Loot-Theory Debunk Thread

For grinders eco matters. For People that know how loot works eco does not matter. They simply hunt only those mobs that have the loot.

This is the truth. The loot system sucks. It's not quite as bad in mining as in hunting. Which is why I gave up hunting. Don't have the peds to be a grinder.
 
You're all still not understanding the basic concept.

Do you understand how difficult it would be to track what you think they're tracking?

How about a mob, where you use two weapons to shoot it? One with 2.95 dpp and one with 2.35. And you do 50/50 damage? How about 85/15 damage?

How about 2 separate avatars who do the above?

How about 3 avatars?

How about a shared mob?

Now multiply this by the amount of looting actions happening every second; minute; hour. You think a server purchased by a low-end game developer could handle those actions?

100% this. I'm often amused and baffled at what people think MA is capable of tracking and calculating. MMOs require an immense amount of events and continuous updates and calculations just to keep the players and objects in sync. I can't imagine the backend power required if you start trying to accrue some of the data such as what you mentioned above. Especially when you consider this game was made ~15 years ago; yes, they ported the front-end engine (quite a few years ago in its own right), but their backend tech is likely still very limited and dated by today's standards.
 
well if it is simple means that eco pretty much matters. simple means hp * multiplier. high eco setup kills 2 mobs where high uneco setup only kills one. meaning you have 2 multiplier "attempts" to hit something rather than just one. you could hit big with one and bad wwith the 2 but its less likely. so eco matters and its pretty obvious. although my opinion is that luck is the most important thing in the long run.
the difference between someone getting 95% return and someone getting 98% return is maybe just one big hof in the long run. if you dont hit it then you dont get the higher return ratio. and you cant say that you need to grind longer to hit the big one. thats bullshit. you need to be lucky to hit something big. obviously. even at 50% chances sometimes you get 50 times in a row one side. id say there is some kind of min return when being quite eco that you will get in the long run but everyone who has more just has more luck.
 
I don't know if this has already been said, but imo if the knife isn't maxed you aren't testing eco vs non-eco.
 
well if it is simple means that eco pretty much matters. simple means hp * multiplier. high eco setup kills 2 mobs where high uneco setup only kills one. meaning you have 2 multiplier "attempts" to hit something rather than just one. you could hit big with one and bad wwith the 2 but its less likely. so eco matters and its pretty obvious. although my opinion is that luck is the most important thing in the long run.
the difference between someone getting 95% return and someone getting 98% return is maybe just one big hof in the long run. if you dont hit it then you dont get the higher return ratio. and you cant say that you need to grind longer to hit the big one. thats bullshit. you need to be lucky to hit something big. obviously. even at 50% chances sometimes you get 50 times in a row one side. id say there is some kind of min return when being quite eco that you will get in the long run but everyone who has more just has more luck.



That :)

+ rep
 
...[Same relatively simple loot theory repeated over and over]...
Actually I agree with you. Pretty much 100%.
I arrived at the same conclusion after a couple of years of hunting.

I have 2 reasons to believe it works like this:
- Occams razor (Well, not sure that really applies to stuff MA made, but still)
- I created a lot of games over the years, most for fun but a few commercial ones as well. If I had to code a TT loot algorytm, I would definitely not put in all kinds of weird complexity, which would make more difficult for me to control the amount the house wins. A simple mechanism based on dmg done, simulated randomness within easily defined limits and then multiplied by a world wide adjustable constant, to make it easy for me (MA) to increase or decrease the average return.

There are ofc also mobs/loot type tables, multipliers and other things taking place, but I believe the core of the loot algorytm is very simple. Complexity is unwanted and even dangerous for MA, since bugs in this could be catastrophic.
For performance reasons, complexity is also bad, but it's not the main reason why I believe it's a simple mechanism.


And my own experiences also seems to confirm that this is indeed how it works. (However, we are all often victims of our brains tendency to confirm the patterns we already imagine. That's just how we work. And that's why we so many believe that some special skill gains are omens of future globals etc.)
 
Last edited:
I've always said MA measure damage done and skill only, maybe damage received as well. The weapon you use versus your skill might 'locally' give you a percentage of damage but MA probably don't care about that.

They probably also measure repairs and pay something back on that in arrears, or if it's 'L' pay you back either auction cost of at least TT, maybe some markup on L...who knows.

Then they pay you back on the damage done over different timed loot cycles to test the strength of your ped card. Pretty simple really. highly likely less loot swings in highly hunted mobs as the loot pool can cope with it.

We know they measure damage as the RT achievements measure that. In other words RT tapped into what was being tracked and made a mission reward out of it.

So personal damage pool, is like personal loot pool, it's just wrapped up differently. ever wondered why people used to waste a ton of ped in the PVP arenas. Did they really waste it?...hee.hee.

Rick
 
Several ppl including OP seem to believe it has to be KISS ("Keep It Simple Stupid"),
  1. because complex formula requires more resources
  2. complex computations run at each looting event increases the total resource requirements exponentially
  3. and finally Occams Razor, it there is a simple way to do the same, it must be true.
So, let's debunk this. ;)

The 1st clue is the "global waves". Everyone has noticed those waves - for a while it's relatively quiet, then suddenly several globals within 20 sec, followed by another period of relative silence. You can almost see the crontab job running in the background... (I have no idea if the server runs on Linux/UNIX, prolly not, but you get the idea)
Nothing is actually calculated at each looting event, it's a fake. Well it's not entirely fake, they prolly used the actual loot algorithm but stripped down anything that requires computational power, simplified it down to the extreme.
The real algorithm only runs after certain intervals and it's essentially a compensation mechanism. That being said, it's not a fair compensation mechanism, it's so-so, fair and unfair at the same time. If the computing power requirements are essential you can set it up to run at longer intervals whenever overall system load increases.


The second clue for the contents of our black box is the security concerns. You could have been a magnificent game developer all your life and you most definitely know how "to do it right" - how to write elegant algorithms that are simple, fast and effective.
None of the games you developed was a RCE. From time to time someone figured out how to cheat the system. They got more loot with less effort, but in the end it wasn't really a big problem as all the cheater would get was a worthless "gold" (game currency). I believe you can already see where I'm going with this... ;)

I don't agree with people who say the code that controls the resources/transactions/gains/losses/balance/etc - the RCE part of the game is extremely simple. The code that controls mob AI in EU is indeed very primitive, the economy system running in the background is not.
I don't believe it's incredibly complex either. It's complex enough for it's purpose. The purpose is to make sure nobody can predict the outcome of the future events.

Y'know, the random number generator is not entirely random, right? You don't have to be a hacker to know about this. There's sequences of numbers that will repeat after some time (oversimplified but you get the point, right?). So, now, if it's really important to avoid the scenario where somebody records the system output, finds the match in the series generated by random number generator and thus will be able to predict the future outputs... you obviously have to make the system more complex, right?​

So, this now becomes our basic requirement. From this point forward we can use KISS and create the most elegant, simple and effective formula we possibly can. As long as the requirements are met, no need to make it any more complex.


Now, I have no idea how the system works or what was used as a seed but I'm trying to imagine what it could be.
What input could we use to make the system unpredictable? It has to be something nobody else can replicate or have access to. So, how about using the collective input from all the online players? You could serialize it and feed the stream back into your system. Nobody will ever be able to predict or even monitor the actions of all the human players. It's a pretty safe bet nobody can possibly "hack" this unless they can control your server or at least the whole network around it.
OK, it's a acrazy idea and prolly wrong, but I thought it's interesting... ;)
 
I don't know if this has already been said, but imo if the knife isn't maxed you aren't testing eco vs non-eco.

True

I've always said MA measure damage done and skill only, maybe damage received as well. The weapon you use versus your skill might 'locally' give you a percentage of damage but MA probably don't care about that.

They probably also measure repairs and pay something back on that in arrears, or if it's 'L' pay you back either auction cost of at least TT, maybe some markup on L...who knows.

Then they pay you back on the damage done over different timed loot cycles to test the strength of your ped card. Pretty simple really. highly likely less loot swings in highly hunted mobs as the loot pool can cope with it.

We know they measure damage as the RT achievements measure that. In other words RT tapped into what was being tracked and made a mission reward out of it.

So personal damage pool, is like personal loot pool, it's just wrapped up differently. ever wondered why people used to waste a ton of ped in the PVP arenas. Did they really waste it?...hee.hee.

Rick

False

Several ppl including OP seem to believe it has to be KISS ("Keep It Simple Stupid"),
  1. because complex formula requires more resources
  2. complex computations run at each looting event increases the total resource requirements exponentially
  3. and finally Occams Razor, it there is a simple way to do the same, it must be true.
So, let's debunk this. ;)

The 1st clue is the "global waves". Everyone has noticed those waves - for a while it's relatively quiet, then suddenly several globals within 20 sec, followed by another period of relative silence. You can almost see the crontab job running in the background... (I have no idea if the server runs on Linux/UNIX, prolly not, but you get the idea)
Nothing is actually calculated at each looting event, it's a fake. Well it's not entirely fake, they prolly used the actual loot algorithm but stripped down anything that requires computational power, simplified it down to the extreme.
The real algorithm only runs after certain intervals and it's essentially a compensation mechanism. That being said, it's not a fair compensation mechanism, it's so-so, fair and unfair at the same time. If the computing power requirements are essential you can set it up to run at longer intervals whenever overall system load increases.


The second clue for the contents of our black box is the security concerns. You could have been a magnificent game developer all your life and you most definitely know how "to do it right" - how to write elegant algorithms that are simple, fast and effective.
None of the games you developed was a RCE. From time to time someone figured out how to cheat the system. They got more loot with less effort, but in the end it wasn't really a big problem as all the cheater would get was a worthless "gold" (game currency). I believe you can already see where I'm going with this... ;)

I don't agree with people who say the code that controls the resources/transactions/gains/losses/balance/etc - the RCE part of the game is extremely simple. The code that controls mob AI in EU is indeed very primitive, the economy system running in the background is not.
I don't believe it's incredibly complex either. It's complex enough for it's purpose. The purpose is to make sure nobody can predict the outcome of the future events.

Y'know, the random number generator is not entirely random, right? You don't have to be a hacker to know about this. There's sequences of numbers that will repeat after some time (oversimplified but you get the point, right?). So, now, if it's really important to avoid the scenario where somebody records the system output, finds the match in the series generated by random number generator and thus will be able to predict the future outputs... you obviously have to make the system more complex, right?​

So, this now becomes our basic requirement. From this point forward we can use KISS and create the most elegant, simple and effective formula we possibly can. As long as the requirements are met, no need to make it any more complex.


Now, I have no idea how the system works or what was used as a seed but I'm trying to imagine what it could be.
What input could we use to make the system unpredictable? It has to be something nobody else can replicate or have access to. So, how about using the collective input from all the online players? You could serialize it and feed the stream back into your system. Nobody will ever be able to predict or even monitor the actions of all the human players. It's a pretty safe bet nobody can possibly "hack" this unless they can control your server or at least the whole network around it.
OK, it's a acrazy idea and prolly wrong, but I thought it's interesting... ;)

So Close


(More later)
 
Y'know, the random number generator is not entirely random, right? You don't have to be a hacker to know about this. There's sequences of numbers that will repeat after some time (oversimplified but you get the point, right?). So, now, if it's really important to avoid the scenario where somebody records the system output, finds the match in the series generated by random number generator and thus will be able to predict the future outputs... you obviously have to make the system more complex, right?​

I'd be surprised if there's no HRNG behind the curtain. They aren't expensive at all.

Well, if you try to mess with a HRNG, you may pick up a signal from outer space. (That book was written 54 (!) years ago; the plot starts with guys making random number tables (a commercial product at that time) from space noise and then receiving complaints along the lines of "your shit isn't random".)

So, pretty please, stop these loot theories before you've triggered an alien invasion. :laugh:
 
Last edited:
suddenly several globals within 20 sec

I haven't noticed this. At least not consistently.

Maybe I'm not looking hard enough :cool:
 
I haven't noticed this. At least not consistently.

Maybe I'm not looking hard enough :cool:
Yeah actually this is where I have to admit I can't prove it's not just another fake pattern that's in reality nothing more than just series of totally random events.
Or maybe I could prove it but I haven't tried. I mean, to figure it out one should analyze huge amounts of data on globals. That's too much like real work and I'm way too lazy for that. :)

Anyway, assuming it's not just a white noise and there actually is some kind of real pattern, I could make it easier for you by making this conspiracy theory even more complex. :wtg:

It seems the basic unit for the loot system is the spawn. It might get hard to notice the pattern if there's lot of ppl online at different spawns. It might be easier to notice how the "focus of the system" shifts to one mob, then another, then a third one. While shifting through those spawns one specific mob produces series of globals, noticeably more than other mobs.

Unless... there's nothing, just random patterns and it only seems to me there's some kind of system behind it. Idk, you tell me... :)
 
Last edited:

Your Evidence is?

FACT....damage is tracked.
FACT there's RT "achievements" confirming how much ammo you've shot.

I don't mind you disagreeing, but just saying "false" just makes you look like a total arse.

Rick
 
While I disagree with going down some of these crazy loot routes it seems are springing up, it is obvious that certain parameters are tracked (input); damage is one, just look at any "force" event. (Damage, killed, healed, etc).

It would be extremely strenuous to link these to ROI on loots when there is a far easier, mathematical way to do so.
 
I don't know if this has already been said, but imo if the knife isn't maxed you aren't testing eco vs non-eco.


Why is that? Are you saying there is some hidden magic behind being maxed on a weapon?

As far as my eyes can see, the only major difference is the ped leak on missed attacks. We could just calculate that and add it to his returns. It wouldn't change the scope of the results by much.
 
Why is that? Are you saying there is some hidden magic behind being maxed on a weapon?

As far as my eyes can see, the only major difference is the ped leak on missed attacks. We could just calculate that and add it to his returns. It wouldn't change the scope of the results by much.

+ the lower damage output on every hit because of the lower min damage
 
+ the lower damage output on every hit because of the lower min damage

Right but how is that different from using a maxed weapon with really shitty dpp?
 
My theory is a little bit different: the algorithm that determines loot is more on the simplistic side, however loot is not determined upon the act of looting, but rather assigned to each mob on being spawned. Maybe just the multiplier, or maybe the exact contents, I don't know, it's mostly irrelevant. My evidence for this is a picture taken from the test server, which I believe McCormick posted years ago, that showed mobs had an individual internal ID number assigned to each. Sure it's not concrete proof, but in a game with such extremely simplistic AI patterns, I have a hard time believing they'd need to track those internal ID numbers in the test server for AI debugging reasons.

This would explain the perceived "global waves" as with any RNG, there would be times when the game assigns more mobs with higher multipliers (maybe once PEDs lost in decay reach a certain threshold, increasing the higher limit and/or frequency on assigned multipliers the RNG can assign to newly spawned mobs), and usually these global waves are on popular mobs, rather than the ones people hunt rarely. And you can often go to a rarely hunted spawn and get amazing loot, then move to a more popular spawn of the same mob shortly thereafter and get crappy returns (or vice-versa, of course).

Potential issues with this theory are mining and crafting. However, I would argue mining actually supports it, as it is widely known that mining an area shortly after somebody else mined it tends to be a pretty bad idea. For crafting, the multipliers could be simply assigned in the form of a queue to individual crafting machines, or even as a whole, and then get distributed and repopulated as people do their clicks.

I may be mistaken, but I believe this would also lighten load on servers and also explains why there's oftentimes planet-wide lag hiccups (although there could obviously be many other things causing these).

But more on-topic: I am fairly positive eco matters. On my research logs I've found mobs killed with uneco weapons tend to give a slightly higher average loot, compared to eco setups, but not enough to make up for the higher costs of using said uneco setup. And yes, that was over thousands of kills.
 
Last edited:
I believe McCormick posted years ago, that showed mobs had an individual internal ID number assigned.

Thats normal, in all of the games/game engines. Everey object in the game get a unique ID number. This is need for the system, and code. This ID is most generated by the engine it self and also get used for the log files generaly.

Loot get generated in the moment you click the loot button. Server algorythm generate this. The ID number have noting to do with the loot.
 
My theory is a little bit different: the algorithm that determines loot is more on the simplistic side, however loot is not determined upon the act of looting, but rather assigned to each mob on being spawned. Maybe just the multiplier, or maybe the exact contents, I don't know, it's mostly irrelevant. My evidence for this is a picture taken from the test server, which I believe McCormick posted years ago, that showed mobs had an individual internal ID number assigned to each. Sure it's not concrete proof, but in a game with such extremely simplistic AI patterns, I have a hard time believing they'd need to track those internal ID numbers in the test server for AI debugging reasons.

This would explain the perceived "global waves" as with any RNG, there would be times when the game assigns more mobs with higher multipliers (maybe once PEDs lost in decay reach a certain threshold, increasing the higher limit and/or frequency on assigned multipliers the RNG can assign to newly spawned mobs), and usually these global waves are on popular mobs, rather than the ones people hunt rarely. And you can often go to a rarely hunted spawn and get amazing loot, then move to a more popular spawn of the same mob shortly thereafter and get crappy returns (or vice-versa, of course).

Potential issues with this theory are mining and crafting. However, I would argue mining actually supports it, as it is widely known that mining an area shortly after somebody else mined it tends to be a pretty bad idea. For crafting, the multipliers could be simply assigned in the form of a queue to individual crafting machines, or even as a whole, and then get distributed and repopulated as people do their clicks.

I may be mistaken, but I believe this would also lighten load on servers and also explains why there's oftentimes planet-wide lag hiccups (although there could obviously be many other things causing these).

But more on-topic: I am fairly positive eco matters. On my research logs I've found mobs killed with uneco weapons tend to give a slightly higher average loot, compared to eco setups, but not enough to make up for the higher costs of using said uneco setup. And yes, that was over thousands of kills.
It is an interesting theory and it can't be disproved (just like any other theory can't be irrefutably "proven"). You missed the best argument btw, the case when (about 10 years ago) someone got a HoF on asteroid, then server crashed immediately after, when he logged in again the HoF was gone, the guy killed the same mob again and got the exact same HoF again. Seems to prove the "loot in the mob" theory... unless server state was serialized upon crash, including RNG state.

As trance already mentioned, all instances of an object (like: mobs) created by the server have to have unique ID's for purely technical reasons. This is how the engine works. Those ID's could be used for loot or not but the fact they exist by itself doesn't prove anything.

I would argue there's 2 observations that seem to contradict the "loot is in the mob" theory.

First, loot lag appears upon looting the mob. This is a special kind of lag, it often appears separately from the area servers (you don't experience any lag while moving around in the game, using items or switching items) - everything seems to be normal, you kill the mob but the loot does not appear. Tells us a) loot server is a separate server b) loot is calculated when the looting event occurs.

Second, the low loot periods that often seem to appear before and/or after big ATH's. Seems to point to some kind of cache building up before (or refilling after) the big payout.
One can also argue that if purely random mechanism was used for ATH's it would require MA to hold immense funds as a backup (should 10 huge ATHs happen within a week and, say, half the guys decide to withdraw). The system that actually draws resources from the lootpool to the special buffer works like built in safety mechanism - it can only pay out as much as was put in. Which system would you prefer when you have a system that deals with a real cash in real time? ;)
From this second observation - there seems to be cache (or even several caches on several levels: local spawn cache and the system-wide cache) - one could make assumption this is the mechanism used for all significant payouts (globals, HoF's and ATH's) simply because it's easier, safer, and doesn't require too much system resources (as opposed to running the full loot algorithm at every looting event).

Then again, this is also just a theory, it's pretty much impossible to prove those things either way... :coffee:
 
Fifth, loot lag is purely down to the game loading icons into RAM. Check this out by clearing your cache then looting, it takes far longer if you haven't seen the items before. This is why special items usually load last.
 
You missed the best argument btw, the case when (about 10 years ago) someone got a HoF on asteroid, then server crashed immediately after, when he logged in again the HoF was gone, the guy killed the same mob again and got the exact same HoF again.

:jawdrop:

Is that confirmed?
 
the case when (about 10 years ago) someone got a HoF on asteroid, then server crashed immediately after, when he logged in again the HoF was gone, the guy killed the same mob again and got the exact same HoF again.

:jawdrop:

Is that confirmed?

Here's a link to mentioned global

Folks need to keep up with their EU history. :)

However, there's also an argument that if loot returns are based solely on our input (I.E. more decay activity = more loot back, which it does) then that would mean that if Haxtor was on the server (or that loot location) pretty much by himself both times and the activity counters were reset to the same amount (which is likely since it seemed there was a rollback) then it follows that the same loot (in terms of ped amount) would occur.

I once had the same basic idea as Huskie but later testing and info seems to point toward a "loot is made after decay is spent, not before" system.
 
Fifth, loot lag is purely down to the game loading icons into RAM. Check this out by clearing your cache then looting, it takes far longer if you haven't seen the items before. This is why special items usually load last.
Sure, yes, there's this special delay that sometimes occurs when something new is in loot and there's no icon cached for it in RAM.

There's also the general loot lag that occurs when nothing new or rare was looted, and it occurs for several different hunters at once - starts at the same time, and goes away at the same time.
 
It is an interesting theory and it can't be disproved (just like any other theory can't be irrefutably "proven"). You missed the best argument btw, the case when (about 10 years ago) someone got a HoF on asteroid, then server crashed immediately after, when he logged in again the HoF was gone, the guy killed the same mob again and got the exact same HoF again. Seems to prove the "loot in the mob" theory... unless server state was serialized upon crash, including RNG state.

Ah, yes, I forgot about that. It's been a long time.

As trance already mentioned, all instances of an object (like: mobs) created by the server have to have unique ID's for purely technical reasons. This is how the engine works. Those ID's could be used for loot or not but the fact they exist by itself doesn't prove anything.

Yes, indeed. But I feel there would be no need to actually have those numbers displayed behind the mob's name, otherwise you'd see those numbers on everything, including terrain features like trees and the like, which was not the case. But yes, I am quite aware this is more of a possible hint, than actual evidence.

I would argue there's 2 observations that seem to contradict the "loot is in the mob" theory.

First, loot lag appears upon looting the mob. This is a special kind of lag, it often appears separately from the area servers (you don't experience any lag while moving around in the game, using items or switching items) - everything seems to be normal, you kill the mob but the loot does not appear. Tells us a) loot server is a separate server b) loot is calculated when the looting event occurs.

I think loot lag could simply be the actual contents of the loot being determined upon the act of looting, whereas the multiplier is determined on spawn. Might be to avoid the fact that if loot content and not just size was determined upon mob spawn could result in mobs holding valuable items from being killed by turrets or drowning and then the item not being spawned again, or at least not for a long time. Also could be a mechanism to lighten stress load on the servers by having only loot contents determined rather than size and content.

Second, the low loot periods that often seem to appear before and/or after big ATH's. Seems to point to some kind of cache building up before (or refilling after) the big payout.
One can also argue that if purely random mechanism was used for ATH's it would require MA to hold immense funds as a backup (should 10 huge ATHs happen within a week and, say, half the guys decide to withdraw). The system that actually draws resources from the lootpool to the special buffer works like built in safety mechanism - it can only pay out as much as was put in. Which system would you prefer when you have a system that deals with a real cash in real time? ;)
From this second observation - there seems to be cache (or even several caches on several levels: local spawn cache and the system-wide cache) - one could make assumption this is the mechanism used for all significant payouts (globals, HoF's and ATH's) simply because it's easier, safer, and doesn't require too much system resources (as opposed to running the full loot algorithm at every looting event).

As I pointed out in my previous post, I'd wager money on the fact that the RNG system has a threshold limiting system to prevent it from paying out more than MA can afford to and limiting size and amount of multipliers the RNG is allowed to distribute at any given time. I agree going purely random without restraint would never work in an RCE as there would be way too many ways it could backfire for MA, even if it was designed to overall make the house win.
 
As I pointed out in my previous post, I'd wager money on the fact that the RNG system has a threshold limiting system to prevent it from paying out more than MA can afford to and limiting size and amount of multipliers the RNG is allowed to distribute at any given time. I agree going purely random without restraint would never work in an RCE as there would be way too many ways it could backfire for MA, even if it was designed to overall make the house win.
Yes, this is a fair argument indeed.
It could work either way and we'll prolly never know for sure... :)
 
I'm 99.9999999% certain loot isn't in the mob when spawned, my observations pretty much confirm it beyond doubt.
 
I'm 99.9999999% certain loot isn't in the mob when spawned, my observations pretty much confirm it beyond doubt.

Interesting. Care to elaborate on how you found a way of testing that? :)
 
Some years back there was a bug that you didn't get any loot off mobs if you looted them while you were dead. I made a support case about it and got an interesting answer

Hi,

We are looking into what is causing this behaviour if you loot when you are dead. In the meantime please do not loot while dead; loot when you have revived and run back. It is not intended that you should be able to do anything when you are dead.


We regret to tell you that we have no way of compensating for this; the creature does not get logged until looted, meaning that we cannot verify what would have been on it. We apologize for the inconvenience.


Kind regards,
Entropia Universe Support
 
Back
Top