DEVs, can we still hope to get a NEW RNG engine for UO ?

[popps]

I remember from a good while back, posts about the need for UO's RNG to be reworked to avoid the problems lamented by many players like the streakiness of the current RNG etc.

Can we still hold hopes that the game we play will get a better and fairer RNG ?

Thanks.

 

[Sevin0oo0]

If I remember right, this IS the new one, as i think they said they were evaluating a newer one. Not sure if they kept it, or reverted. a good Ask-A-Dev question.

 

[Raptor85]

better yet instead of a pure RNG how about replacing it with a weighted system, each failure increases chance of success, each success increases chance of failure enforcing that the "chances" as stated actually happen instead of simply "technically" being that over long periods of time. Those are far more fair, and FAR more fun than relying on things to "even out" in the long term. For instance now a 120 swords vs a 120 swords, technically both have a 50% chance of hitting each other, stand right next to each other and attack, one may get 10 hits in a row, one may get 10 misses in a row, it's completely random as to the outcome. In a weighted system the chances of "streaks" goes down with each sucessful hit/miss, making your "luck" with the RNG a lot less of a factor in the fight.

 

[O'Brien]

Like with the virtue arty drop system?

Wow, has it already been 5 years since they added those?

better yet instead of a pure RNG how about replacing it with a weighted system, each failure increases chance of success, each success increases chance of failure enforcing that the "chances" as stated actually happen instead of simply "technically" being that over long periods of time. Those are far more fair, and FAR more fun than relying on things to "even out" in the long term. For instance now a 120 swords vs a 120 swords, technically both have a 50% chance of hitting each other, stand right next to each other and attack, one may get 10 hits in a row, one may get 10 misses in a row, it's completely random as to the outcome. In a weighted system the chances of "streaks" goes down with each sucessful hit/miss, making your "luck" with the RNG a lot less of a factor in the fight.

 

[MalagAste]

better yet instead of a pure RNG how about replacing it with a weighted system, each failure increases chance of success, each success increases chance of failure enforcing that the "chances" as stated actually happen instead of simply "technically" being that over long periods of time. Those are far more fair, and FAR more fun than relying on things to "even out" in the long term. For instance now a 120 swords vs a 120 swords, technically both have a 50% chance of hitting each other, stand right next to each other and attack, one may get 10 hits in a row, one may get 10 misses in a row, it's completely random as to the outcome. In a weighted system the chances of "streaks" goes down with each sucessful hit/miss, making your "luck" with the RNG a lot less of a factor in the fight.

There has been much talk about the RNG in the past... one DEV even admitted it was borked... they at one time I swear said they had some sort of fix or idea about it.... but I don't think anything was EVER acted upon.

I rather like the weighted system... though I think the RNG is stuck on negative numbers for me and has been for a great many years.

 

[Lord Frodo]

Just made a 2050 luck suit and (in shame) I am seeing no difference then when I run a 1000 luck suit. To me this is unsat because we are spending real money (reforge tool) to make them and they do not appear to make any difference. I do see a difference with no luck and having some luck but none really with a leet luck suit.

 

[Zyon Rockler]

I agree with Raptor. If the RNG is made to be random then anything it does really has no outcome. It's just the system trying to produce a prediction of what it thinks its' programmers wanted to produce, like i'm thinking of a number between 1 and 10. The problem is the computer doesn't realize its already chosen certain numbers, so it will keep guessing certain ones. Also some numbers depending on how many are used have a better chance of coming up than other numbers. I have always wondered about numbers in there relationship to each other. If you look at it from 1 to 10 and then you go 11, you now have 4 1's and when you add the number 12 you now have the number 2, 2x. So, as you look at numbers as they're growing, you'll begin to see that it is repetitive.

So, even though there is a difference between the amount the number is the same. So, by looking at the smaller sets of numbers, for example: The power of 10, you can see much larger numbers just by knowing to add zeros or by just adding the larger number, like, 8,000 and 8,000 is 16 because you can use your mind to figure out that you are still in the thousands category. So, then you have to wonder how a computer reasons these different types of numbers.

So, for me, not knowing for sure what's happening I would use patterns because patterns are basically self contained, mathematical structures or instructions. For example: A pyramid. If you look at the top of a pyramid, you have a triangle and if you add the next 3 groups, you still have a triangle but it's larger. So, basically what you have done is taken the math and put it into a shape and as you produce patterns you can make them abstract or use a pattern in a way that the numbers will give you results in the way the pattern is made but you have to understand that a computer can only do what it's programmed to do.

So, you would still have the same thing basically happening over and over unless every pattern was different and I don't know if that's possible because even molecules have memory that control an outcome, meaning they have also been programmed to perform a task. So, it's possible there is no such thing as random more so, something like 1 to infinity but in theory you know everything in between.

So, the best way would be to help the computer to reason so that it understands like a human and to do that you have to tell it no that you have already picked that number and when it asks again, you can tell it again you have already picked that number so that it begins to eliminate its' choices and can finally pick the number that you were thinking. Now, that doesn't necessarily mean there won't be streaks because you could pick the same number twice. So, if you increase the chances of using a different number the computer will use all of the numbers. So, if you use the number 1 the next time the choices are between 2 and 10. Or better yet 2 and 9.

 

[CovenantX]

Just made a 2050 luck suit and (in shame) I am seeing no difference then when I run a 1000 luck suit. To me this is unsat because we are spending real money (reforge tool) to make them and they do not appear to make any difference. I do see a difference with no luck and having some luck but none really with a leet luck suit.

More luck doesn't necessarily mean every item will have tons of properties & high properties, it just means more often then not you will get loot that's improved over loot you would have gotten without having luck.

3800-4000 luck [roughly a 100% chance to increase item budget].
1000-1200 luck [roughy a 50% chance to increase item budget].
^ seems to have quite the diminishing returns eh?

There's no difference between 1000 luck and 4000 luck as far as how "good" the loot would be.
it simply means that the more luck you have, the more often your loot will be of higher quality/quantity. [maybe just quality I think they removed the extra items generated via luck but I could be wrong].

Luck does seem make a difference with the Doom gauntlet system... I just wish there was a way to view your "Doom points'

 

[Cogniac]

I remember the last time we had this discussion in UHall I ended up writing a Random Number Generator article on UOGuide. It was right after Draconi started talking about redoing the RNG (he was the aforementioned dev that admitted it was "borked"). Hard to believe that was two (2) years ago.

Long story short: The RNG is not quite as bad as we all make it out, a little bit of it is just in our heads, but overall, it is pretty damn terrible.

 

[NuSair]

I see a large difference in the number of legendary artifacts I get from Cove and Cora with a luck suit than I did without one.

 

[O'Brien]

I remember the last time we had this discussion in UHall I ended up writing a Random Number Generator article on UOGuide. It was right after Draconi started talking about redoing the RNG (he was the aforementioned dev that admitted it was "borked"). Hard to believe that was two (2) years ago.

Long story short: The RNG is not quite as bad as we all make it out, a little bit of it is just in our heads, but overall, it is pretty damn terrible.

I read that article. Working in a casino, I see people all day who think the system is broken or against them somehow. None of it is based on reality. It's human nature to think we're special, and we should get better-than-average or predictable results. And when we don't, something is wrong with the system. We fail to account that the odds are against us and results are subject to volatility.

Nothing in that article you wrote indicates or shows to me that there is anything wrong or "borked" with the system. It more demonstrates to me what I see in a casino all day exists also in UO.

 

[MalagAste]

I read that article. Working in a casino, I see people all day who think the system is broken or against them somehow. None of it is based on reality. It's human nature to think we're special, and we should get better-than-average or predictable results. And when we don't, something is wrong with the system. We fail to account that the odds are against us and results are subject to volatility.

Nothing in that article you wrote indicates or shows to me that there is anything wrong or "borked" with the system. It more demonstrates to me what I see in a casino all day exists also in UO.

Ok if that's the case explain this to me.... For Years I was in a guild and we had regular hunts every Saturday... and quite often we went to Doom.... week after week month after month... Now... randomly everyone who went save me got an arty... Several of the other members would get arties almost every time we went... or at least once a month... And even though I was doing just as much damage if not more I never got one ...... month after month after month.... Infact we had one guy would ALWAYS get one.... every single time infact this one particular fella even got 2 Ornies 15 min apart.... enough to make a person want to puke his luck was that unreal... He pulled I don't know how many arties out of doom. After about 2 or 3 years I finally got my first arti..... a stinking DRAGON LANCE..... the most useless arti ... and I didn't even have a fencer. THAT is the kind of luck I have...

It came again when we did 100's and 100's of runs of Travisty and Lady M..... everyone else in the guild had gotten at least 2 or 3 Crimmies before I ever got one. Infact some had gotten 10 or more before I got one. So convince me that there is no such thing as bad luck.... or that the RNG isn't against some folk.

 

[Driven Insane]

Ok if that's the case explain this to me.... For Years I was in a guild and we had regular hunts every Saturday... and quite often we went to Doom.... week after week month after month... Now... randomly everyone who went save me got an arty... Several of the other members would get arties almost every time we went... or at least once a month... And even though I was doing just as much damage if not more I never got one ...... month after month after month.... Infact we had one guy would ALWAYS get one.... every single time infact this one particular fella even got 2 Ornies 15 min apart.... enough to make a person want to puke his luck was that unreal... He pulled I don't know how many arties out of doom. After about 2 or 3 years I finally got my first arti..... a stinking DRAGON LANCE..... the most useless arti ... and I didn't even have a fencer. THAT is the kind of luck I have...

It came again when we did 100's and 100's of runs of Travisty and Lady M..... everyone else in the guild had gotten at least 2 or 3 Crimmies before I ever got one. Infact some had gotten 10 or more before I got one. So convince me that there is no such thing as bad luck.... or that the RNG isn't against some folk.

A perfect example of why they need to add the "new" Doom arti drop system to any high level drop.

Recently I've killed Cora approx 30 times and got jack crap for my trouble. I'm to the point that I've given up even trying to get that shield for now. However if I knew that the drop system for Cora was set up like Doom where the more times I kill her the higher my chances for a drop get, then I'd feel like my efforts weren't being wasted waiting for a drop that may never come.

 

[Zyon Rockler]

There are alot of stories about things like that and if you do the numbers you would find it to be so not possible. So what i like is that you can not prove what chance is can you? You can say once you flip a coin and it lands on heads that it was chance but how can you prove it was chance at all? Maybe something caused it to land on heads and you are just not able to understand how it works. Can you say for sure it was chance or could it be it was meant to be that way at that time and on that flip and that nothing could of changed the fact it landed as heads because of its connection to the world ,time and space ,and other factors.

When we look at things we think they look random but nothing is nothing could be. Like leaves on a plant change as they go up to the top of the plant but they do this for sun light. So is it possible that everything is doing something because of something or for something. Like any thing in math or science is based on what makes it happen so are you saying that it is all simply chance?

So its kinda funny to me that this is what is shown and alot of times ppl just do not like to say it is the case. But think about things like DNA they can tell you ahead of time what is going to happen and it is true for anything everything has a blueprint even our universe and smaller things like atoms.

To take it even more out of what is or what is not can you prove if what is happening right now has already happened or not or if we have already dreamed it in our minds so it can be played out. And this is the ? of them all, can you change what has or what is going to happen?

 

[O'Brien]

Folks ask me how it's possible that someone else got 2 royal flushes in 1 day, but they haven't seen one in years. All I can say is I sympathize with you. But a broken RNG? No... Not the fairest system for determining wins/rewards? Now we're getting closer.

On average, a royal flush comes in 1 out of 40,000 hands. But that's just an average. Sometimes folks get it when they just sit down at a machine, and for some it could take 1,000,000 hands. Is luck or some other deep philosophical question at play? Well, pick your poison. But none of them mean that the system is borked.

Randomization isn't as difficult for computers as many think. I remember learning QBASIC programming in 8th grade, and randomization even with that antiquated language was even good enough for most intents and purposes. Argue all you want that nothing in the universe is truly random. But computers easily generate all of the randomness we need in software, just like throwing dice is all the randomness that is needed for a craps table, and flipping a coin is all we need at the Super Bowl for awarding who gets to decide to receive or kick first.

MalagAste, I hope your guildmates shared the wealth with you If not, consider joining a new guild.

Ok if that's the case explain this to me.... For Years I was in a guild and we had regular hunts every Saturday... and quite often we went to Doom.... week after week month after month... Now... randomly everyone who went save me got an arty... Several of the other members would get arties almost every time we went... or at least once a month... And even though I was doing just as much damage if not more I never got one ...... month after month after month.... Infact we had one guy would ALWAYS get one.... every single time infact this one particular fella even got 2 Ornies 15 min apart.... enough to make a person want to puke his luck was that unreal... He pulled I don't know how many arties out of doom. After about 2 or 3 years I finally got my first arti..... a stinking DRAGON LANCE..... the most useless arti ... and I didn't even have a fencer. THAT is the kind of luck I have...

It came again when we did 100's and 100's of runs of Travisty and Lady M..... everyone else in the guild had gotten at least 2 or 3 Crimmies before I ever got one. Infact some had gotten 10 or more before I got one. So convince me that there is no such thing as bad luck.... or that the RNG isn't against some folk.

There are alot of stories about things like that and if you do the numbers you would find it to be so not possible. So what i like is that you can not prove what chance is can you? You can say once you flip a coin and it lands on heads that it was chance but how can you prove it was chance at all? Maybe something caused it to land on heads and you are just not able to understand how it works. Can you say for sure it was chance or could it be it was meant to be that way at that time and on that flip and that nothing could of changed the fact it landed as heads because of its connection to the world ,time and space ,and other factors.

When we look at things we think they look random but nothing is nothing could be. Like leaves on a plant change as they go up to the top of the plant but they do this for sun light. So is it possible that everything is doing something because of something or for something. Like any thing in math or science is based on what makes it happen so are you saying that it is all simply chance?

So its kinda funny to me that this is what is shown and alot of times ppl just do not like to say it is the case. But think about things like DNA they can tell you ahead of time what is going to happen and it is true for anything everything has a blueprint even our universe and smaller things like atoms.

To take it even more out of what is or what is not can you prove if what is happening right now has already happened or not or if we have already dreamed it in our minds so it can be played out. And this is the ? of them all, can you change what has or what is going to happen?

 

[Raptor85]

Folks ask me how it's possible that someone else got 2 royal flushes in 1 day, but they haven't seen one in years. All I can say is I sympathize with you. But a broken RNG? No... Not the fairest system for determining wins/rewards? Now we're getting closer.

On average, a royal flush comes in 1 out of 40,000 hands. But that's just an average. Sometimes folks get it when they just sit down at a machine, and for some it could take 1,000,000 hands. Is luck or some other deep philosophical question at play? Well, pick your poison. But none of them mean that the system is borked.

Randomization isn't as difficult for computers as many think. I remember learning QBASIC programming in 8th grade, and randomization even with that antiquated language was even good enough for most intents and purposes. Argue all you want that nothing in the universe is truly random. But computers easily generate all of the randomness we need in software, just like throwing dice is all the randomness that is needed for a craps table, and flipping a coin is all we need at the Super Bowl for awarding who gets to decide to receive or kick first.

MalagAste, I hope your guildmates shared the wealth with you If not, consider joining a new guild.

negative, there is very much such a thing as a broken RNG, they're found all the time, particularly in older software and older systems that didn't have a hardware RNG like newer ones do. (newer ones use stuff like temperature and current voltage to further seed the RNG and make it harder to guess)

Any RNG on a computer though is generated mathmatically via a formula using seed values, it is NOT random, the proper term is actually pseudo-random when talking about a software RNG, and if too few or guessable seed values are used on a weak formula you can guess the sequence. Some variations of them also are "streaky" in that they have a tendency to generate a lot of similar values in bunches as it moves on (eg 10, 12, 9,13,22,25,20) which is generally from using a poor or outdated formula (common in some older games as the weaker formulas were quite a bit faster than calculating using more seeds/etc.

There's TONS of computer science articles in the field, and plenty of examples around of different methods of "random number" generation and what their pitfalls and strengths are, and UO's definitely has the "feel" of a RNG prone to streaks, even the devs in the past have said that UO's RNG was broken so it's not something that's really an unknown....it would just be nice if it were fixed...

 

[O'Brien]

I guess I should clarify that I personally haven't seen any evidence of a broken RNG in what I'm talking about (UO and casinos.)

True, that doesn't mean it doesn't exist. And granted, there are plenty of things broken in UO to where it would be no surprise to me if there was a bug where some people were being given crimmies like no tomorrow, and others nada. I've just seen no evidence to believe so. But do show me if there's evidence out there that I don't know about. If you still have a quote from one of those Devs, please post.

Again, you could argue that nothing in the universe is truly random, but it is easily "random enough" for our intents and purposes. So pseudo-random vs random is pretty irrelevant here, isn't it? But I am no CS or math grad. Just someone trying to clear out all of the paranoia and reveal the facts.

negative, there is very much such a thing as a broken RNG, they're found all the time, particularly in older software and older systems that didn't have a hardware RNG like newer ones do. (newer ones use stuff like temperature and current voltage to further seed the RNG and make it harder to guess)

Any RNG on a computer though is generated mathmatically via a formula using seed values, it is NOT random, the proper term is actually pseudo-random when talking about a software RNG, and if too few or guessable seed values are used on a weak formula you can guess the sequence. Some variations of them also are "streaky" in that they have a tendency to generate a lot of similar values in bunches as it moves on (eg 10, 12, 9,13,22,25,20) which is generally from using a poor or outdated formula (common in some older games as the weaker formulas were quite a bit faster than calculating using more seeds/etc.

There's TONS of computer science articles in the field, and plenty of examples around of different methods of "random number" generation and what their pitfalls and strengths are, and UO's definitely has the "feel" of a RNG prone to streaks, even the devs in the past have said that UO's RNG was broken so it's not something that's really an unknown....it would just be nice if it were fixed...

 

[Viquire]

If we still had a community manager they would jump all over me, but I would ask Mag if you have your characters eating. They have told us for years it doesn't matter, but my own humble opinion leads me to believe that that code is right down there at the very nitty gritty of how the world turns. I think I may have had this discussion with Mag before. If that is the case I will offer apology and claim a lack of caffeine in my defense.

 

[Zyon Rockler]

I don't think there is really a right or wrong until you find an absolute. The idea is to find out how to fix the RNG so that we know for sure. I think there are very valid points on both sides to consider. You kind of have to argue both ends to come up with a solution.

It's a fascinating topic and the more we think about it, the more likely we are to understand it better. Like, look at death. No one knows when they are going to die but look at all the factors that control death. It doesn't mean that it couldn't happen right this second but it probably will not.

Like i've always thought about this one extreme where a meteor travels for billions of years and turns into this little pebble as it enters the atmosphere and bonks someone on the head and kills them. Now, that rock came an aweful long way and it's a way of saying, That either this is a very accurate system. One that defines time and space or this guy was just really, really unlucky.

But what really gets me is that if things were so random or if they are? Why can't we produce something random? Even with a coin, you have certain rules. You have heads or tails and then the possibility of it landing on its' side but when you use the word random, you've already made that impossible because of the rules. It can only be one of those three things.

The same with with cards or slots. It's a design rule system that you can change and automatically the results will change. For example: Add some aces. So, it's not really random but rather a set of event that is given a rule to decide a winner or loser.

But the idea is to make it so that people can not win. So, what will happen is, more people will lose than win but if i've learned anything from that, it's that there are sets of events that cause other events or other sets of events and that makes sense, kind of like an avalanche. So, one action could produce a set. Meaning that it's a chain reaction of rules. Another example: House of Cards. To remove one of the bottom cards would affect the entire house.

 

[O'Brien]

If we still had a community manager they would jump all over me, but I would ask Mag if you have your characters eating. They have told us for years it doesn't matter, but my own humble opinion leads me to believe that that code is right down there at the very nitty gritty of how the world turns. I think I may have had this discussion with Mag before. If that is the case I will offer apology and claim a lack of caffeine in my defense.

While you're at it, you might as well pay tribute to the RNG gatekeeper/keymaster. I've started a collection fund. Just let me know what shard you're on, and I'll come by and collect.

 

[Viquire]

While you're at it, you might as well pay tribute to the RNG gatekeeper/keymaster. I've started a collection fund. Just let me know what shard you're on, and I'll come by and collect.

I do, but I have a bloody alter for that. The UO gods will be appeased in no other fashion.

 

[Raptor85]

I guess I should clarify that I personally haven't seen any evidence of a broken RNG in what I'm talking about (UO and casinos.)

True, that doesn't mean it doesn't exist. And granted, there are plenty of things broken in UO to where it would be no surprise to me if there was a bug where some people were being given crimmies like no tomorrow, and others nada. I've just seen no evidence to believe so. But do show me if there's evidence out there that I don't know about. If you still have a quote from one of those Devs, please post.

Again, you could argue that nothing in the universe is truly random, but it is easily "random enough" for our intents and purposes. So pseudo-random vs random is pretty irrelevant here, isn't it? But I am no CS or math grad. Just someone trying to clear out all of the paranoia and reveal the facts.

There's always the philosophy of "if it ain't broke, don't fix it." This doesn't really work out well for a 12 year old game that *still* has things most MMOs long ago gave up on achieving. We have to understand and improve our own systems if we hope to improve the experience for everyone.

The RNG is one of those areas. It was originally implemented in the simplest way possible, because UO was, after all, an experiment and the RNG *worked* (or had the appearance of working) and they moved on to more important things, like, say, reducing lag between client/server messages.

I see that one minor improvement occurred when they moved the servers to Linux, but the wrapper code around the call to the RNG was still poorly implemented, skewing the results. We recently discovered that just fixing the wrapper code solves the visible problems without having to replace the entire RNG.

Unfortunately, though they found issues with it, even as draconi was heading out they had not yet implemented it but he assured us that the rest of the team had it and would fix it soon. See the problem with a borked RNG is it's NOT "random enough"...it's like playing craps with a lopsided die or roulette with some slots for the ball larger than others (and from what I understand UO's rng problem wasn't the RNG itself...but the fact that the results were being truncated in the wrapper

 

[Zyon Rockler]

I have one if you have a die that has 6 sides but every number is the same is each roll going to be random?

 

[Driven Insane]

You guys can argue about whether there's truely a RNG system that works all day and night for all I care.

The problem is that this a game, a game where we supposedly pay money for enjoyment. And it's not enjoyable to spend over 100 hours in Doom and not be able to get an Orni, HoM, etc. Doing over 100 peerless bosses and not getting a single crimson is BS.

So we're not talking about the randomness that includes winning the million dollar jackpot or the lottery. We're talking about supposed entertainment where in my opinion time spent should equal a reward. But for some of us time spent equals nothing but frustration and disappointment.

 

[RaDian FlGith]

better yet instead of a pure RNG how about replacing it with a weighted system, each failure increases chance of success, each success increases chance of failure enforcing that the "chances" as stated actually happen instead of simply "technically" being that over long periods of time. Those are far more fair, and FAR more fun than relying on things to "even out" in the long term. For instance now a 120 swords vs a 120 swords, technically both have a 50% chance of hitting each other, stand right next to each other and attack, one may get 10 hits in a row, one may get 10 misses in a row, it's completely random as to the outcome. In a weighted system the chances of "streaks" goes down with each sucessful hit/miss, making your "luck" with the RNG a lot less of a factor in the fight.

Just curious... how do you implement a weighted system without an RNG to determine success or failure in the first place?

Truthfully, a weighted system would have to be significantly complex in order to make it work in a manner that we would expect it to. Consider that sure, swords vs. swords is a 50/50. We could wait it so that each failure raises the chances 2%, each success lowers the chance 2%, but then you end up with an entirely different system that by definition would work nearly the same as the RNG. Plus, you'd have to adjust things based on other factors. Say you compare Swords A vs. Parry B... Sword A scores success, but Parry B scores success too... does that cut the damage to zero, or just a percentage? Either way, Swords A goes down, Parry B goes down as well.

And, at its core, this would all still rely on a good RNG. By the way, play RNGs as much as you want... some appear to work "better" than others, but there aren't any really good ways of measuring an RNG except by way of streakiness. But any programming to correct streakiness begins to take the random out of RNG.

Plus, at what level do they stop the success vs. failure? What about damage? Would you weight low damage against high damage to ensure critical hits at some point?

It's very easy to theorize how it should all work... but at the end of the day, the RNG is in control. And any weighted system that went up and down, by definition, would have the same mean as an RNG.

 

[RaDian FlGith]

I agree with Raptor. If the RNG is made to be random then anything it does really has no outcome. It's just the system trying to produce a prediction of what it thinks its' programmers wanted to produce, like i'm thinking of a number between 1 and 10. The problem is the computer doesn't realize its already chosen certain numbers, so it will keep guessing certain ones. Also some numbers depending on how many are used have a better chance of coming up than other numbers. I have always wondered about numbers in there relationship to each other. If you look at it from 1 to 10 and then you go 11, you now have 4 1's and when you add the number 12 you now have the number 2, 2x. So, as you look at numbers as they're growing, you'll begin to see that it is repetitive. [...] So, the best way would be to help the computer to reason so that it understands like a human and to do that you have to tell it no that you have already picked that number and when it asks again, you can tell it again you have already picked that number so that it begins to eliminate its' choices and can finally pick the number that you were thinking. Now, that doesn't necessarily mean there won't be streaks because you could pick the same number twice. So, if you increase the chances of using a different number the computer will use all of the numbers. So, if you use the number 1 the next time the choices are between 2 and 10. Or better yet 2 and 9.

Well, here's the thing... probably every roll in the game calls the same RNG, and feeds in a high and a low end. Yes, you could have it track -- briefly -- what numbers it had produced to help reduce streakiness, but I suspect the RNG is one of the most used systems in the game. Nearly every game function performs a random check. And depending on the ranges being fed to it, you might end up guaranteeing good rolls in certain instances depending on how it falls along the line.

And... for a severely used system like an RNG, you don't really want to make it overly complex, because the more complex the system is, the slower the whole thing gets with the more requests made for it to perform. Not saying the current system might not need a tweak or two, but you definitely wouldn't want to make it so sophisticated that it became a noticeable in-game issue.

Certain systems, like Doom, and so forth, actually do attach something to characters to keep track of overall successes and failures so that success is basically inevitable.

But at some point, we also have to keep in mind, failure is also an option in the range of random.

 

[Lord Frodo]

Go to shame with a 2260 Luck suit, %73 chance, and see how well you do. They said they fixed Luck in the revamped dungeons but believe me they have not. You can not tell me that the RNG is beating me out with a %23 chance majotity of the time. If UO wants people to upgrade thier suits with RL $ and wants to keep people interested in these new dungeons then they really need to look into this or stop selling upgrade items for RL $.

 

[RaDian FlGith]

Here's the problem with Luck and any portion of the game:

There is absolutely no way for us to measure the effectiveness of Luck against the results obtained. Yes, Luck is supposed to give chances to improve things, but how do you measure that satisfactorily on your end? There's no real way for us to look at a given mob and say, "Yes, Luck worked." Because even if Luck worked, we might still not see a result. And on the result that was visible, we might not see an effective result on that result that would say to us, "Yes, Luck worked."

Luck, unfortunately, is all about our perception of how it works. I mean, sure, they could send us system messages:

Your Luck caused you to gain a chance at an extra item.
Your Luck failed to cause you to gain an extra item.
On the three items you did receive, your Luck:
- Did add one property to the first item.
-- Your Luck failed to increase the intensity of that property.
- Did not add any properties to the second item.
- Did add one property to the third item.
-- Your Luck increased the intensity of that property by 5%.
-- Since the RNG resulted in a value of 1%, the intensity is now 6%.
-- This appears that Luck did not work, but we assure you it did.

But I'll be honest... I don't want that level of detail.
None of us would ever see the left hand of our screen again in the CC.
None of us would ever see a conversation again in the EC.
And regardless...

We still wouldn't believe that Luck was working as intended.

 

[Warpig Inc]

Five grains of sand doubled to ten is not the handfull the spoiled want all the time. Everyone should win a Corvette every other month. Everything in game that is an artifact, legendary or magical rares should be renamed common.

Sad thing is the rare level 12 arty damaged painting. The rare few that know the few minute window that it spawns each year. Are they lucky? Weight system would take the earning aspect out of a drop just as much. Now moonstone jewel owners know what it is to have earned luck. Was busy with other interest and like others would like to taste the grind. Just bring the moonstone jewels back with 6 mod 600 cap blank imbuer. Leave it 1 in 25000.

 

[Sevin0oo0]

as stated before, I'm against most "RNG" ascribed connections. Anyone that tried to collect the virtue armor set knows it's broken. Maybe the code is working as designed, the end point is being able to receive the items, all of them, in a timely fashion. They know too, since the quest was removed that used the armors.
Using the virtue drops as an example... I've gotten all but 1-2 items several times over, those couple things, I've never gotten - Broken? or bad design?

Since the system already keeps my 'score', and when I hit a predetermined point, I receive something. I say at that point, the item should NOT be Random selection, it should cycle thru, picking the next one, in order. Repeating the circle when all have been received. That way all items possible can be received and we will know our progress towards getting any others.

 

[Viquire]

as stated before, I'm against most "RNG" ascribed connections. Anyone that tried to collect the virtue armor set knows it's broken. Maybe the code is working as designed, the end point is being able to receive the items, all of them, in a timely fashion. They know too, since the quest was removed that used the armors.
Using the virtue drops as an example... I've gotten all but 1-2 items several times over, those couple things, I've never gotten - Broken? or bad design?

Since the system already keeps my 'score', and when I hit a predetermined point, I receive something. I say at that point, the item should NOT be Random selection, it should cycle thru, picking the next one, in order. Repeating the circle when all have been received. That way all items possible can be received and we will know our progress towards getting any others.

Not for sure, under this system everyone would always receive the first thing first. Playing again to those who powergame whatever new reward system we might be looking at. Unless, perhaps, everyone was allowed to enter the circle at a random point. Hmmmmm

 

[Zyon Rockler]

What if each char had a gear of some type or a shape that could turn that would change the rng into a force that is the first step of a set of possible out comes. Like the other day last night I got two arties in row a about 10 min apart that are the same. If the arties had a gear on the char and each tooth was an arty then the tooth might of moved to the next possible arty.

Like something that can remember for each char what set it is on. could even have a smaller gear that would cause some to be more rare where it would move back to one that was more possible. This is gonna take time yet but also in pvp you could set up something like this but am thinking not so much of a gear but a shape for example give each die a shape so that as it rolls some things are more or less going to happen then others.

So you would add something like you swing a large ax and get a tag that changes the shape of the die because you are a bit tired or it moves the next swing off a bit so basically a dexxer would get tired. lol ,...

I am trying to look at it from the roll side of things not so much skills, str, dex but the what ifs like what if the guy misses what if he hits. The way it is now you drop some dex or just miss. What if tho that missed hit makes a set of happenings.

So again the char would have some type of change that would change the out come depending on the chars state.

 

[Sevin0oo0]

Unless, perhaps, everyone was allowed to enter the circle at a random point.

works for me. As for the music gears, they're 'rare' because they have influenced the RNG. In the same way, they could say Frostwood or val is a 'rare.'
Just a huge grind, and I'll never do another one, or if the harpsichord has the same method, no thanks, I'd rather have a dancin pony.

 

[Zyon Rockler]

A list will fix the items in a way you get them all but it will not make things random. Also somthing that could use all systems would be better , you could still have a set of gears like a list but it needs to be part of the game not just a list. The first part would be like a life bar that fills up once the bar is full you have a greater chance to get a drop but if you move from one list to the next you would need a new bar to keep track of the list. Anther way would be to use one list but then you would get doom drops in ilish or tok drop in doom.. So aia think you would try for somthing that would work off the rng in even pvp.

 

[Warpig Inc]

Daemon Bone Helm nuff said.

RNG works fine. Made BOD run last night for my three off same NPC. Last two 10n halberd. Had enough gold to fill one buying halberds from the NPC. Returned the filled BOD for another, and yep. 10n Halberd. Many times over the years getting back to back same weapon or cloth BOD.

Virtue armor drops just fine doing the Void Pool defence. Been many years that I did the cloak quest part of the set.

ALways like to think with items like the gears. Karma and Fame dicided what group of gears you would get more often. Something that would get players from boring themselves using the same character or going out and trading. Even a paladin's squire needs to talk with shadey characters to see if the world still goes around.

 

[Sevin0oo0]

ALways like to think with items like the gears. Karma and Fame dicided what group of gears you would get more often.

I've seen F/K affect which holiday gift you receive, a very biased "random", imo. If they are going to use that, they should say so - with the 'rarer' ones going to the ones that the extreme ends (Glorious/Dread) with all else falling in between, You can sacrifice fame easily enough for a quick change.
Do I need to become a kitten killer to get the rest of my gears?

 

[O'Brien]

Do I need to become a kitten killer to get the rest of my gears?

Uh oh... I hope Sauteed Onion isn't reading

 

[Llewen]

What we all need is a genuine hardware quantum mechanics rng...

 

[Sauteed Onion]

Whoa A Kitten killer!! At least wait til they are full grown cats ! But, on the meowside, you notice kitties like to set on paper especially the kind that people are USING! It wouldn't surprise me that the devs decide to make the more intelligent species of britannia be the ones humanz and elves and gargolyes have to kill in order to get their music sheets.

Do I need to become a kitten killer to get the rest of my gears?

 

[Sevin0oo0]

Whoa A Kitten killer!! At least wait til they are full grown cats !

I can't do it then either. When I'm farming Trogs, I always tame and release them out of the cave before I get started

 

[Zyon Rockler]

I do not know if you can base the rng off of to many things but maybe you could add more factors .. Like if its day or night what type of moon, high- low tide, days of the week. The year of.. So many factors that you can use and not make it so complex. like if you used the days of the week for arty drops you could pick so many that would fall on that day only. Then if you log in every day you could have better chance of a complete suit !,, So now you have incentive to log in everyday of the week. Then you can change the day so the new Sunday is Monday. Plus you have a better chance of not having the same things drop.

You could also add things that are like luck like moods, so if you are mad bro you might get a type of drop or more chance to fail maybe.

So you could base mood off of color, like colors mix and change color so would your mood change. You would have basic moods like happy might be yellow, sad would be blue, mad would be red. As the color mix you would have other moods like relaxed, peace, frustration. The colors would be part of what you are doing like if you fail or succeed. As the colors change so does your chances. It adds another way to change the out come and it is part of your char not some random number.

Things like hunger could change your mood and affect pvp and pvm changing your chance to hit. Mood could play a powerful role in the out come.

The things you add to the char or the world would be a direct effect on your daily life. There are alot of things to think about like time and space but those might be to complex as well as looking at the make up of things like walls or the ground.

In a much more advanced system tho it would work well like ground tiles change how you move they could effect your mood like wet or hot and dry. And you could add thirst. Being in a desert would make you tired and hot. So you can keep expanding and adding more to the mood or to the way things effect the char.

We can see day and night changing alot of things so you could have fear. Now we can change pvp and pvm like looking at something really ugly or being out numbered could make a chars fear grow. So you would build off of that like when its dark and you can not see it might make you move slow or if you run you fall down and take damage.

So even tho we are getting past the fact it should be random we still need something to change the state we are in.

 

[Ludes]

Uh oh... I hope Sauteed Onion isn't reading

I love starting my day with a good chuckle!

 

[Basara]

I read that article. Working in a casino, I see people all day who think the system is broken or against them somehow. None of it is based on reality. It's human nature to think we're special, and we should get better-than-average or predictable results. And when we don't, something is wrong with the system. We fail to account that the odds are against us and results are subject to volatility.

Nothing in that article you wrote indicates or shows to me that there is anything wrong or "borked" with the system. It more demonstrates to me what I see in a casino all day exists also in UO.

The official word: The RNG is prone to streaks in short-term use, be they good or ill for the player.

Draconi looked at it, and discovered that while "fair" in the long term, samples taken of groups of sequential checks were showing streaks of similar results more often than a fair random generator would allow. Because these streaks were possible at ANY POINT, they balanced out in groups of thousands or millions of samples - but were really off if you looked at, say, 20 or 50 rolls in a row.

The confusion lies with the following events:
1. Draconi was the primary one looking into it.
2. He posted the discovery of just what went wrong on his FACEBOOK page that was created for official feedback from him, where he was posting a series of essays on things about UO.
3. Within a few weeks (if not days), he was left go from EA, as were a bunch of others, due to EA's budget cuts where they were gutting other divisions to compensate for losses in their console game branches. That page immediately disappeared from Facebook, replaced by his own personal FB page that had NO UO information.
4. Given the timing, it is unlikely that his plans were implemented on more than a personal test shard at work, that of course got shut down and probably deleted when he was laid off.

 

[Kylie Kinslayer]

The official word: The RNG is prone to streaks in short-term use, be they good or ill for the player.

I just know I will be glad when I get off the "snow elemental slayer" kick the RNG has me on. Two scrapper's that are snow elle, got 2 blight of the tundra drops ... and yepp... both snow ele slayers

 

[RaDian FlGith]

The official word: The RNG is prone to streaks in short-term use, be they good or ill for the player.

Draconi looked at it, and discovered that while "fair" in the long term, samples taken of groups of sequential checks were showing streaks of similar results more often than a fair random generator would allow. Because these streaks were possible at ANY POINT, they balanced out in groups of thousands or millions of samples - but were really off if you looked at, say, 20 or 50 rolls in a row.

There is something entirely different to consider as well, and something that I'm not entirely sure Draconi took into consideration.

If I sit down with an RNG, and I start requesting either manually or programatically a bunch of random numbers, yes, I might experience streaks. Let's say I ask it for 100 random numbers in the course of 1 second, and it spews them out to me, and it shows that there are streaks. It's still useless data, because that doesn't demonstrate in any way, shape, or form how the RNG is practically applied to a real-time environment.

What I mean is this:

If I ask for 100 numbers and get bunches of streaks, it proves that the RNG is subjected to streaks. What it does not prove is that there is any long-term, adverse affect on it. Consider it this way: At any given moment, assume there's 100 players doing things across the game. Let's say they all simultaneously perform something that hits the RNG. They're queued up, and the RNG returns numbers. To 100 different people in 100 different places doing 100 potentially different things.

The flip side of the coin is that "streakiness" could just as easily be defined as because there are a few hundred requests constantly hitting the RNG, some results will be repetitive based on when those calls are made and sent back simply because it's generated a ton of random numbers in between that particular person's calls.

At the end of the day, there aren't going to be a ton of perfect solutions to make a random number generator that produces what everyone expects "random" to look like. And things that adjust the random nature of things to help prevent streaks either in sequence or directly on the caller will, by definition, reduce the nature of "random."

 

[Bleak]

I remember from a good while back, posts about the need for UO's RNG to be reworked to avoid the problems lamented by many players like the streakiness of the current RNG etc.

Can we still hold hopes that the game we play will get a better and fairer RNG ?

Thanks.

Yes, there is still hope! I am surprised that there have not been many threads on noticing a difference in the RNG . Since Publish 75 all shards have been using an updated PRNG(Mersenne Twister) which had been baking on Origin for over a year. This thread has been a joy to read.

 

[Basara]

Radian: The effects DID balance out, long-term. That's immaterial.

The problem was that you had more than expected short-term streakiness that adversely affected game play in terms of causing hardship and discontent among some players at any given time. A crafter that fails FIVE TIMES IN A ROW on a 99% smelt chance, for example - and has it happen several times in an hour. Sure, someone else might get 5 hits in a row about the same time, when they should have averaged 1 at best, but that doesn't benefit the crafter who got 16 ingots from smelting a pile of 256 bronze ore in the near term - especially when you consider if the crafter succeeds the first time, the whole pile is converted, and there are NO further checks to take on that ore pile - so for checks such as those, streaks are only apparent on the bad side of the equation.

 

[Tina Small]

Yes, there is still hope! I am surprised that there have not been many threads on noticing a difference in the RNG . Since Publish 75 all shards have been using an updated PRNG(Mersenne Twister) which had been baking on Origin for over a year. This thread has been a joy to read.

Bleak, just a bit ago this evening, it took a character of mine on Baja with 112.1 real taming skill (she had no taming jewelry on) 11 successive attempts (i.e., 10 failures in a row followed by a success) to tame a unicorn that I know had never previously been tamed. The last time I took that character out to get a gain was five days ago. I did notice after she finally got the unicorn tamed that when she ate something, she was "extremely hungry."

If this was the only character that I was training taming on that's at that level, I'd say this was a fluke occurrence. But I've noticed similar strings of failures with other characters, some of whom have even higher levels of taming.

Did you make taming unicorns more prone to failure recently, could being "extremely hungry" have been a factor in the successive failures, or is this example of a long string of failed attempts an example of streakiness that is still present in the RNG even after you updated it?

 

[Nimuaq]

The probability distribution of a random process provides how probable certain outcomes are. When the probability distribution is uniform (uniform distribution, uniformly distributed, equidistributed, etc), it means that all outcomes have the same probability or each number is as likely to be generated as any other. In other words, when you roll a dice, each outcome (from 1 to 6) is as likely to appear. (In an unrelated note, there is no way to predict the outcome of a truly random process.)

A pseudo-random number generator (PRNG) generates pseudo-random numbers instead of random numbers, which basically share the same statistical properties of random numbers, in other words, have no distinct patterns (or streaks as it is called before). PRNGs usually start with an initial state and by using a transition function on these states it produces an output. The PRNG algorithm mentioned by Bleak is proposed to generate uniform pseudo-random numbers.

A very very old example of a PRNG is the linear congruential generator: next_number = (a*(current_number) + b) mod m. Let a = 2, b = 7 and m = 10 with initial_number (first current_number) = 1 . next_number = (a*(current_number) + b) mod m = (2*(1) + 7) mod 10 = 9. Now, with current_number = 9, the next_number is : 7. If we continue, we get: 1, 9, 7, 3, 4, 6, 1, 9, 7, 3...

As you have already noticed, it looped back after 6, so we don't have the best algorithm here but still the initial sequence up to 6 looks random. Java programming language actually uses a variant of the algorithm above to generate random numbers (java.util.Random, with a = 25214903917, b = 11 and m = 2^48). However, we only checked it to see a basic PRNG.

One common programming error on PRNGs is not checking for thread safety, or the manipulation of the shared current state by separate processes, especially when using it in parallel. While this might explain "streaks", it is an unlikely error.

Also, as we already discussed with O'Brien before, a random sized subset (eg. Tina's set of ten non-consecutive numbers generated by the PRNG from above) of a uniformly distributed random sequence, especially if the size is rather small, does not guarantee a uniform distribution nor does it mean it will be free of patterns. He pointed out the effects of clustering illusion on that discussion, and while I don't agree that it explains the so called patterns, I still agree that we tend to remember what appears to be patterns more often.

 

[RaDian FlGith]

Radian: The effects DID balance out, long-term. That's immaterial.

The problem was that you had more than expected short-term streakiness that adversely affected game play in terms of causing hardship and discontent among some players at any given time. A crafter that fails FIVE TIMES IN A ROW on a 99% smelt chance, for example - and has it happen several times in an hour. Sure, someone else might get 5 hits in a row about the same time, when they should have averaged 1 at best, but that doesn't benefit the crafter who got 16 ingots from smelting a pile of 256 bronze ore in the near term - especially when you consider if the crafter succeeds the first time, the whole pile is converted, and there are NO further checks to take on that ore pile - so for checks such as those, streaks are only apparent on the bad side of the equation.

Again, Basara, that's at least as much a function of the number of calls being made as it is the RNG itself. It's entirely possible to pick up the DnD equivalent of a 100-sided-die and roll 98, 97, 99, 98, 98. Not likely, but possible. That would be "streakiness," and yet entirely possible.

And no, long-term balance is not at all immaterial. That's how "random" works whether it's on a computer or not. You could flip a coin 10 times and get 10 heads in a row. That doesn't change that long-term you'd get heads and tails roughly 50% of the time each.

What you're sort of arguing with your mining example is a bit silly. Yeah, if you succeed, it's all converted. Would you prefer that on a failure it's all lost? To me, multiple attempts is better than complete loss from the get-go.