GM smithing/tailoring success rates w/ Mastery 3, possible bug

[Xenephex]

You've got to be kidding me. You're gonna drink that SOE Kool-Aid and tell me everything is "working as intended"?

Not a chance. I didn't spend at least 24-48 RL HOURS of my life getting from 250-300, Pay 300K or more on Emeralds, Rubys, and Saphires, and do countless thousands of combines to then fail on 4 of 6 combines.

It's broken or collectively SOE Devs are no smarter than a doorknob.

Give me a break.

Trying to do this as politely as possible...

The operation of the RNG is counter-intuitive. Everyone who does trade-skills or reads these boards or both is certainly well aware of that. If people are asked to write down a string of random results (true / false, heads/tails, success/failure - whatever) they will not write down anything like what true randomness will produce; it's as simple as that.

So it's somewhat understandable for you to think that what happened to you is unlikely, even though, in fact, it is quite likely. If you skilled up to 300 in smithing, you certainly noticed counter-intuitive strings of skill-ups and lack thereof in the course of your run.

It is unfortunate and unlucky and you have every right to be upset about it and curse your luck, but it was well within the range of possibility / probability.

What you don't have a right to do is to insult the developers or leap to the conclusion that it is broken based on your own lack of understanding of how random numbers work. To continue to insist that you are right and (most) everyone else is wrong will eventually drift into the realm of self-mockery.

[drtyaty]

Trying to do this as politely as possible...

The operation of the RNG is counter-intuitive. Everyone who does trade-skills or reads these boards or both is certainly well aware of that. If people are asked to write down a string of random results (true / false, heads/tails, success/failure - whatever) they will not write down anything like what true randomness will produce; it's as simple as that.

So it's somewhat understandable for you to think that what happened to you is unlikely, even though, in fact, it is quite likely. If you skilled up to 300 in smithing, you certainly noticed counter-intuitive strings of skill-ups and lack thereof in the course of your run.

It is unfortunate and unlucky and you have every right to be upset about it and curse your luck, but it was well within the range of possibility / probability.

What you don't have a right to do is to insult the developers or leap to the conclusion that it is broken based on your own lack of understanding of how random numbers work. To continue to insist that you are right and (most) everyone else is wrong will eventually drift into the realm of self-mockery.

I understand random numbers just fine, and am a very educated individual. The reason I say it is broken is based on the investigation I did AFTER (unfortunately) this "string of bad luck" as you might call it occurred. For instance, I saw a thread, and I will try to locate it, wherein the author suggested that the probability of something like this happening is .00001 or something like that.

Give me a break. Should I have a string of "good luck" (as again, you might call it) I will report back and say that my conclusions are similar to others on this board who have said that absolutely nothing is wrong.

Edit: I over exaggerated a little. Here is the post from Reorxx at http://eqforums.station.sony.com/eq/...ending&page=1:

If the calculator at EQTraders is to be believed (and my math for that matter), a 300+5% attempt with a 386 trivial has a success rate of 77%. For 10 attempts your chance at failing all 10 is (0.23 ** 10) = 4.14e-7 = 0.00004%. You'd expect to see 10 such failures in a row every few million combines. Even a lower success rate of 50% you end up with only a 0.1% chance of 10 failures in a row. With the mastery3 AA 10 failures in a row is even more rare...you'd need several billion combines before having a good chance of it happening.

Furthermore, according to the EQ Traders Calculator, shouldn't my success rate be near 80%?

[Xulan]

The problem is is that the RNG does not give true randomness, and is therefore prone to streakiness. Therefore those probabilities are not correct.

It is just how it works in EQ. We've come to live with streakiness. If you want the RNG redone, that is an entire other beast. I remember Absor stating (some time ago) that they would try to look at evening-out things a bit more, so that the crests and troughs would not be so far apart. I do not know if this was done or not.

As far as success rate...

Here are "approximate" success rates as they are capped for gm armor and symbols:

bp/leg: ~60%
arms: ~70%
boots/helm/gloves ~80%
wrists: ~90%

Those are actually a couple of % off here and there, but it does represent the rough caps. See my post a couple of posts back where I listed my success rates.

[Xenephex]

I understand random numbers just fine, and am a very educated individual. The reason I say it is broken is based on the investigation I did AFTER (unfortunately) this "string of bad luck" as you might call it occurred. For instance, I saw a thread, and I will try to locate it, wherein the author suggested that the probability of something like this happening is .00001 or something like that.

Give me a break. Should I have a string of "good luck" (as again, you might call it) I will report back and say that my conclusions are similar to others on this board who have said that absolutely nothing is wrong.

Edit: I over exaggerated a little. Here is the post from Reorxx at http://eqforums.station.sony.com/eq/...ending&page=1:



Furthermore, according to the EQ Traders Calculator, shouldn't my success rate be near 80%?

As has been stated numerous times on these boards, the success rate for gm cultural combines is capped. 60 percent for bp and legs, somewhat higher percentages for the other items.

You are quoting a post about the chances of having 10 failures in a row, but your example said you failed 4 out of 6. There's a huge difference statistically.

[Bobaten]

The problem of course is the success calculator is not right for these (success capped) combines. It might be a good idea to add a note about success capped combines to the calculator to help people better understand what to expect.
As far as going 0 for 4 on your combines, you have about a 1 in 200 chance of doing that. To put that in perspective the odds of going 0 for 2 on a 95% success chance combines is 1 in 400, and I have done that more than once.
However, those odds assume those are the only combines you did(which of course is true so far in your case). For example the chance of failing 2 in a row on a set of 3 95% success combines is 1 in 210. Sadly, you started out at the bottom, but don't worry, time has shown (even to the people who started this thread 5 months ago) that in the end success rates are right where they are expected to be, so just keep trying and you will succeed.