations to be made.

 

Saying that 1% of unlucky players will need 34~35 tries to get their weapon is just binomial distribution. You take the probability of failure (0.875), and raise it to the power of 34, you end up with ~1.07%. It means that chance of failure @34 trials are ~1.07%, which means that 1.07% of players are likely to fail obtaining their weapon after opening the chest 34 times. This is a purely theoretical value, it doesn't mean that it will always happen, nor that the game tracks it. But over a large sample of players, if we were to look at all chests opened, we should roughly find a similar failure rate (excluding blue keys).

Basically this :)

No, it can't be. You dont get to have your cake and eat it too.

 

Either it is Linear distribution as you said "rolling an 8 faced die" where EACH side is equally probably of coming up like this:

http://anydice.com/program/4b1

 

Which each probable outcome is 12.5%, or it isnt. Please make up your mind.

 

IF you have, like you claim, 100 people rolling this d8 then probability states that each of the faces will come up 12~13 times

 

This means that 12 of those 100 players will get their weapon, and 87 will not, with a margin of error being 1 player.

 

So when you are talking about binomial distribution you are NOT talking about rolls being independent, but a cumulative based on how many rolls have come to get a set series of outcomes.

 

THIS means, that it is NOT independent on a linear curve such as an eight-sided dice, but instead factors in PREVIOUS rolls in order to determine if it is time for the "positive" result to appear. Now whether this "possitive" result appears as a function of the class generating the RN or just based on total number of "rolls" is what is in question.

 

Currently you have described the RNG as to being a "claw machine" which are ALWAYS RIGGED so that after only X amount of tries using it will it produce the positive result.

 

This "rigged claw machine" mechanic and analogy is EXACTLY what people can see in the game as the results are produced, AND agrees with your 1% of players having to do it so many times.

 

When you see independent mentioned in binomial distribution it just means all the dice aren't rolled at once, but I guarantee you for ANY single die rolled you will ALWAYS get a linear curve for the probability because each side has EQUAL chance of coming up as any other side. So binomial distribution is checking for the result against ALL past rolls to see how often that singular result came up over a course.

 

Your d8 example will always yield a linear curve because only 1 object is being tested.

 

Quibbler is correct in that EACH should appear independently of a previous outcome, and  that is how it should be.

 

Binomial distribution will tell stats based on the number of tries to see if the game is designed to allow people to be able to get enough of the drops from the boxes. It is a part of probability analysis, and especially useful in computers where an RNG is NOT infallible and certain factors such as the computers clock, flops, etc can affect the outcome of the RNG such that at a certain time of day or server load, it could more often come up with one result over the others.

 

I have been doing this with both RPGs and computers since 1982, so i know what all the math is, just didn't want to drag it all out for the masses of non-English speaking persons on the forum, nor those players that may not have reached this education level yet to get bogged down with it.

 

But I can break out full graphs to compare if NCSoft wants to do the same to actually should their distribution ratio of drops in not only the weapon chests, but ALL other RNG "boxes" including loot auctions. I doubt that confidential intellectual property and "trade secret" is something NCSoft wishes to release.  :) (and I frankly don't require it but figuring the math beyond  basics does require it)

[TheQuibbler]

3 hours ago, Shadzar said:

Quibbler is correct in that EACH should appear independently of a previous outcome, and  that is how it should be.

Unless I am missunderstanding the CM, that is also excatly what he is saying is going on inside the game.

Each class weapon has 1/8th chance to drop, what he is using the bionomial distribution for, is just to show people here, when you theoritcally should have seen the item drop, or how unlucky you need to be, to not see it for 34 tries.

 

Quote

In probability theory and statistics, the binomial distribution with parameters n and p is the discrete probability distribution of the number of successes in a sequence of n independent yes/no experiments

However, I do see how this could be confusing for some players on the forum. Making it seem like they should have 100% chance to see an item after x amount of kills, or chests, or whatever is in question.

[Shukran]

numbers have no memory.