-
Hail Guest! We're looking for Community Content Contribuitors to Stratics. If you would like to write articles, fan fiction, do guild or shard event recaps, it's simple. Find out how in this thread: Community Contributions
-
Greetings Guest, Having Login Issues? Check this thread!
-
Hail Guest!, Please take a moment to read this post reminding you all of the importance of Account Security.
-
Hail Guest! Please read the new announcement concerning the upcoming addition to Stratics. You can find the announcement Here!
holy crap - they actually said they had a new RNG in testing
How nervous would you be knowing sensitivity of the coding of this game when time comes to flip the switch and roll out the new RNG.
Also makes one curious as How Much testing will be done before release.
*Revert*
Also makes one curious as How Much testing will be done before release.
*Revert*
F
Fayled Dhreams
Guest
it can't be any worse - I haven't gotten a rare drop of any kind off any monster I've killed in the last 3 months - not 1 - and I've killed a ton of different stuff - nothing is dropping - can't be worse than zero - there is no negative numbers (although i'm sure they'll find a way to invent a way to take stuff out of my pack)
C
Cloak&Dagger
Guest
lmfao. This would be how I feel right now haha.
M
Muu Bin
Guest
They have... it's called your monthly fee...
I wonder if Luck with work any better with a new RNG.
Do we even know if its actually 'working as intended' now?
But yes, please let this bake on TC for a good long while before it touches a production shard.
Do we even know if its actually 'working as intended' now?
But yes, please let this bake on TC for a good long while before it touches a production shard.
W
Whinemaker
Guest
I was trying to imbue something the other day with a 99.6% success chance... and I failed 6 times in a row.
I'll take any new RNG.
I'll take any new RNG.
I agree.
The RNG I see as the "heart" of the game. If it is a good one and it works well, players will be happy and see statistics fairly applied to their time spent in the game.
If it is a bad RNG, it will make players upset and risk them leaving the game thus losing subscriptions.
But more than that, I am extremely happy by these news because they show to me that Ultima Online is getting real support, not just patched up one.
To me, this shows that there is will and effort to make Ultima Online endeavour for still a long time..
Yes, this is extremely good news !!
I bet they could switch to a new RNG without telling anyone and players will continue to complain the same as they always have. UO does not have a bad RNG, but it also is not perfect.
Bulk Order Deeds are the perfect example of players hating the RNG for no reason. You are not going to get a valorite runic BOD 1 out of 100 times. You have a 1% chance each time you get one. That chance does not change the more you get. Developers in the past did mass pulls of BODs and they came out exactly as the percentages predicted.
Bulk Order Deeds are the perfect example of players hating the RNG for no reason. You are not going to get a valorite runic BOD 1 out of 100 times. You have a 1% chance each time you get one. That chance does not change the more you get. Developers in the past did mass pulls of BODs and they came out exactly as the percentages predicted.
the hell with that - anything is better than what they've got right now - put it in tomorrow - jeez - if I have to "test" it on my home shard - that's fine - nothing could be worse - it's impossible - just stick something / anything in immediately and then tweak it if you have to - just do SOMETHING
With their track record of bugged-as-hell publishes this year? Are you serious?
No thank you, I'd rather retain what works crapily - but we KNOW "works" before trying out something this truly GAME CHANGING without testing. That would be like them revamping treasure maps without something to check for a valid location...oops...they did that.
Definitely just give us the new RNG to replace the bad one we have now. It certainly can't be worse.
I have yet to ever get an Ancient SOS. 1 in 25? Riught....
The imbueing percentages are definitely a bad joke.
The top three wood flavors? Only way I have gotten them is using a treestump.
I have yet to ever get an Ancient SOS. 1 in 25? Riught....
The imbueing percentages are definitely a bad joke.
The top three wood flavors? Only way I have gotten them is using a treestump.
R
RavenWinterHawk
Guest
Yes please fix the RNG so I get a great chance of getting the same old stuff from the same old monsters.
Forget random adventure and all that. As long as the RNG provides me with a bit less randomization and more favor... maybe make it less random. Great.
After all the true monster to battle in this game is the RNG. Ive been wrong all these years.
I want I want I want.... wooha now I have, decrease the drop rates please!
Forget random adventure and all that. As long as the RNG provides me with a bit less randomization and more favor... maybe make it less random. Great.
After all the true monster to battle in this game is the RNG. Ive been wrong all these years.
I want I want I want.... wooha now I have, decrease the drop rates please!
This is the only change i think is needed. As i understand it, luck only helps with how much "trash loot" (Its all worthless nowadays) you get, and Artifact drops. Since most of the of the SA and HS loot isnt artifact, does luck have any effect on it? Or are they all considered minor arties? I guess for anyone to have a definitive answer he would have to know the code or whatever.
Personally i wouldnt like to see a system like was introduced to doom where its guaranteed that you will get an arty as long as you stay / camp enough, but i would like to see Luck effect the drop rates, if it doesnt already.
C
Coppelia
Guest
What a game needs isn't a truly random numbers generator, but a number generator with a fair pattern. Real random sucks. It sucks as failing several time in a row a high percentage chance of success.
If you have 1/3 chance of success, it will look like that only if you have approximately 1 success every 3 attempts, not if like from a real RNG can give you get 5 success in a row and then like 12 failure in a row. Particularily if you make only few attempts (for a costy thing for example). Typically when enhancing that luck 100 +4 other mods you found on a mob, you don't care if on 10,000 attempts the stats are correct.
If you have 1/3 chance of success, it will look like that only if you have approximately 1 success every 3 attempts, not if like from a real RNG can give you get 5 success in a row and then like 12 failure in a row. Particularily if you make only few attempts (for a costy thing for example). Typically when enhancing that luck 100 +4 other mods you found on a mob, you don't care if on 10,000 attempts the stats are correct.
The RNG is NOT that bad...let's at least be somewhat serious.
T
Trebr Drab
Guest
I've noticed that when this happens, all the lights on those machines in Exodus's lair go dim.
I think I'm with Martyna on this... Yes I want this fix very, very badly... but not bad enough that I want something just "tossed" out there and not seriously tested...
I mean I would like to not fail as bad as I do all the time now... I swear the current RNG picks some folk to just be "failures" all the time... I know some folk have some sort of favor with the RNG and I don't know why. I know one guy who always gets stuff... He's so lucky with the RNG it makes you want to puke... and it seems weighted in some way in his favor.
I'd hate to have it thrown out there and things just becoming insanely "easy" to get... I like how things work in Doom now where you are rewarded for your continued effort. I'd like to see THAT sort of system added to other things... like champ spawns and mini-champ spawns... and things of that nature.
I don't like things that reward the highest damager.... since some folk like myself suffer from Poor internet service... it's not possible for me to ever be the top damager... My internet is far too sluggish for me to compete in such things. I prefer the random reward but with some sort of system in play that the more often you work the spawn without reward the higher and higher your chance becomes.
I mean I would like to not fail as bad as I do all the time now... I swear the current RNG picks some folk to just be "failures" all the time... I know some folk have some sort of favor with the RNG and I don't know why. I know one guy who always gets stuff... He's so lucky with the RNG it makes you want to puke... and it seems weighted in some way in his favor.
I'd hate to have it thrown out there and things just becoming insanely "easy" to get... I like how things work in Doom now where you are rewarded for your continued effort. I'd like to see THAT sort of system added to other things... like champ spawns and mini-champ spawns... and things of that nature.
I don't like things that reward the highest damager.... since some folk like myself suffer from Poor internet service... it's not possible for me to ever be the top damager... My internet is far too sluggish for me to compete in such things. I prefer the random reward but with some sort of system in play that the more often you work the spawn without reward the higher and higher your chance becomes.
Too little, too late...
new RNG cant hurt, and if it fixes/fully utilizes luck then great!
but as was said, RNG system is linked in with almost every other thing in game, so be careful. Think of the RNG as one of the bottom jenga pieces in a 13 year old game of jenga..... thats a load bearin piece there so take it eeeeeasy.
but as was said, RNG system is linked in with almost every other thing in game, so be careful. Think of the RNG as one of the bottom jenga pieces in a 13 year old game of jenga..... thats a load bearin piece there so take it eeeeeasy.
C
Cloak&Dagger
Guest
The rng controls every aspect of the game, not just drop rates. Aside from that point alone, Random adventure? That part has always been created by the player. If you find you adventure lacking, why not create some? Why not cause things to be harder for you. Also, even if they tried to please you which sounds impossible to me, what would control these random encounters?
It is not about a "great" chance to get anything, it is about a fair or balanced chance, it is also not just loot, every swing you take, each amount of damage you do, every time you parry, each gain you get in a skill, the skills and stats of the monsters, am I getting the point of what this does across yet?
I would however like to know what random adventure you are talking about that any amount of developer time could possibly ever fix.
C
Cloak&Dagger
Guest
Honestly the only way a new rng could break anything in game to such a serious extent is if they find a way to allow it to roll negative numbers, which in theory would simply crash the server. That sounds worse than it really is, server already crash so not much change, and the chance of them making it so you can roll negative numbers should be relatively low. Although I agree it should be tested, I am just not sure how much damage they can do by just swapping it out and never saying a word.
F
Fortyniner
Guest
After 12 years of denying there was anythig wrong with it?
I know every game has crabby no-life poopsockers who love nothing more than running around the forums screaming at everyone else for not having an adequate work ethic when it comes to slaying internet hobgoblins or whatever, but seriously? Did you just try to flame a bunch of people as being overentitled or lazy for wanting a random number generator that works properly?
Why on earth should anyone anywhere take anything you have to say seriously ever again, when it's obvious that you're either some sort of troll or just blatantly crazy?
I'd ban you on the spot, because it's patently obvious that no useful output is ever going to come from you. You're just going to run around screaming "YARR EVERYONE NEEDS TO BE MORE SERIOUS ABOUT UO! STOP BEING LAZY!" in every thread, no matter how esoteric or technical the argument or how little your ranting applies to the subject.
Devs: We might add a vendor search.
RWH: OMG LAZY PLAYERS!
Devs: We might fix the RNG.
RWH: STOP BEING LAZY EVERYONE!
Devs: We're adding some new dyes.
RWH: YOU WANT IT ALL FOR NOTHING!
Devs: We fixed a bug.
RWH: LAZY LAZY LAZYYYY *head explodes to reveal wires and robot parts*
dood your crazy. most people consider whats best for the game overall. its not a matter of "yes sounds like a good idea, we'll put it into action and eventually perfect it." What weve seen many times is things get put in, dont work, never get fixed. Or things get put in, takes too much time away from more important issues. its not a game where theres a team of a few hundred people to work on old bugs, and new systems. Theres like less than 10 people... I just want the game to work. If they can rework the rng without it taking too much time away from other things, thats great, but if not, then so be it.
And like i said before, i just hope its not like doom, or people can basically just kill something till they get what they want, like a little kid begging his mum untill he gets a peice of candy.
C
Cloak&Dagger
Guest
I am not saying anything is wrong with the RNG, but I will say the RNG is the core mechanic of the game itself, you can not have any type of D&D based game with out "dice rolls" and you can not have that with out the RNG, if the devs think the RNG needs work then it should be fairly high on the list of things to get done. It might not be that the RNG is broken, but perhaps not efficient enough, or perhaps to out dated and they wish to improve on this game by updating it.
Saying the RNG might take away from other aspects is like saying fixing a graphical glitch is more important than a glitch that crashes the server. The RNG is basically the game, with out it nothing else works and thus nothing else could be as important as it, if it was broken of course.
R
RavenWinterHawk
Guest
BINGO well said
R
RavenWinterHawk
Guest
NO lazy is the vendor search.
This thread is just silly. RNG is not 1 out of 1000 its 1% which could mean you do something 1928 times and the get your roll 2x's in a row.
Or it could be like my gypsy ring i got maybe 10 or 20 gypsys.
But I do bless all that is holy that I will finally have a new RNG so I can here people whine its still broken.
Woohoo
Oh I have lots of good input. Check all my old stuff. Goes back a bit. Enjoy.
R
RavenWinterHawk
Guest
Yep. Now if I said that...
I wrote our page on the Random Number Generator a while back, and the article ended up being dominated by its problems, both real and perceived. I'm glad they're finally trying to replace it.
Who knows, there might be something wrong with the rng, but i just hope they aren't "fixing something that isnt broke" just to make players who complain about loot happy. (again the doom revamp is what im hoping doesnt happen)
If theres really something wrong with it, then by all means fix it. It is probably outdated and could be improved.
Personally i dont see anything wrong with it. I dont play hours a day, but still get what i think is a normal amount of drops, a soul seeker here, tangle, crimson, renowned drops there, useless stuff, blah blah blah. So if there is a problem, its something that i dont think is very evident to players. Im just suspicious about the luck thing, i like the idea of luck as it takes stats away from a suit in favor of better loot, making it a bit harder to kill things. Im just hoping that the luck stat is actually being taken into some type of consideration.
If theres really something wrong with it, then by all means fix it. It is probably outdated and could be improved.
Personally i dont see anything wrong with it. I dont play hours a day, but still get what i think is a normal amount of drops, a soul seeker here, tangle, crimson, renowned drops there, useless stuff, blah blah blah. So if there is a problem, its something that i dont think is very evident to players. Im just suspicious about the luck thing, i like the idea of luck as it takes stats away from a suit in favor of better loot, making it a bit harder to kill things. Im just hoping that the luck stat is actually being taken into some type of consideration.
Testing, WTH is that? LMAO!! I say the day it goes live run out and do EVERYTHING you can. One of two things will happen, it will be grossly F'ed up and you will get nothing ever. Or it will be grossly F'ed up and everyone will get tons of drops for a day or two.
The latter is my bet.Edit: Ok, after a moments thought, it will probably be designed to drop crap like it is going out of style for a couple days, in the hopes people will like it, then after complaints, it will be turned to zero drops.
Raven, the stats are telling though. I agree, even with a 1% chance it can happen 2 times in a row, but the odds of that happening are 1 in 10,000, not unheard of.
Just to be clear, up top the 1 in 1000 chance is only .1%, as in 1/10th of one percent, so, two of those in a row is a 1 in 1,000,000 chance.
ANYTHING is possible, but we need to look at how PROBABLE it is. My worst RNG experience was having about a 10% chance to imbue an item and requiring 73 tries, before success. I do not have a calculator, but those odds would be 9 to the 73rd power over 10 to the 73rd power = VERY long odds to fail that many times in a row, but not impossible. To bad I did not play the lottery that day.
I remember the first knight they made the changes to doom.
15 arties in about 2 hours. That was nice. LOL
15 arties in about 2 hours. That was nice. LOL
B
Babble
Guest
Bah another RNG which will you give the illusion of having a chance to get a good item
We lost every coin toss one football season. We lost the toss 12 times in a row. Obviously a coin is a very poor rng device.
B
Babble
Guest
I would either check the coin or the one who throws it.
Sounds rigged
Sounds rigged
Basara
UO Forum Moderator
Moderator
Professional
Governor
Stratics Veteran
Stratics Legend
Wiki Moderator
UNLEASHED
Campaign Supporter
Sorry, but you're dead wrong.
The summer before he left UO, Draconi actually wrote a report on his examination of the UO RNG and how it IS BROKEN. IT was on his (no longer extant) official Facebook page for his UO projects.
He was planning on working on a replacement back then, but then he departed, and the team is just now getting around to working on it.
The problem with the RNG was that, while it is "fair" over 100s of thousands of iterations, there was something about the RNG seed that was causing streaks with lots of similar numbers, in small sample ranges.
This is what leads to the (FAR more common than a normal distribution) streaks of similar results in low-probability outcomes at both extremes. But, since the streaks occur at every potential step of the curve, over the long run it balances out, disguising the fact that the curve is not, in fact, a far probability distribution for short term consecutive (or near-consecutive) samples.
A way to illustrate (way over simplified):
Imagine if the RNG was asked to generate numbers, with equal probability.
Instead of the numbers being truly random, there were periods of a time period we'll call "x" that made up the distribution, that is defined by the starting or ending of a sub-process in the seeding process.
In any given "x" period, the RNG is giving numbers that tend to clump around a specific value, in a way not intended. In any "x" period, there is an obvious "mode" (one value more common than the others) in the distribution that should not be there. When the attribute that defined that "x" period changed, the base position and modal value for the NEXT "x" period of streaky values changed as well.
However, since this was occurring at all points along the distribution, looking at just the accumulated data, or samples taken from multiple "x" periods, would not show the short-term streaks.
It would be like claiming that a data set of two digit numbers (with the second digit "x" being a random 0-9)
1x,1x,1x,1x,1x,1x,3x,3x,3x,3x,3x,3x,2x,2x,2x,2x,2x,2x,4x,4x,4x,4x,4x,4x,5x,5x,5x,5x,5x,5x
generated IN THAT ORDER, would produce a "fair" sample, if sampled once, for 4 consecutive results within the span of the 30 items in the population(instead of 4 numbers taken from random positions, multiple times). There's a couple spots where something approximating the mean of the distribution would occur, if the single sample was taken there, but most other places, it would not.
(like I said, I'm oversimplifying this almost to the point of absurdity, but there's not an easy way to describe the type of streakiness that Draconi found - and that players had been relating ancedotal evidence of for over 10 years, by the time he looked for and found the cause)
B
Babble
Guest
But we might lose our Urban Legends with a new RNG
No more eating for better gains or success chances. No more wait 5 minutes until RNG resets, or move to a different spot...
no more checking the moonphases and if Trammel Is alligned with ... uhm I am rambling now .. but ... Fayhled where are you when i need rambling?
No more eating for better gains or success chances. No more wait 5 minutes until RNG resets, or move to a different spot...
no more checking the moonphases and if Trammel Is alligned with ... uhm I am rambling now .. but ... Fayhled where are you when i need rambling?
Here's the thread where Draconi discusses the RNG and why it is broken:
http://vboards.stratics.com/uhall/144665-random-number-generator-questions.html
About the BOD testing:
http://vboards.stratics.com/uhall/144665-random-number-generator-questions.html
Oh, I never did.
Although I *am* saying that 12 years ago some poor soul decided to use a pretty naive method that doesn't work very well.
About the BOD testing:
At the 90th post, he says they've found the source of the problem and a fix is forthcoming:
This last post is months before his farewell but I'm not sure if these changes are implemeted.Forthcoming, we found the source of the problem two days ago
There's a pretty easy way to fix the RNG in UO: play a mage, haha.
But seriously, seems odd that something that should be rather simple, like a RNG, could have such serious flaws. But a true RNG, as it has been explained, gives a fixed percent chance, which is unnatural and misleading over a small sample size. However, I don't see people giving up on Las Vegas, or Atlantic City, or their local lottery drawings, either. The sample size changes for the local lotto, and a slot machine, for example, is influenced by the current position of all the rows, but the general odds do not change. Yet, while black jack is by far the best odds, people love to waste their money on roulette, craps, and all of the other poor win ratio games in the casino.
Anyway, the RNG is a must for certain aspects of the game, whereas point systems fulfill other aspects of the game. To fix a RNG you'd need to force a success, such as how the Doom Gauntlet artifacts were changed. A true RNG is never going to be satisfying for someone who wants to "achieve" something, only for those who want to "win." That's why the GGS system is nice to have around, because it lets you know you will eventually get a gain if you are currently struggling.
Eh, last time I did a Goblin mini-spawn I got an arty drop off of each of the two end bosses. Then I did two Corgul runs and got the sash and the rare blades drop. But when I tried to get a specific arty from a Cavern Rat boss, a friend and me killed over 30 of that specific boss before we were like, "f that." All I can say is when you expect to win something you want to achieve, your morale goes straight to the toilet and you give up ;P.
But seriously, seems odd that something that should be rather simple, like a RNG, could have such serious flaws. But a true RNG, as it has been explained, gives a fixed percent chance, which is unnatural and misleading over a small sample size. However, I don't see people giving up on Las Vegas, or Atlantic City, or their local lottery drawings, either. The sample size changes for the local lotto, and a slot machine, for example, is influenced by the current position of all the rows, but the general odds do not change. Yet, while black jack is by far the best odds, people love to waste their money on roulette, craps, and all of the other poor win ratio games in the casino.
Anyway, the RNG is a must for certain aspects of the game, whereas point systems fulfill other aspects of the game. To fix a RNG you'd need to force a success, such as how the Doom Gauntlet artifacts were changed. A true RNG is never going to be satisfying for someone who wants to "achieve" something, only for those who want to "win." That's why the GGS system is nice to have around, because it lets you know you will eventually get a gain if you are currently struggling.
Eh, last time I did a Goblin mini-spawn I got an arty drop off of each of the two end bosses. Then I did two Corgul runs and got the sash and the rare blades drop. But when I tried to get a specific arty from a Cavern Rat boss, a friend and me killed over 30 of that specific boss before we were like, "f that." All I can say is when you expect to win something you want to achieve, your morale goes straight to the toilet and you give up ;P.
Anyone who seriously thinks the RNG is not broken are the ones making out with good RNG rolls.
One of my guildies got 3 pairs of Leggings of the Insane Tinker within a 24 hour period, not grinding the spawn. I grinded the spawn for several hours, several times, even solo'd it, and didn't receive a single pair.
During Navrey runs or Champ Spawns, generally there is one specific person in our guild that gets 80% of the 120 drops every time, as well as replicas.
The RNG is neither fair or effective, as Draconi researched before leaving. Unfortunately we have lost him, but we still have good Devs to look into these things and fix them.
Does anyone have a time for when they MENTIONED changing the RNG btw? I watched the entire Dev video and yet I could find where they mentioned this.
One of my guildies got 3 pairs of Leggings of the Insane Tinker within a 24 hour period, not grinding the spawn. I grinded the spawn for several hours, several times, even solo'd it, and didn't receive a single pair.
During Navrey runs or Champ Spawns, generally there is one specific person in our guild that gets 80% of the 120 drops every time, as well as replicas.
The RNG is neither fair or effective, as Draconi researched before leaving. Unfortunately we have lost him, but we still have good Devs to look into these things and fix them.
Does anyone have a time for when they MENTIONED changing the RNG btw? I watched the entire Dev video and yet I could find where they mentioned this.
Which means it will either be better six times in a row or worse 10 times in a row using the current RNG as a guide...
you DO know .... there is only a 30% chance that it will be "better" ...
let me stir this pot with a giant spoon then - what if they have a flag or an adjustable slider of some sort on everyone's account that they can change to make it easier or harder to get the rare drops and let's say that they adjust it by how annoyed the GMs get when you page them too much or by how often their 3rd party detection database says you're cheating - I could see them playing favorites like that with their friends and secretly screwing over the people that annoy them - just so they don't have to actually ban people outright but they still get their satisfaction of screwing them over because it does seem like certain people always get the drops and more than that - certain other people never get the drops - ever - a while back in I think DAOC, 1 player had a flag set on his account that made sure he never got a drop - I think it was the Oui flag (we flag) - i spelled his character name wrong - this guy had done hundreds and hundreds of things that should have gotten him a drop until 1 day he complained on the boards and the devs investigated and found out that they had that flag set on his account so he could never ever get any drops - after they fixed it - things went back to normal for him - just wondering if a modified / more finely tuned version of that is going on here (perhaps even intentionally) - *end huge pot stirring* hehe
K
Kylas
Guest
This would be the definition of what a random number generator ISN'T.
NIST has well known published tests you can execute easily to see if you are generating really random bits. If the developers are spending time writing a new Pseudo random number generator is means their current implementation fails those tests.
The problem isn't the random source as I think Coppelia was trying to state. It's how those random inputs are applied. What you want is a pattern generator that has a random input component that can skew a result slightly from that pattern.
It should look like a signal with a noise component. Not just pure noise.
Dig