Grand Master vs AA Mastery

[Kronepsis]

Krone, the worse fail rate you can get is 95%, where is your 300% fail rate coming from?

What Twistagain said.

What Kron was saying was that all the percentage chances are calculations. At some point in the calculation, the server says "Hey...that's over 95%..so we're gonna make it exactly 95%." The question is "When does the server do this?"

[...]

In other words, if the AA calculation was performed BEFORE the "minimal chance" correction, there would be no difference in this particular case.

In other words, we DON'T know where in the formula AA Mastery kicks in. Someone should test it before people start going berserk about it.

And as for minimal success rates, all I can say is that if there wasn't one, how did I make a Grandmaster Smith's Hammer at 160 Blacksmithing skill?

[Kronepsis]

Dunthor,

When the JCM aa's first came out, I believe thats actually how it worked. I recall people getting near 100% success on valorium rings. Then they changed to how it is now. Instead of 72% chance going to 108% chance success (and then capped to 95%), it went 28% fail to 14% fail, thus 86% chance success.

I might be wrong on that, but I remember there being a big outcry of nerf.

The big outcry of nerf was that the JCM AA...stopped working. Period. They fixed that.

[Eileah]

Eileah,

If you purchase 1 rank of Brewing Mastery, that should be a 10% reduction in fails. If you then take the cleric, remove gear (as to minimize skill up rate) and create 1000 mino brews. You'll still end up with a bunch of skill ups, but Mino Brew should give no more then a 5% chance up till 136 skill, and you shouldn't cross that.

With no aa, 5% chance, would get 50 success.
If aa is 10% reduced fails, then a 95% fail reduces to 85.5% fail, or 145 success.
If the aa is as Kronepsis says, then you should get a >5% chance at skill 125 with aa. Since I doubt you'll see 124 skill ups on 1000 combines, you should still get about 50 success.


While the RNG could still have a wide margin of error on this, the difference between 50 and 145 success should be noticable.

Ok, will grind her 3 AA's starting tomorrow, and then sit down at the barrel, still don't know anything about parsing or how to do it though =/

[Eleena Transient]

very easy to "parse".

a) don't have any brew on you when you start.
b) don't drink any brew while you click
c) take the first 8 successes and put them in separate stacks, this will make sure that you don't accidently drop any with auto-invetory (up till 160 at least, and if you get that many we know its giving you a 15% success rate instaed of 5%)
d) keep count of how much yeast you buy (1 used per combine)
e) once you used 1000 yeast (50 stacks), count how mana mino brews you have in bags total


If its closer to 50, then you stayed at 5% regardless. If its closer to 145, then the 10% reduction was applied directly to the 95% fail rate. If its somewhere in the middle, we'll need someone else to do 1000 combines with no skill and similar stats but no aa for a baseline.

[Eileah]

Well she has her 3 AA now and lots of empty bags, so I will be brewing in the morning and will post the results then....

E

[Eileah]

Naked Cleric, Brew Mastery 1 and no buffs.

1000 combines, 41 successes and 103 skill ups.

Didn't seem to do much for success rate that I can see.

Eileah

[Humlaor]

Thank you so much for doing this test Eileah! I'm glad to see that they didn't devalue skill like I was worried they had!

Boomlaor

[Cartright]

oops sorry didnt see all the post

[Boleslav]

Thank you very much. The time you invested in grinding out 3 AA, and grinding out all those combines is appreciated.

I retract my previous rant on the subject.

Boleslav Forgehammer
Paladin of Brell in his 66th Campaign
E'ci - Sacred Destiny

[Verdandi]

I wonder if it's the description of the AAs that's flawed. Perhaps they modify your chance of failure on a percentage basis. So if you have a 5% chance, and get all three levels of the AA, your chance becomes 5% plus 50% of 5%, or 7.5%. If your chance of success is 50% then it becomes 75%, etc. I confess to having not followed all details of this conversation because the equations make my eyes glaze over, but, well, just a thought.

[Eleena Transient]

It is theorized the 'success' formula follows the equation:

S = Skill - (0.75*triv) + 51.5

I think a better way to think of it is instead as a fail chance:

F = 100 - S = 100 - (Skill - (0.75*trtiv) + 51.5)
F = 100 - 51.5 -Skill + 0.75*triv
F = 48.5 + 0.75*triv - Skill

Then, this fail rate is capped. Generally 5% as a lowest (with new exception for very trivial that scale down to 0%), and 9x% as highest. Its assumed that 95% fail rate is the highest it goes.

I think its the base F that gets modified by AA mastery, before its capped at 5/95.

So, assuming skill 50, making Mino brews:
F = 48.5 + .75 * 242 - 50
F = 48.5 + 151.5 - 50
F = 180

Thats a 180% chance to fail! Now, with brew mastery, we knock 10% of, down to 162%. Still caps to max fail rate.

Lets assume skil 100 (easy to get) making triv 335 things with mastery 3
F = 0.5 * (48.5 + .75*335 - 100)
F = 0.5 * (199.75)
F = 99.875

Still caps to max fail rate.

Lets assume a skill of 200 making triv 335 with mastery 3
F = 0.5 * (48.5 + .75 * 335 - 200)
F = 0.5 * (99.75)
F = 49.875

If this *is* how they coded it, then we needn't worry about the 'unskilled' using the AA to major advantage. It would put a 200 skill with 18 aa spent on par with a base 250 skill without the mastery (though a geerlok would push the 200 ahead). That seems relatively balanced.

[Eileah]

the equations make my eyes glaze over.

Me too, and I have a (very dusty) degree in math =P

[Dunthor Warsmith]

Me too, and I have a (very dusty) degree in math =P

I have a feeling a discussion of basic regular ol' algebra would make a lot of math majors' eyes gloss over (even moreso ones with no interest in teaching), whether their degrees are dusty or yet to be printed

[Boleslav]

Algebra discussions put me to sleep.

Boleslav Forgehammer (posessor of a dusty degree in Statistics)

[Verdandi]

I got a B in Basic Astrophysics six years ago. Amazing how quickly the brain turns to mush.

At any rate, Eleena's formulas are showing that it's entirely plausible that it's the existing caps that prevent people with low skill from having runaway success rates. That fits the existing description of the AAs better than my proposal, although it also means they are still somewhat misleading. I'm satisfied that this isn't a cause for great concern anymore.