Martis Running in streaks?

[Basara]

I had a pretty good night down at Melissa - I got TEN minor artifacts from paragon balrons. I'm turning them in, so it doesn't matter to much to me, EXCEPT....

The first two Martis I got were the Bow of the Juka King.
I then got a Violet courage
I then got FOUR Nox Ranger's Crossbows in a row, over 90 minutes
Followed by TWO Staff of Power in about a half-hour
then finishing with an orcish visage.

Isn't this a pretty good indication that whatever randomization factor is supposed to be at work, isn't really functioning that well?

Note that one of my Guildmasters had about 9 of his first 11 or 12 marties as Arctic Death Dealers, when Minors were introduced way back when, and ADDs are STILL about half the drops he gets from the original set.

 

[Tazar]

I had a pretty good night down at Melissa - I got TEN minor artifacts from paragon balrons. I'm turning them in, so it doesn't matter to much to me, EXCEPT....

The first two Martis I got were the Bow of the Juka King.
I then got a Violet courage
I then got FOUR Nox Ranger's Crossbows in a row, over 90 minutes
Followed by TWO Staff of Power in about a half-hour
then finishing with an orcish visage.

Isn't this a pretty good indication that whatever randomization factor is supposed to be at work, isn't really functioning that well?

Note that one of my Guildmasters had about 9 of his first 11 or 12 marties as Arctic Death Dealers, when Minors were introduced way back when, and ADDs are STILL about half the drops he gets from the original set.

Sounds like our RNG (Random Number Generator) is acting like it always has....

 

[Limlight]

I think its just RNG...

I have gotten 7 Virtue Armor Chest pieces in a row...

YES....7 in a row.

In destard fel...

 

[Kiminality]

If you consider it, you don't have a personal random number generator.
Everyone gets numbers from the same RNG.
So, for every random number that's generated for you, there are a bunch generated in between, for other people.
So, any streak represented for you, is not necessarily an overall streak.

To consider streaking overall, you'd want to track the awarding of hundreds of marties.

 

[Lord Richard]

One morning I recieved 4 virtue armor pieces in less than 4 hours, but the nexd day it was maybe one per 4 hours...or more.

 

[Farsight]

If you consider it, you don't have a personal random number generator.
Everyone gets numbers from the same RNG.
So, for every random number that's generated for you, there are a bunch generated in between, for other people.
So, any streak represented for you, is not necessarily an overall streak.

To consider streaking overall, you'd want to track the awarding of hundreds of marties.

Believe me, with our weather right now, I'm considering streaking overall.

As for the RNG, I believe that our character has something to do with it.

For example, I have 115 swords, the person I hunt with most has 115 swords. We attack the same creature. Sometimes I miss every swing, she hits every swing. Five minutes later, it's the other way around.

Then there's the paragon thing. I believe that paragons spawn in certain time intervals rather than randomly. For example, if I'm hunting Melissa with my tamer and a rune beetle and bake kitsune, and another person has a dread horse/mare combination, the balrons in there can die extremely quickly. If a paragon spawns, and we kill the paragon balron (best time, the pair of us and 2 dragons was 15 seconds to a dead paragon, usually it's at least 30), then the next balron to spawn is nearly always a paragon.

As for the artie drops, I hunted Miasma for several months before I got a single drop, then I got 8 drops in about 2 hours, and SIX of them were rightous anger (this was a couple months ago). So there's something fishy about the RNG.

And I'm so convinced about that that if I have that "unlucky" feeling, I won't even play. And if I start getting drops, it's hard to pull me away from the game.

 

[phantus]

Martini's and steaks?!!.........oh.....nm

 

[Basara]

My suspicion is that the RNG uses a string from the character (or a character present, for multiple people present) as the "seed" number for the RNG (computer RNGs require a seed number to play with). Along the way, it also tends to "trend" certain ways that produces an Abnormal distribution, that can look (without looking at how the results clump) "normal" if one just looks at the end data, not how it fell during the experiment.

For example, while a 2D4 experiment should result in a mean of 5, if you end up with 16-result test that goes, in order of roll (NOT sorted for result),

2 3 3 4 4 4 5 5 5 5 6 6 6 7 7 8

You'd suspect that there is some sort of error in the RNG code.

 

[Erekose]

If you consider it, you don't have a personal random number generator.
Everyone gets numbers from the same RNG.
So, for every random number that's generated for you, there are a bunch generated in between, for other people.
So, any streak represented for you, is not necessarily an overall streak.

To consider streaking overall, you'd want to track the awarding of hundreds of marties.

That is speculation; none of us are privy to the RNG code so we do not know if it's somehow instanced to the character or not. As someone else noted it may be seeded uniquely for each character as well.

All we DO know is that the RNG has been horribly broken for years.

BTW I got 3 Talon Bites in a row from Miasma the other day. From 3 consecutive kills.

 

[Sarsmi]

Well first off you would need a larger sampling. And secondly you would need to want to make a post about the RNG *before* you get results. Consider this: 100 people decide to go fight monsters and get mini artifacts. 99 of them get a random assortment of items and the 100th person gets 4 of the same sort in a row. Now, out of those 100 who is the most likely person to post about it? This skews results for Joe shmoe forum reader, because really only the people who get 'streaky' numbers are going to post about it. No one is going to make a thread that says 'well heck I got totally random drops all night long'.

So yeah if you really want to test it out, plan on testing it out ahead of time and posting your results, and do it with 100 or so drops, not just a dozen.

 

[uoBuoY]

That is speculation; none of us are privy to the RNG code so we do not know if it's somehow instanced to the character or not. As someone else noted it may be seeded uniquely for each character as well.

If it is seeded uniquely then there must be an additional process to randomize the unique seed. Because a given RNG will always produce the exact same result or series of numbers when given the same seed number.

In RNG-using code that I've written I used this formula:
current (second x minute x hour x day x month)
to generate a fairly unique and random seed.

 

[Basara]

Actually, Sarsi, we've tested it with other aspects that are subject to the RNG... Mining (smelting ore), for example, over hundreds and thousands of checks.

It's pretty conclusive that, while overall the numbers come out to the expected mean, they also come in result clusters that aren't apparent from looking at the totalled data, only from viewing the list of experiments and looking for trends.

It's also interesting to note that, on tests of around 1000 Tailor BODs at 120 skill, instead of the 50/50 or 60/40 Exceptional/Normal ratio that is expected, the actual results are trending toward something between 35/65 & 30/70. (Smith BODs, on the other hand, are dropping as they are supposed to, after the fixes a few months ago, for the first time in years)

So, there's a number of RNG-based systems in game that have odd issues, that may or may not be releated.

 

[Tycolby]

Note that one of my Guildmasters had about 9 of his first 11 or 12 marties as Arctic Death Dealers, when Minors were introduced way back when, and ADDs are STILL about half the drops he gets from the original set.

RNG my butt!!!!! After 29 Artic Death Dealers I truely believe I am cursed!!!!!!!

:loser:

 

[Violence]

There you go reviving the Urban Legend of Personal Character Luck.

 

[Redxpanda]

I have noticed similar oddities with RNG. It's like everytime the server comes back up, it goes from favorable to unfavorable. I reached a point where i would only do activities every other day for the best luck.

 

[RoycroftLS]

It's pretty conclusive that, while overall the numbers come out to the expected mean, they also come in result clusters that aren't apparent from looking at the totalled data, only from viewing the list of experiments and looking for trends.

From personal experience, I totally agree. For a short period of time, I had a 99% LRC suit while I was searching for that elusive final piece of armor. While over time it probably did average out to 99%, I would frequently get the "need more reagents" message twice in a row. And twice in one day I got that message three times in a row. So I am inclined to believe that the game's RNG isn't quite statistically random.

 

[Sunrise]

I So remember thy days when I didn't need a math degree to play a silly game

But you all bring up good points....:bowdown:

 

[Erekose]

If it is seeded uniquely then there must be an additional process to randomize the unique seed. Because a given RNG will always produce the exact same result or series of numbers when given the same seed number.

In RNG-using code that I've written I used this formula:
current (second x minute x hour x day x month)
to generate a fairly unique and random seed.

That is pretty much how I've seeded every RNG I've ever coded with as well. RNG is a misnomer actually, the actual term is 'Pseudo RNG' because there really is no such thing as a random number generator.

 

[Llewen]

If you consider it, you don't have a personal random number generator.
Everyone gets numbers from the same RNG.
So, for every random number that's generated for you, there are a bunch generated in between, for other people.
So, any streak represented for you, is not necessarily an overall streak.

To consider streaking overall, you'd want to track the awarding of hundreds of marties.

You know, I don't think this is how it works. I think the rng is somehow tied to accounts, I also think it is somewhat broken, especially for certain things. UO is just too consistently streaky. Someone posted about getting virtue breastplates, well I know at least half of the dozens of 10th Anniversary arties have been gorgets. Something is up with the rng.

edit: That's what I get for not reading the whole thread. I see others have already made similar observations. Yes, the rng should be based on a seed which is based on the computer time, to as many places as is reasonable (ie. it shouldn't be in days/months/years, it should depend on, at least, thousandths of a second). That is the only way to get true random number generation with computers.

edit again: What made me think of this thread is I am collecting map fragments for the minotaur quest. When I first tried, I killed what seemed like dozens before I got a map fragment, then when I did it again, it seemed I was getting something like 2 for ever 3 I killed. Now I am on a streak where I am getting maybe 1 for every 6.

As far as game play goes, this is actually quite a serious issue, in game terms. It is just an annoyance when it comes to situations such as training, or looking for quest items, but if a similar thing is happening in pvp, which I am sure it is, it can have a major effect on the outcome of battles.

 

[Erekose]

You cannot get 'true' random number generation with computers. The best you can do is make it look random to the casual observer. That being said, the RNG code in UO has been bad like...forever.

 

[Llewen]

If you are basing your "seed" on say the last 12 places from .xx0000000000000 to .xx999999999999 of a second in computer time, that will be just about as random as you can get (I'm not sure how many places computer time is calculated to, but I know it is quite a few).

 

[Erekose]

I was just being pedantic. As I said, to the observer a properly done pseudo rng will look random. UO's is not properly done; we've all seen examples of that fact in gameplay.

 

[Llewen]

Speaking of pedantic, it should be possible to create a real hardware rng for computer using some kind of crystal and quantum mechanics.

Don't know what the proper name for it is, but an example of this is when you have water coming out of a tap at a low rate, you will get a steady drip at a constant rate. If you turn up the water very high, you will also get a steady flow that is not random, but if you find just the right rate in between water drops will drip from the tap in a truly random fashion (within a set boundary).

The thing is, you can do that same sort of things with electrical current, and certain kinds of crystals. Those random rates are explained by quantum mechanics theory (I hope I'm right on this... ). Using these principles it should be possible to create a true, hardware random number generator.

Of course this has nothing really to do with the topic at hand...