How "honest" do you feel the RNG in POE actually is?

So many "experts" with tinfoil hats on this forum recently.

"

beaniac wrote:

The only way a seed can cause changes to drop rates is if the same seed is used over and over again every time the random number generator is instantiated. In this case you will get exactly the same results every time, not some nebulous change in the chances of getting good gear. Also any first year computer science student knows not to do this, it's a stupid noob mistake. Though it does sound like that is what happened with the resonator bug.... -_- You're kill'n me here GGG

I feel like you need to get through comp sci get your first job in game development and find out how many dev studios use players account numbers as part of their seed, or "master seed" which is used to spin off more temporary seed numbers so that they keep an identifier that helps with tracking and anti cheating measures. Like figuring out what account made a bunch of dupes.

GGG has stated publicly that they can and do "turn down" the loot of known botters this means our accounts contain a variable that can lower loot drops. it would be the easiest thing in the world to manipulate this variable conditionally for the purpose of extending "player retention". Despite some humorous attempts by some to suggest that this requires mind reading, just having sets of conditions that automatically adjust the variable (such as x numbers of card found within X time frame too quick = lower the drop rate) or even making these conditions only apply to things that are unlikely to be statistically analyzed by the player base. So for example cards that are going to be farmed over and over by the same people regardless of if you've found a set already are very likely to have their drop rates studied and published. Extreme outlyiers like going from ~30 minutes per card for 5 cards and then 22 hours for the last card are much less likely to be seen as suspicious for sets that will be one off Cards that you're only ever going to farm one or maybe 2 sets of...(which BTW 22 hours was how long it ended up being it took 3 hours to farm the first 5 and 22 to farm th last, this much of an extreme outlier is theoretically possible its ludicrously unlikely and in fact suspicious given how many other times I've experienced "extreme theoretical outliers") these are prime candidates to have conditional variables put in place. Give the player a taste of success, let them see the end goal and then make them grind. I know, I know attributing this sort of dirty "player retention" tactic to GGG as beloved as they are is heresy (Hell even I think they are the best game studio I'v ever experienced in nearing 30 years of PC gaming but I'm not going to call them infallible or saintly).

They also admitted publicly that loot seeds can be bugged to have dramatically unfair outcomes. AKA the Delve 4 socket resonator debacle where some players found them multiple times a night while others found none.

Lost in the loot seed debacle was the very clear implication: TOO MUCH VARIABILITY is "bugged behavior" but "behavour" from a loot seed implies that there is some designed in variability inherent to the system; In other words they themselves implied that we don't all have the same drop chance for 4 socket resonators, OR an exalt when everything else is equal we have some "behavour" in our loot seeds that gives variability. We aren't All playing on an even playing field.

"Tin foil hat" "you're just paranoid" "why would they even do this". Well have you never heard Chris talk about "multi axis randomness"? AKA RNG within RNG? What is more mutli axis RNG than having your base chance of a drop be a variable? Its also wonderful for your player retention metrics: do players with "good loot seeds" play longer or shorter than players with Bad loot seed? You might think thats obvious but is it? In a game where some people choose to play SSF or private leagues to extend play? For a company so obsessed with retention and studying metrics when literally their lives work and paying their rent from month to month relies on knowing the answers to "who spends more money over the longest time frame: unlucky players or lucky players?". If I was coding a game in a garage I would absolutely want to be able to scientifically ask those questions, and all I would have to do is put a variable into the game that allowed me to easily ask and answer.

"

beaniac wrote:

"

DarthSki44 wrote:

"

beaniac wrote:

You guys do know that Malichus was being facetious right?

It is worth noting that GGG does not publish drop rates, and it would be silly to think that they never adjust them.

It's even more silly (to the point of paranoia) to think that they change drop rates based on account parameters. The real truth is that if they increased drop rates the people who would be the first to quit are the players who play 16hrs a day. The people who spend the most on the game tend to be the people who enjoy the challenge of it.

That's not entirely true. Chris has admitted to adjusting rates based on player count and retention.

Also it's been revealed that GGG uses seeds to control drops.

https://www.reddit.com/r/pathofexile/comments/9gcclj/that_talk_of_rngseeds_has_put_me_on_edge/

Ok, let's be clear here. Adjusting drop rates for general balancing (balance is connected to retention rates) reasons falls into the first thing I said, and is very different from changing drops on an individual account basis.

Secondly, all random number generators use a seed. That is how the underlying math works. Generally that seed is set at runtime or when a new routine is instantiated which means at the very least every time you restart the game you get a new one. And beyond that the value of the seed being used will not have any effect on your drop rate as that value is set externally to the random number generator.

The only way a seed can cause changes to drop rates is if the same seed is used over and over again every time the random number generator is instantiated. In this case you will get exactly the same results every time, not some nebulous change in the chances of getting good gear. Also any first year computer science student knows not to do this, it's a stupid noob mistake. Though it does sound like that is what happened with the resonator bug.... -_- You're kill'n me here GGG

Making a seed based on a character Id is dumber than dumb and this comes from a long time developer. Unless the system clock can't be used as a seed for some reason, which I have no idea why it could be used in PoE, there are other ways to make the seed less deterministic. On top of this, usually you want to reseed as often as possible if you are trying to be as random as possible, unless you are trying to do weighting. Now that later part is the key. If you are using a deterministic seed, it is because you are trying to apply a "Weight" to the generator. That could be for good or bad reasons. In game design this is a good thing in many games. For example making it less likely a player gets hit by attacks as they get lower in life but keeping that fact from the player. It gives a sense of epic-ness to the fights to feel like you just came through by the skin of your teeth. The real truth is that as the fight progresses, the "random" attacks of the fight started getting weighted in favor of the player as the fight got longer.

When you tie a seed value to an account or character ID... you can introduce weights that could be used for bad design.

The variance between those in my group of players that get a lot of loot and those that don't is extremely consistent for 7 years now. To a point that one of our guild members plays 2-3x every league versus one of our guys that plays pretty casually. In this case there has never been a league where the casual player did not have more EX, better map drops/sustainability, etc. Even played the same builds and honestly the 2-3x guy is a very good player with generally 9+ 95+ level characters and usually a couple of 100 level builds.

We can pretty much go into any league and know who will have what drops. We have even charted who would get multiple of what uniques, so that we can plan our builds. Never has failed.

So, I do not pretend to know what or how they do their RNG. However it is a vast valley between those that have good RNG and those that don't. With 7 years of data, it makes it pretty crazy.

GGG is a business. They have a free to play game and still need to make money. They are not immune to the less than ethical practices that all FtP games have to deal with. There is a reason seminars like this exist:

https://www.youtube.com/watch?v=xNjI03CGkb4

Literally teaching companies how to maximize profits in the FtP business model of gaming. Some game companies go for full maximizing of short term profits and some slight press the issue.

From a business stand-point it makes absolute perfect business sense to tweak RNG for some accounts and not others. Streamers and other highly visible PR agents of the game are basically ambassadors that bring in more players. More players means more chances to make money. It makes sense to tweak the RNG slightly in their favor.

Same thing goes for big times whales. Someone dumping hundreds or thousands on support packs it would be a good business idea to give them a slightly more favorable gaming experience. They are likely to spend more if they enjoy the game. That is what WHALING means in the gaming business industry. Trying to land those whales that will spend a lot.

Again, it can be a variable tweak that changes itself from time to time. A famous streamer can have the RNG tweak in their favor for a 5 days out of the week and then goes the other way for 2 days. Makes it seem more "fair" to the viewers and still helps to create a sense of surprise for the players.

It also makes sense to make RNG slightly worse from time to time for non-playing characters with an occasional "pop" that gives them a very huge high when it happens. It makes people think, "Wow! that just happened! Maybe I really like this game enough to buy something now!"

I am not accusing or saying GGG is doing any of these things. I am saying that these things are done and are well known business practices in the industry. If done right it is very hard to prove. There is no laws right now regulating anything like this and most gaming companies are fighting tooth and nail to prevent any such laws from coming to fruition as well. I am saying that as long as there are no laws to break, companies looking to maximize profits will do things some people find less than ethical.

"

Viscar wrote:

Making a seed based on a character Id is dumber than dumb and this comes from a long time developer. Unless the system clock can't be used as a seed for some reason, which I have no idea why it could be used in PoE, there are other ways to make the seed less deterministic. On top of this, usually you want to reseed as often as possible if you are trying to be as random as possible, unless you are trying to do weighting. Now that later part is the key. If you are using a deterministic seed, it is because you are trying to apply a "Weight" to the generator. That could be for good or bad reasons. In game design this is a good thing in many games. For example making it less likely a player gets hit by attacks as they get lower in life but keeping that fact from the player. It gives a sense of epic-ness to the fights to feel like you just came through by the skin of your teeth. The real truth is that as the fight progresses, the "random" attacks of the fight started getting weighted in favor of the player as the fight got longer.

When you tie a seed value to an account or character ID... you can introduce weights that could be used for bad design.

You also introduce the ability track transaction strings or potentially RMT of items, detect duping and who did it, and create scientific metrics that allow you to tune your game for maximum player engagement and retention.

Using a players identifier in loot generation and storing it in the item database dates back to the late 90's post Ultima Online and the massive duping issues they had that permanently ruined their live economies. It is used in many games, and in fact we know 100% for sure that something like it is in POE because of the Delve loot seed bug but also because GGG has stated they turn down the loot of botting accounts (which requires a variable assigned to the account that can then be "turned down") which in itself is proof that the game references an account wide variable while generating loot.

it could be as simple as the first 2 digits of everyones account ID/master loot seed being a base quantity variable that can differ for be edited manually or dynamically through automated conditions.

"

DS_Deadman wrote:

The variance between those in my group of players that get a lot of loot and those that don't is extremely consistent for 7 years now. To a point that one of our guild members plays 2-3x every league versus one of our guys that plays pretty casually. In this case there has never been a league where the casual player did not have more EX, better map drops/sustainability, etc. Even played the same builds and honestly the 2-3x guy is a very good player with generally 9+ 95+ level characters and usually a couple of 100 level builds.

We can pretty much go into any league and know who will have what drops. We have even charted who would get multiple of what uniques, so that we can plan our builds. Never has failed.

So, I do not pretend to know what or how they do their RNG. However it is a vast valley between those that have good RNG and those that don't. With 7 years of data, it makes it pretty crazy.

I also think its interesting how amazing the RNG of UberDan who seems to have no struggle hitting almost anything he needs when compared to say RaizQT (who always struggles but never moreso than this league). There's consistency here, both players play an immense amount, one of them has RNG struggles I recognize from my own gameplay experience, the other is playing a version of POE I don't recognize at all.

"

alhazred70 wrote:

"

Viscar wrote:

Making a seed based on a character Id is dumber than dumb and this comes from a long time developer. Unless the system clock can't be used as a seed for some reason, which I have no idea why it could be used in PoE, there are other ways to make the seed less deterministic. On top of this, usually you want to reseed as often as possible if you are trying to be as random as possible, unless you are trying to do weighting. Now that later part is the key. If you are using a deterministic seed, it is because you are trying to apply a "Weight" to the generator. That could be for good or bad reasons. In game design this is a good thing in many games. For example making it less likely a player gets hit by attacks as they get lower in life but keeping that fact from the player. It gives a sense of epic-ness to the fights to feel like you just came through by the skin of your teeth. The real truth is that as the fight progresses, the "random" attacks of the fight started getting weighted in favor of the player as the fight got longer.

When you tie a seed value to an account or character ID... you can introduce weights that could be used for bad design.

You also introduce the ability track transaction strings or potentially RMT of items, detect duping and who did it, and create scientific metrics that allow you to tune your game for maximum player engagement and retention.

Using a players identifier in loot generation and storing it in the item database dates back to the late 90's post Ultima Online and the massive duping issues they had that permanently ruined their live economies. It is used in many games, and in fact we know 100% for sure that something like it is in POE because of the Delve loot seed bug but also because GGG has stated they turn down the loot of botting accounts (which requires a variable assigned to the account that can then be "turned down") which in itself is proof that the game references an account wide variable while generating loot.

it could be as simple as the first 2 digits of everyones account ID/master loot seed being a base quantity variable that can differ for be edited manually or dynamically through automated conditions.

That are other ways to track all that, which is why you have a character Id in the first place. You don't need to use it as a seed or salt for a seed at all.

From my experience the seed has to be based on the system clock. Reason being, when I roll items I'll more often than not see the same mod twice or the same combination of mods twice or thrice in a row when I go quick. When I go slower I don't, so that's suspicious.

However, in regards to "is our RNG rigged" I have to say: no - and if you do believe it is, you should find a better use for your tinfoil hat.

"

ArtCrusade wrote:

I have to say: no - and if you do believe it is, you should find a better use for your tinfoil hat.

Why? Because you say so? Provide a rationale beyond "everything that sounds tin foil hat must be false".

I understand skepticism but have a rational basis for it right?

Looking past GGG admitting they can and do manipulate drops on an account by account basis. Looking past GGG accidentally doing exactly that due to a Bug with their loot seeds. Looking past really obvious motives that are in fact industry wide practices and even things other game devs have been caught doing such as Blizzard with Raid drop RNG in WoW that was found out to be unfairly distributed, denied but then later admitted to (and who knows where else in WoW such code has been used).

Its not reasonable to dismiss all this with "meh I don't like it and i don't want it to be true, so I disbelieve". You can doubt all you want you can disbelieve all you want but if you're going to argue, at least have a rational reason backing you up.

Right now we have empirical evidence of unintentional fuckery (Delve loot seed bug), we have admitted capability (anti botting measures). We have multiple good motives, including marketing the game, attracting players with immediate success and then stretching it out; tuning the item economy; gathering scientific metrics etc. etc. and we have lot of plausible reasons including bugs, industry practices and past industry examples of this sort of fuckery happening. And maybe most of all we have Chris Wilson publicly talking about how much he loves "multi axis randomness" which a drop rate variable in our loot seed that we KNOW exists (but in fairness we don't know if it varies outside of botters getting it turned down) certainly plays right into.

I think skepticism is healthy but in the face of evidence outright dismissal is just naivety.