Dear ZOS, lets finally talk about RNG and the drop system

GilGalad · May 2017

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

Please stop referring to it as RNG cause nothing about it is random as you and hundreds of others have shown. It's a loot table

It is a loot table with drop chances for every item, say two items with 40% and 60% drop chance. When you now take your loot the game will calculate a random number between 0 and 1. if it is within 0.0-0.4 you get item one, if it is within 0.4-1.0 you get item 2. So the loot table only determines the range within a random number has to be in order for you to get the item, but the system itself is based on a random number generator.
Where have I shown that this system is not random???
Show me 2 people with exactly the same loot every time...

@GilGalad

No......that's why I'm saying stop calling it RNG
It's not a random number, it's a set and filtered result by location, NPC, etc so the results within that loot table aren't random.

It's very obvious from a few very detailed threads, the loot possibilities are pre set.
Basically it's if this and then that player 0% of this, has 20% to get this with x% of said traits, 60% to get this with x% of said trait but if one previous happens, it excludes any other result due to how the tiered and filters work in their tables.

The traits aren't even random as they also have pre set % chances so it's misleading to write RNG cause their tables are a few pre-set chances with specific excluded chances of other items.

If it were random then those who ...for example....ran vMSA over 1,000 times would more than likely have found their desired item but that's not the case. It was either in the forums or on ESO Live way back over a year ago when they talked about adjusting their loot tables and they literally talked about certain items and certain traits having pre-set drop chances set but where if x happened y couldn't. That's not random at all

The way it works is...
Rejuvenate staff may have an 80% drop rate and powered trait may have 5% chance but the rejuvenate staff and powered trait are only possible if the other staves don't drop.

Ok....
So then you realize each staff has its own different % but can only drop from one NPC or chest.

Ok...
So then there are those other weapons and each have their own % chances with differing trait % chances

Ok....
So then you realize ZOS has changed the % of each for all more than once since vMSA released.

Ok...
Then for set items.....limited to X dungeon or X location above ground and only drops from 1,2 or 3 NPCs but in a group each player has their own different chances but can share loot drops.

Ok....
So then player decide to run a vet dungeon for one specific helm with 3 others and they run it for a month 4-5 times each day and out of all those runs the helm with their trait never drops for any player.

So.....if a specific weapon is chosen which can be argued as RNG, the probability of that weapon being selected over other weapons now goes into another filter where the trait has a % among the other existing lootable traits. Some have all while others have been removed but where the % of each all differ.

More clear is a key drops and a player takes 100 keys and opens the same chest for monster shoulders but the shoulders can be argued as random but the traits for the random shoulder aren't random as they too are pre-set.

I know the item and the traits are pre-set on a % which isn't 1-100% among other gear of the same type

How is that random?
It's a loot table or what's described as games of chance not random, right.
That's why I'm saying stop calling it RNG

Chance can't be calculated to a % where random should be calculated to a %

At least thats how I discuss the two

1. Every chance can be calculated to a %.
2. Random is nothing you express in % its a characterization of a process. If you have a process with a fixed input (open loot box) and non-fixed output (different possible items) its a random process. How often you get an item after a large number of repetitions of the process is determined by a prefixed probability, the "chance to get the item". As long as you don't know what you are looting before you loot it, it is a random process and therefore related to RNG.

No it can't, that was the point of that long wall of text. That makes it chance which is explained similar but different due to not being able to calculate it.

Each person can quantify their chances results but there isn't a standard % that applies to all
It's a chances game of chances with unlimited random results.

The point of your long wall of text is that you have a different thinking about RNG, random processes and chances than me and probably everyone I met during my physics master's program. Since nothing of that contributes to the topic, let's just agree to disagree.

@GilGalad
No disrespect to your education, the instituation or your colleagues....but I'm suggesting the way you're describing it, isn't how it works.

I don't disagree or want to challenge your education or level of understanding this as your OP exhibits a great degree of understanding RNG.
I'm simply pointing out, it's not truly random by development intent.

If a dice has 12 sides, and you roll it 100 times, each side always has the same chance of landing up. That you can calculate and regardless of who is rolling or how they roll, the chance of each side always remains the same...again...that which can be calculated.

The results, may differ but that's not what we are talking about because while people don't like their results, the concern isn't the results, it's the system in place that predicates the "possible results".

ZOS by design has a 12 sided dice where each side has a pre-set chance of facing up. On top of that, they decided that this loot system wouldn't have the same dice roll with those pre-set chances per side. They have multiple dice, some with 12, some with 4, some with 16 and all have pre set chances per side.

On top of that, they have another dice again, with pre set side chances that determines which dice has a chance to be rolled.

That isn't random at all because the possibilities change every time and with or less possibilities which removes any opportunity to calculate accurate results which is why so many players have extremely different possibilities , opinions and time spent for loot

Rolling a dice is one of the most famous random processes of all, even with different weights for each side. Rolling a dice and rolling another one based on the previous result is still a random process and can be calculated perfectly with a tree diagram. Hell even if the probabilies chance with every try its still a random process.

The reason why you have such high differences between people with 5 runs and others with 500+ is because the drop chance for single items is very low and there are always some that are lucky. You can calculate the standard deviation and such easily if you assume a normal distribution (see central limit theorem). If you did that you would also see how far the edges go and that it is easily possible to have people with 5 and others with 500+ runs.

NewBlacksmurf · May 2017

GilGalad wrote: »

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

GilGalad wrote: »

NewBlacksmurf wrote: »

Please stop referring to it as RNG cause nothing about it is random as you and hundreds of others have shown. It's a loot table

It is a loot table with drop chances for every item, say two items with 40% and 60% drop chance. When you now take your loot the game will calculate a random number between 0 and 1. if it is within 0.0-0.4 you get item one, if it is within 0.4-1.0 you get item 2. So the loot table only determines the range within a random number has to be in order for you to get the item, but the system itself is based on a random number generator.
Where have I shown that this system is not random???
Show me 2 people with exactly the same loot every time...

@GilGalad

No......that's why I'm saying stop calling it RNG
It's not a random number, it's a set and filtered result by location, NPC, etc so the results within that loot table aren't random.

It's very obvious from a few very detailed threads, the loot possibilities are pre set.
Basically it's if this and then that player 0% of this, has 20% to get this with x% of said traits, 60% to get this with x% of said trait but if one previous happens, it excludes any other result due to how the tiered and filters work in their tables.

The traits aren't even random as they also have pre set % chances so it's misleading to write RNG cause their tables are a few pre-set chances with specific excluded chances of other items.

If it were random then those who ...for example....ran vMSA over 1,000 times would more than likely have found their desired item but that's not the case. It was either in the forums or on ESO Live way back over a year ago when they talked about adjusting their loot tables and they literally talked about certain items and certain traits having pre-set drop chances set but where if x happened y couldn't. That's not random at all

The way it works is...
Rejuvenate staff may have an 80% drop rate and powered trait may have 5% chance but the rejuvenate staff and powered trait are only possible if the other staves don't drop.

Ok....
So then you realize each staff has its own different % but can only drop from one NPC or chest.

Ok...
So then there are those other weapons and each have their own % chances with differing trait % chances

Ok....
So then you realize ZOS has changed the % of each for all more than once since vMSA released.

Ok...
Then for set items.....limited to X dungeon or X location above ground and only drops from 1,2 or 3 NPCs but in a group each player has their own different chances but can share loot drops.

Ok....
So then player decide to run a vet dungeon for one specific helm with 3 others and they run it for a month 4-5 times each day and out of all those runs the helm with their trait never drops for any player.

So.....if a specific weapon is chosen which can be argued as RNG, the probability of that weapon being selected over other weapons now goes into another filter where the trait has a % among the other existing lootable traits. Some have all while others have been removed but where the % of each all differ.

More clear is a key drops and a player takes 100 keys and opens the same chest for monster shoulders but the shoulders can be argued as random but the traits for the random shoulder aren't random as they too are pre-set.

I know the item and the traits are pre-set on a % which isn't 1-100% among other gear of the same type

How is that random?
It's a loot table or what's described as games of chance not random, right.
That's why I'm saying stop calling it RNG

Chance can't be calculated to a % where random should be calculated to a %

At least thats how I discuss the two

1. Every chance can be calculated to a %.
2. Random is nothing you express in % its a characterization of a process. If you have a process with a fixed input (open loot box) and non-fixed output (different possible items) its a random process. How often you get an item after a large number of repetitions of the process is determined by a prefixed probability, the "chance to get the item". As long as you don't know what you are looting before you loot it, it is a random process and therefore related to RNG.

No it can't, that was the point of that long wall of text. That makes it chance which is explained similar but different due to not being able to calculate it.

Each person can quantify their chances results but there isn't a standard % that applies to all
It's a chances game of chances with unlimited random results.

The point of your long wall of text is that you have a different thinking about RNG, random processes and chances than me and probably everyone I met during my physics master's program. Since nothing of that contributes to the topic, let's just agree to disagree.

@GilGalad
No disrespect to your education, the instituation or your colleagues....but I'm suggesting the way you're describing it, isn't how it works.

I don't disagree or want to challenge your education or level of understanding this as your OP exhibits a great degree of understanding RNG.
I'm simply pointing out, it's not truly random by development intent.

If a dice has 12 sides, and you roll it 100 times, each side always has the same chance of landing up. That you can calculate and regardless of who is rolling or how they roll, the chance of each side always remains the same...again...that which can be calculated.

The results, may differ but that's not what we are talking about because while people don't like their results, the concern isn't the results, it's the system in place that predicates the "possible results".

ZOS by design has a 12 sided dice where each side has a pre-set chance of facing up. On top of that, they decided that this loot system wouldn't have the same dice roll with those pre-set chances per side. They have multiple dice, some with 12, some with 4, some with 16 and all have pre set chances per side.

On top of that, they have another dice again, with pre set side chances that determines which dice has a chance to be rolled.

That isn't random at all because the possibilities change every time and with or less possibilities which removes any opportunity to calculate accurate results which is why so many players have extremely different possibilities , opinions and time spent for loot

Rolling a dice is one of the most famous random processes of all, even with different weights for each side. Rolling a dice and rolling another one based on the previous result is still a random process and can be calculated perfectly with a tree diagram. Hell even if the probabilies chance with every try its still a random process.

The reason why you have such high differences between people with 5 runs and others with 500+ is because the drop chance for single items is very low and there are always some that are lucky. You can calculate the standard deviation and such easily if you assume a normal distribution (see central limit theorem). If you did that you would also see how far the edges go and that it is easily possible to have people with 5 and others with 500+ runs.

Man.....you actually understand this stuff far better than most but you have overlooked the most important part.
RNG does not mean the same as random.
RNG or the point of such a system is so the results cannot be reasonable predictable.....

Even if dice is weighted, that same dice is always weighted in the same manner but that's not this system.

Let go of the loot results and concentrate on the loot system or that which controls the possibilities of those results. That's the conversation I'm having with you....

If the possible results change each time, that's not random when the reason they change is based on designed pre-sets that add or eliminate the next tier of chance.

Lottery tickets....
Let's say a lottery has 17 numeric and alpha combinations and each week the winner has to align with a always changing result. That's something you can calculate but that's not this system.

Kill Boss in ESO and it does this:
Roll part 1
1-chance for getting helm, weapon, ring, trash
2-chance of getting helm, ring, trash and no weapon
3-chance of getting helm only, and no trash or other items
........that's tier 1 of the dice but the sides all have different pre-set % to be chosen but it's not out of 100% as each has its own 1%-100% chance

Roll part 2
1-if helm chance is won cause it's 100% the next roll is helm type heavy, medium, light
1-b if helm chance is won cause it's 100% the next roll is which monster helm less 0% pre-set chances
1-c if helm chance is won cause it's 100% and the above rolls pick any one of three pre-sets influenced by any one of many possible monster sets...then

Roll part 3
1- chance of trait based on pre-set trait chances which are influenced by roll 2

..... that's what happens to each individual item per kill......

That's not random at all, it's a tiered loot table system

Didgerion · May 2017

RNG is sooo wrong!!!

You just shrink your player base with it.

I remember @FENGRUSH's frustrations with his countless VMA runs for his weapon.

Let me tell you ZOS - @FENGRUSH is a dedicated player with lots of patience for this game - also he is a streamer and he goes a little bit over his patience limitations to not disappoint his community.

Do you really thing that all ESO players are as patient and dedicated? How can you let an RNG system control your end-game gear?

Are you sure your RNG system is targeting players right? Are you sure that your RNG gives the least patient player item faster?

You have to keep all the players in the game the patient and impatient ones!!

Please design a healthy gear drop system!

Violynne · May 2017

NewBlacksmurf wrote: »

RNG does not mean the same as random.

Yep, but...

RNG or the point of such a system is so the results cannot be reasonable predictable.....

You have this backward.

The purpose of RNG is to have a reasonable predictable result, which is why the "G" part, generator, accepts parameters which limits the random part of the attempt.

So far, the discussion has limited our examples to the 6-sided dice. If a generator's purpose is to simulate the methodology, by not having a reasonable prediction could lead to infinite results (as clearly exampled by the loading screen

)

Assume, for example, we're looking for a "5", but the generator can return any number of possibilities (true randomness). The calculated result of hitting the "5" would be impossible to obtain since there is no limit on the generator (both positive and negative).

The CLT graph link is the more accurate representation of the RNG system at play. Theoretically, it is quite possible for a result to return "0" (or non-match) in every attempt, while the predictable results will show more would hit the return than not.

Computers cannot perform true random sequences. They require a generator. If we use one to example a 6-sided dice roll over a span of 100 iterations, it should come extremely close to matching the same results if done through mechanical means or human generation (which would take into consideration those who try to "cheat" to roll specific numbers by not thoroughly performing the "generator" part - watch clever Craps players for examples of this).

NewBlacksmurf · May 2017

Violynne wrote: »

NewBlacksmurf wrote: »

RNG does not mean the same as random.

Yep, but...

RNG or the point of such a system is so the results cannot be reasonable predictable.....

You have this backward.

The purpose of RNG is to have a reasonable predictable result, which is why the "G" part, generator, accepts parameters which limits the random part of the attempt.

So far, the discussion has limited our examples to the 6-sided dice. If a generator's purpose is to simulate the methodology, by not having a reasonable prediction could lead to infinite results (as clearly exampled by the loading screen )

Assume, for example, we're looking for a "5", but the generator can return any number of possibilities (true randomness). The calculated result of hitting the "5" would be impossible to obtain since there is no limit on the generator (both positive and negative).

The CLT graph link is the more accurate representation of the RNG system at play. Theoretically, it is quite possible for a result to return "0" (or non-match) in every attempt, while the predictable results will show more would hit the return than not.

Computers cannot perform true random sequences. They require a generator. If we use one to example a 6-sided dice roll over a span of 100 iterations, it should come extremely close to matching the same results if done through mechanical means or human generation (which would take into consideration those who try to "cheat" to roll specific numbers by not thoroughly performing the "generator" part - watch clever Craps players for examples of this).

@Violynne

O K so I have the understanding of the term RNG and it's use wrong.

So you and others and saying the results are predictable......but then the community all has very different results

Can you talk more about that?
Honest question because as you see above, my understanding is that the results aren't predictable and causing endless posssibilities

I'm not arguing that the graph is wrong, I'm saying the methodology to how the graph was generated is creating an incorrect understanding or possibly creating an assumption that doesn't align with actual results and how ZOS talks about it

If we all used the same process, we would have different charts but no better results, so something is off

ButtersEP · May 2017

There really needs to be a token system, an easy way could be a merchant in each dungeon or trial that obviously sold that dungeon or trial gear, today I just finished a 3 day grind for automaton BODY pieces, it took 3 days to get a good chest to boots in the right traits... talos have mercy on those that grind the weapons....

amasuriel · May 2017

The saddest part about all this is, we have super dedicated players like you and Alcast, and lost of other amazon folks on this forum. They provide math, graphs, analysis of the issue. Something like a token or currency system they already have in game, it's not technically complicated.

We get nothing in response. 1 ESO live comment that will never be spoken of again in all probability.

It's the same with housing. All that work, largely pointless because writes / horse trainer / banker means you have no reason to visit your house except for target dummies and roleplay.

Same with Traits. Some are harder to balance, some are just math that people have shown again and again.

Same with under-performing sets. Some need complex changes, some just need simple number tweaks.

Probably 1 decent programmer could fix some huge pain points in 3-4 weeks, but instead we will get silence, and some more broken features.

Anyway, I'll stop ranting, I just find it so depressing how much pain is being caused by issues that are not complicated to finish. There must be some pointy haired bosses at ZOS having a gleeful chuckle.

Feanor · May 2017

@GilGalad

You know they implemented "smart loot"? And all this does is removing a weapon from the loot table if it "cannot proc the 5pc bonus of the specific set in any circumstance"? That's why we have so many stam weapons on Magicka sets and vice versa. It's just bad design.

coplannb16_ESO · May 2017

So I get an infused Inferno staff on my third run....

17% stronger enchantment is okay, but the 40 % cooldown reduction are wasted on like all of those weapons. Why is this even in the loottable?

Violynne · May 2017

NewBlacksmurf wrote: »

So you and others and saying the results are predictable......but then the community all has very different results

Can you talk more about that?

I wish I could, honestly, but there's no way to know for 100% sure how the numbers are weighted by ZoS.

But I have an extremely strong opinion one of the factors is the actual number of days since last "drop".

I pretty much collect every Runestone I find, and when I approach a node, I sing a little jingle in my head "Kuta! Kuta! Kuta! Gooooooo Kuta!"

Ta.

But, here's the kicker: I happened to notice I get far more Kuta returns every 4 days than I do if I just ran around every single day. Moreso, in one instance, I found three Kuta stones in a row, and absurdly rare event I'll most likely never see again.

Don't take this as a fact, but if there's any substantial validity in my own observations, this means ZoS is really weighting down those returns, making it much more difficult to get the expectation the rest of us believe should occur.

Which is why I say in this game, RNG isn't here to reward us. It's here to punish us. I firmly believe this, which is why I'm never, ever going to chase dungeon gear.

I've better things to do than fight against an unfair system.

NewBlacksmurf · May 2017

Violynne wrote: »

NewBlacksmurf wrote: »

So you and others and saying the results are predictable......but then the community all has very different results

Can you talk more about that?

I wish I could, honestly, but there's no way to know for 100% sure how the numbers are weighted by ZoS.

But I have an extremely strong opinion one of the factors is the actual number of days since last "drop".

I pretty much collect every Runestone I find, and when I approach a node, I sing a little jingle in my head "Kuta! Kuta! Kuta! Gooooooo Kuta!"

Ta.

But, here's the kicker: I happened to notice I get far more Kuta returns every 4 days than I do if I just ran around every single day. Moreso, in one instance, I found three Kuta stones in a row, and absurdly rare event I'll most likely never see again.

Don't take this as a fact, but if there's any substantial validity in my own observations, this means ZoS is really weighting down those returns, making it much more difficult to get the expectation the rest of us believe should occur.

Which is why I say in this game, RNG isn't here to reward us. It's here to punish us. I firmly believe this, which is why I'm never, ever going to chase dungeon gear.

I've better things to do than fight against an unfair system.

@Violynne

Gotcha...and I'm not suggesting I've done enough tracking at the level of the OP but I do know because of older ESO Live show that their system does eliminate things from the loot table based on certain criteria or circumstance so there's no way anyone can calculate or come up with any predictable % for specific loot drops.

The other thing is.....
The different variables for nodes to dungeon loot, to vMSA to Dolemans, etc. it's all different.

I have noticed and would agree that number of days is involved but that may also align with others who suggest the loot results could be a large table and anytime X drops, it's removed until Y happens and so on and so forth at different rates for Z timeframes per server rather than per player or at X%

I just need people to stop thinking it's supports to be predictable or something like Powered heal staff drops 40% and light staff precise drops 27% cause it definitely doesn't work like that

Violynne · May 2017

NewBlacksmurf wrote: »

I just need people to stop thinking it's supports to be predictable or something like Powered heal staff drops 40% and light staff precise drops 27% cause it definitely doesn't work like that

Probably not going to happen, but I understand.

The majority of people may not get "numbers", but there's no denying when they know something is wrong. It's hard to be upset with them when they can't prove their feeling.

We've all been there, which ironically, makes us the majority to know RNG is massively broken as a whole.

NewBlacksmurf · May 2017

Violynne wrote: »

NewBlacksmurf wrote: »

I just need people to stop thinking it's supports to be predictable or something like Powered heal staff drops 40% and light staff precise drops 27% cause it definitely doesn't work like that

Probably not going to happen, but I understand.

The majority of people may not get "numbers", but there's no denying when they know something is wrong. It's hard to be upset with them when they can't prove their feeling.

We've all been there, which ironically, makes us the majority to know RNG is massively broken as a whole.

Rittings · May 2017

I maintain a token system would be best... one for dungeons, one for trials and one for overland sets...

You could spend tokens to gamble for trait - or save up a LOT of tokens to get the specific trait.

GilGalad · May 2017

Feanor wrote: »

@GilGalad

You know they implemented "smart loot"? And all this does is removing a weapon from the loot table if it "cannot proc the 5pc bonus of the specific set in any circumstance"? That's why we have so many stam weapons on Magicka sets and vice versa. It's just bad design.

I think they only restricted the drops for sets like hawk eye, which are only active with a specific weapon equipped. Since it is technically possible to proc sets like BSW with a bow equipped they did not change that, which is really frustrating.

NewBlacksmurf wrote: »

I just need people to stop thinking it's supports to be predictable or something like Powered heal staff drops 40% and light staff precise drops 27% cause it definitely doesn't work like that

Violynne wrote: »

Probably not going to happen, but I understand.

The majority of people may not get "numbers", but there's no denying when they know something is wrong. It's hard to be upset with them when they can't prove their feeling.

We've all been there, which ironically, makes us the majority to know RNG is massively broken as a whole.

First about the statement RNG and random is not the same:
RNG stands for (pseudo) random number generator, wich is in the best case very close to real randomness. I find it very very unlikely that a broken random number generator is responsible for the variations in drops for different players. All the data I have seen so far indicates that ZOS is using a fixed loot table with equal weights for traits, weapon types and shoulders.

The reason why most people don't see it is because no one is even close to a reasonable sample size to derive the drop chances. When you have a chance of 1/288 (undaunted shoulders) you would need at least 10k observations to get somewhat close to the underlying drop chances.
You don't see many people with highly different drops because the system is rigged or wrong or because you don't believe in fixed probabilities. But you see so many different drop distributions because the drop chances for the items are so insanely low, and the loot tables are huge leading to big deviations.

Just for the fun a distribution of 10k undaunted shoulder drops using a random number generator:
https://docs.google.com/spreadsheets/d/11zZoUgkeFXJez-HHUhOSjT4Svb7fwrVeuWLLUIUKE40

Does this look like an equal distribution?
No it doesn't, even though the random number generator used equal probabilities. So you clearly see that 10k observations is way to low to get even close to the average!

I also simulated 288 draws 3 times and these were the number of unique shoulders that were missing:

Test 1	107
Test 2	110
Test 3	119

So after 288 keys you would still not have 1/3 of the 288 shoulders.

I am 100% sure we have ordinary loot tables with fixed probabilities for items and equal probabilities to traits, but the possible outcomes are just way too many in order for anyone to get even close to the average, meaning you only get the item you want if you are lucky, no matter how much skill you have or if you believe in RNGsus or not.

Lord_Eomer · May 2017

I accepted defeat and no longer needs lighting sharpen after 140+ VMA runs,

Just got twice infused in a row. I am not going to waste any further time!

NewBlacksmurf · May 2017

GilGalad wrote: »
Feanor wrote: »

@GilGalad

You know they implemented "smart loot"? And all this does is removing a weapon from the loot table if it "cannot proc the 5pc bonus of the specific set in any circumstance"? That's why we have so many stam weapons on Magicka sets and vice versa. It's just bad design.

I think they only restricted the drops for sets like hawk eye, which are only active with a specific weapon equipped. Since it is technically possible to proc sets like BSW with a bow equipped they did not change that, which is really frustrating.

NewBlacksmurf wrote: »

I just need people to stop thinking it's supports to be predictable or something like Powered heal staff drops 40% and light staff precise drops 27% cause it definitely doesn't work like that

Violynne wrote: »

Probably not going to happen, but I understand.

The majority of people may not get "numbers", but there's no denying when they know something is wrong. It's hard to be upset with them when they can't prove their feeling.

We've all been there, which ironically, makes us the majority to know RNG is massively broken as a whole.

First about the statement RNG and random is not the same:
RNG stands for (pseudo) random number generator, wich is in the best case very close to real randomness. I find it very very unlikely that a broken random number generator is responsible for the variations in drops for different players. All the data I have seen so far indicates that ZOS is using a fixed loot table with equal weights for traits, weapon types and shoulders.

The reason why most people don't see it is because no one is even close to a reasonable sample size to derive the drop chances. When you have a chance of 1/288 (undaunted shoulders) you would need at least 10k observations to get somewhat close to the underlying drop chances.
You don't see many people with highly different drops because the system is rigged or wrong or because you don't believe in fixed probabilities. But you see so many different drop distributions because the drop chances for the items are so insanely low, and the loot tables are huge leading to big deviations.

Just for the fun a distribution of 10k undaunted shoulder drops using a random number generator:
https://docs.google.com/spreadsheets/d/11zZoUgkeFXJez-HHUhOSjT4Svb7fwrVeuWLLUIUKE40

Does this look like an equal distribution?
No it doesn't, even though the random number generator used equal probabilities. So you clearly see that 10k observations is way to low to get even close to the average!

I also simulated 288 draws 3 times and these were the number of unique shoulders that were missing:
Test 1	107
Test 2	110
Test 3	119
So after 288 keys you would still not have 1/3 of the 288 shoulders.

I am 100% sure we have ordinary loot tables with fixed probabilities for items and equal probabilities to traits, but the possible outcomes are just way too many in order for anyone to get even close to the average, meaning you only get the item you want if you are lucky, no matter how much skill you have or if you believe in RNGsus or not.

First I hope you're taking my comments as discussion and not arguments cause that's how I intend itto come off.

Typing doesn't always do that

How can you be so sure when "ZOS" told us certain items have pre-set percentages which come after another pre-set percentage chance?

I do follow your comment and to me, you're doing the explanation in terms of math correctly however it's not conclusive that it's actually random or pseudo-random.

I know it's not an equal distribution and that was my first comment right off.
But that's why I commented about the pre-set % per items within each table for loot chances.

It's just a chance system....it's not random or RNG cause it's set up so that the odds are against everyone based on "highly desired results which makes it impossible to come up with consistent % or chance of drop calculations.

The results are endless cause if 1,000 of us did 10k chances, there should draw a % right.

So if the experiment were done again by a different group, we should begin to see trends...right? Or and I off or wrong?

GilGalad · May 2017

@NewBlacksmurf

This is my last comment about this since it does not provide anything useful to the topic apart from the fact that the deviation is so big that people don't believe in it's randomness anymore. Even when the drop chances are not equal it is a random process. It is random as long as you the player don't know 100% of the time what item you will loot next. That is the definition of randomness.

Give me a record of 50k used undaunted keys or loot of a final dungeon bosses and I can tell you the approximate drop chance for every item or check if their actual random number generator is broken (what I highly doubt). The problem is that no one goes on the forums and posts "oh I got my sharp BSW staff after 5 runs". The data you find on the forums is highly biased due to mostly frustrated people posting.

If you want to discuss this further go to my math of RNG thread you find in my signature.

kargen27 · May 2017

Gothlander wrote: »

Deltia quit because of this. Deltia was a cool dude. He will be missed.

Nah he quit because it was time for him to move on. He and others who have put as much time as he has in the game have done about all they can do and still be entertained. So they are moving to something else to fill their time. Leaving was inevitable for some of the people that have been around since the beginning they are just finding something to call the last straw to justify walking away.

Zewks · May 2017

NewBlacksmurf wrote: »

Please stop referring to it as RNG cause nothing about it is random as you and hundreds of others have shown. It's a loot table

To a casual gamer such as myself, RNG simply means this:

"I have zero control over the drops that I get, and no matter how many times I farm for a specific item, I may never actually get it."

I dont care about the true meaning of "Random" or "RNG". In fact, up until this point, RNG was just a term I figured a joking gamer coined in WoW or EQ or UO or something when they were trying to find a catch phrase term to describe the annoying loot drop system. I doubt they researched any real math behind the meaning of "Random". More likely, they just knew the same thing that I stated above: They have zero control over what drops.

All of this talk about what "Random" means and such distracts from the main point too. That point being, Players should be able to reach the goals they put in the work and effort trying to achieve. If a "Random" or "RNG" or "Loot Table" system is preventing players from realistically reaching these goals, then there is something wrong.

t3hdubzy · May 2017

As long as people keep playing the content, its working as intended.

ChunkyCat · May 2017

D0ntevenL1ft wrote: »

[

Hate to say it but I've ran VMA about 15-18 times and got sharpened greatsword , lightning, and inferno; yet I haven't seen a single bow drop yet.

Bro, I have 4 sharpened bows. No joke.

I lost count of how many runs I've done, i'm guessing I'm closing in on a hundred, and I've only seen 1 inferno staff (charged).

wildbear247 · May 2017

Great analysis OP. It's disheartening that after 3 years ZOS still hasn't implemented a better loot system like what you describe. I know they have a vision for their game & applaud them for the evolution of ESO to date. But what I really would like to see is for them to maintain a transparent & running "Top 10" list for PvE & PvP that is based on extensive player feedback, which they work towards over time in their patches. Rome wasn't built in a day...but seriously 3 years we've been giving feedback & here we are still waiting for a better loot system???

ForsakenSin · May 2017

what really PI$$@# me off ive been doing daily's for couple of weeks now with two characters and opening the same damm chest and get F#$^en everything i dont need except one piece ... in couple of weeks opening the same chest only 2 times did the Infernal Guardian drop in light and medium training , i got 3 pirate shoulders in training in a fricken row !!

Why can't i just buy the damm thing in divines by using the damm keys instead of wasting my time every single day with the GF that dose not work in order to open the chest and get crap out of it come on!

i would not even care if you had option " open light medium heavy chest" fine at least i can keep on opening the weight i need but don't give me BS in training or some other crap trait.... i don't even bother to decon anymore i just destroy it

SublimeSparo · May 2017

Lord_Dexter wrote: »

I accepted defeat and no longer needs lighting sharpen after 140+ VMA runs,

Just got twice infused in a row. I am not going to waste any further time!

I once got 11 or 13 charged weapons in a row, an that was before guaranteed drops so like 30+ runs smh

Totes-Bode · May 2017

I decline

S1ipperyJim · May 2017

If all the time people spent whining about RNG drops on the forum was spent farming for them they would have had their "BIS" weapons a long time ago

wildbear247 · May 2017

S1ipperyJim wrote: »

If all the time people spent whining about RNG drops on the forum was spent farming for them they would have had their "BIS" weapons a long time ago

Nice to get a good chuckle every now & again. Over 1000 posts from players for 3 years involving RNG and your constructive feedback on the matter amounts to "shut-up and grind".

Care to provide a counter-point to any of the excellent points made by the OP? Also if you feel that RNG is good for the game because it keeps players coming back for more, then I suggest a @Phinix1 thread in which he dismantles that idea:
https://forums.elderscrollsonline.com/en/discussion/337496/dungeon-token-system-when/p1