Tae Ew Failure rate increased HARDCORE

[Lling]

You just had a run of bad luck - I swear. It happens with everything, not just Tae Ew. It happens with cultural, it happens with PoP - and you want to throw up when it happens with elemental. Soon you will have a GOOD streak, so save up and celebrate.

Best of luck.

[Dvinn]

It's just the RNG.

Or should I say...the lack of a RNG. The RNG is no way random, it's full of hit or miss streaks. You got victimized by it, but don't bother telling SOE, they will just give you some garbage aboout how it's working as intended. If it was truly random you wouldn't go 0/10. Never.

[Kaaba]

Yeah you're unlucky, I tried 4 tunics the other night at 188 skill and I was 2-for-4.

[Hethral]

I think there's some misconception about what a random number generator is supposed to do here.

It is not supposed to guarantee that every single person gets a nice average non streaky average success set of combines for every set of 10 or even 20 combines.

It is supposed to generate numbers that considered over many thousands of values average out to the number midway between the high and low point of the numbers being generated. Each value generated in the sequence is also supposed to not be predictably relative to the past few values without looking at the sequence as a whole. This means that you shouldn't see 1,2,3,4,5,6 or 7,5,3,1 when generating numbers from 1-10 if you're the only person pulling numbers from that random number generator.

Now also remember that the server is generating the numbers, and those numbers are being used for attack rolls, damage rolls, tradeskill combines, concentration checks, resistance rolls, randomizing the length of spawns over 28 minutes. Every value that's generated by the random number generator is fed to whichever of these events comes up next. This means that the number of other people are doing tradeskills or engaged in combat changes how large the gap is between values in the sequence that you're getting. So if enough people are shooting arrows in between your combines it's entirely possible for you to get the same value every single time for your combine from the random number generator. Of course the number of people doing this would have to be absolutely immense.

Random numbers are most often generated by asking the CPU for one. In intel chips about 8 years old for example there's a sequence of 2^16 - 1 numbers, I think it's been a while since I did this in school. That sequence corrosponds to a formula that someone did their math thesis on proving that it generates the most random sequence of numbers over however many numbers it is. The remainer of this number divided by the range you want to generate numbers over is your random number. In order to not start at the same place every time, usually the placement in the list is seeded with the time to generate more random numbers. But seeing as how every event in everquest is ultimately started by a human in different orders ever time it's about as random as it can get by the time a dozen people log in and start killing things.

Basically noone's come up with a better way to make a random number generator than what SOE is using, odds are it's the same one everyone else is using.

If it was truly random you wouldn't go 0/10. Never.

If the random number generator could be truely random, then it's guaranteed that someone would eventually see a streak like that. I picked up my 36 d6 last night and rolled them getting 32 of them showing 6, the odds of it happening are unbeleivable, but if they're truely random it will eventually happen.

[Velurian]

I think there's some misconception about what a random number generator is supposed to do here.
(...)
So if enough people are shooting arrows in between your combines it's entirely possible for you to get the same value every single time for your combine from the random number generator. Of course the number of people doing this would have to be absolutely immense.

No, it really doesn't. A good pseudo-random number generator (a computer can't generate *truly* random numbers) will generate the same number multiple times in a row from time to time -- in fact, approximately as often as you would expect from truly random numbers.

From a good PRNG, about one time in a million, you would expect to get 10 on a /random 10 six times in a row. It doesn't matter what else is happening. Now, as you pointed out, think about how many things happen in the game -- or even in a single zone -- that need to be randomized. Probably literally tens, maybe hundreds, of thousands -- and in some zones, millions -- of random numbers over the course of a day. Chances that are literally one in a million will happen a few times a day. Probably not to you, but to someone.

There are random number generation algorithms that come pretty **** close to the theoretical ideal here, and many of them aren't terribly computationally expensive, so there's no reason Sony wouldn't use one of them. That said, I have no idea if they actually do -- they may have Bob, a third cousin of one of the programmers, holed up in a truck picking numbers out of a hat.

But in general, humans are positively awful at spotting true randomness, or creating it ourselves. Our brains are wired to recognize patterns, and that often leads to seeing them where they aren't. And that's where most of the "stealth nerfs" and talk about EQ having a terrible RNG comes from -- people recognizing patterns that aren't really there.

(That said, I *do* think that there's something wrong with the RNG, I just don't think it's nearly as severe as many people make it out to be.)

I picked up my 36 d6 last night and rolled them getting 32 of them showing 6, the odds of it happening are unbeleivable, but if they're truely random it will eventually happen.

Forgive my skepticism, but unless you show me a video of you doing that, I'm not going to believe that -- and even if you do, I'll think you doctored the tape. I'm not entirely sure that my math is correct (been too long since I've done any combinatorics), but you're talking about something that the odds are, I believe, quintillions ( = 10^18 ) to one against.

(Unless I'm misunderstanding what you're saying here...)

[Vapors]

You might not want to belive it. course.. you have any idea of the probability of someone useing a wooden stick to hit a round ball traveling over 100mph? Its one of the most improbable things in the world, but it happens all the time. As for eq's RNG it dosnt always feel fair but they have stated at severel times they have checked it to see if it was truly random. As for what system there useing I dont exactly know call to a chip probley, though i wouldnt mind seeing them going to something like a electromagnitic static detector off of a piece uranium.

[Hethral]

I can't say I blame you for not beleiving me, I know I wouldn't. Was the ****dest thing I've ever seen with dice. I'm not in the habit of videotaping my gaming sessions with friends, but seeing as I work at a place with video editing equipment I'm sure I could whip you up something given a few hours .

With up to 100,000 people simultaneously online at once, something that's a one in a million chance will be happening at least every few minutes on one of the servers.

[Reflan]

Incredibly unlikely things do happen. In a seven-handed game of 7-card stud one evening, I got as my first three cards(two down, one up) three 9's three times in a row.

[Silound]

I've played pen and paper D&D since conception of 2nd Ed. ruleset, many thousands of times. To this day, I will ALWAYS roll a natural 20 on my first combat roll vs and mob my level or higher. My friend who GM's and has a thief/rogue/whatever you rationalize him to be has a dagger+1, he ALWAYS hits for 6 when he throws it (1d4+1+1, 2nd +1 is str mod in 3rd edition). Weird, but its all in hte wrist

Silound
Minstrel of Prexus

[sulyen]

Assumption: that many players are pulling from a single pseudo random number generator.

I think this is the case. I don't have access to SOE internal information.

Hethral wrote concerning random number generators:

It is supposed to generate numbers that considered over many thousands of values average out to the number midway between the high and low point of the numbers being generated. Each value generated in the sequence is also supposed to not be predictably relative to the past few values without looking at the sequence as a whole. This means that you shouldn't see 1,2,3,4,5,6 or 7,5,3,1 when generating numbers from 1-10 if you're the only person pulling numbers from that random number generator.

Then each player will receive a sequence of numbers which FAILS the definition of a random number generator.

Many of us have read comments or have personal experience with the streaky patterns produced by the EQ RNG. This is further indication that the EQ RNG is not truly random on an individual character basis.

I can easily believe that Verant/SOE looked at this problem and decided not to commit resources to fixing it. Heck, I can believe that Verant/SOE isn't even aware of the problem.

One possible solution is white noise random number generators and the like. Here is a sample link. www.cryogenius.com/hardware/isarng/ Not the best example, but I hope you get the idea.

And I just had to comment on this.

Vapors wrote:

... Its one of the most improbable things in the world, but it happens all the time. ..

If it's improbable then by definition it doesn't happen all the time. Hence, what you wrote while a nice read is untrue.

[Kimba]

Failure rate

I am a GM smith and have had the worst failure rate ever since the last patch. I've been making pretty simple heraldic armor - the 1 infused ring items - and have made 2 items out of 30+ attempts. This is abnormally high :<

Kimba

[Lling]

Re: Failure rate

I am a GM smith and have had the worst failure rate ever since the last patch. I've been making pretty simple heraldic armor - the 1 infused ring items - and have made 2 items out of 30+ attempts. This is abnormally high :<

Kimba

At an unadjusted 250, I was in tears over the failure rate of heraldic. Then my streak was over and I've failed maybe 1 out of 10 attempts.

It happens.

It you get more than 3 failures in a row, force yourself to stop and try another day - not for the sake of the RNG but to stop slamming your head against the wall. It will get better - it always does.

[Reflan]

Yes, but the law of large numbers says that the short run numbers will not be a reproduction of the smooth average you see over a large run of numbers. When you look at a small set of numbers(like your thousands or even tens of thousands of combines over an EQ toon's life), it may never have anything that looks remotely smooth and even about it.

In fact, you may indeed virtually ALWAYS have horrible luck for the life of your toon's character. That number of combines is a drop in the bucket.

[burnttoastv]

Blew all my luck for a month this evening. Someone brought me enough for two combines at blessed full mithril bp. Both combines succeeded. The tae ew shield I had tried 10 minutes earlier had failed..go figure. First smithing I've done in a few weeks.