Random Crafting Question Thread I: HQ secret cracked!

[Damane]

Well I finally got a skillup, so it was just a really huge string of bad luck. Nothing like a string of bad luck when you're spending 300k/synth, right SE?

wait until you hit 104 and then 105... nothing feels better then spending 2million hopeing for a HQ (T1) getting no skill ups and haveing a nonexistent market for the HQ piece lol, the salvage ugprades to usukane +1 further destroyed tenryu +1 pieces from the smithing side ugh

[Taint]

wait until you hit 104 and then 105... nothing feels better then spending 2million hopeing for a HQ (T1) getting no skill ups and haveing a nonexistent market for the HQ piece lol, the salvage ugprades to usukane +1 further destroyed tenryu +1 pieces from the smithing side ugh

This ^

I was going to bite the bullet and finally take smithing to 110 but Usukane+1 pretty much crapped on anything that smithing had going for it. (not much to begin with)

[NynJa]

I'm just going to 104, none of my crafts have touched anything past 104 really, SE's idea of how they want shit to work from 105-110 is fucking retarded.

[dasva]

Yeah seriously. The only recipe to get gold from 108 to 110 costs like 12-13mil a synth isn't T1able currently and is garbage.

At least put these stuff on some nms that can be low manned. None of this allaince only crap

[Buffy]

I took cooking and woodworking to 110. But both have realistic paths to get there. And arguably the increased hq tier on red curry buns will pay for cooking over time. Woodworking was a total loss though. Cloth craft is next, but no t1 while skilling up because no torque blows.

[cdgreg]

Cloth craft is next, but no t1 while skilling up because no torque blows.

A friend has a 93 cloth mule I offered to skill to 110. Once I realized there is no torque to T1 post-100 it really turned me off from the idea lol. It will probably sit around 100ish until the torque is introduced.

[urat]

Are any of the pieces of cursed togi set worth making and selling for HQs? I dont play jobs that use it much so I'm not sure if it's worthwhile trying to make the set.

[Byrthnoth]

Has anyone ever had the Trituration Holy Water recipe yield 12 Holy Waters?

I just noticed tonight that out of all the Holy Water I've made with that recipe (over a hundred stacks), I've never seen the supposed HQ3 yield and see the HQ2 yield slightly more often than I'd expect otherwise. I'm going to guess that the HQ3 is actually 9 Holy Water.

[Xerlic]

Is spam feeding earth fewell on nuggets still the best way to powerlevel synergy? All the info I've found on Synergy is circa 2010.

[Buffy]

Yes, essentially. Can use a crafting stall to cap 80.

[Buffy]

I'm gonna hate myself for this...using the same method I am using for the shaper's shawl testing, what would be a good recipe to test the effect of the different Rusk(s)?

What I'm envisioning:

Two characters, same level (70 alchemy), same recipe (Potion - 40 cap).
One using food, one not.
Craft at same time.

Potion would be target because the characters would be 11 over cap, but not 31 over cap, and because the recipe has four results that are identifiable. Plus its fucking cheap as hell.
Potion, +1, +2, +3, etc.

One of the rumors of Rusk is that one helps you with T3 while another helps with T1, etc.

Am I overlooking anything?

And can a strong math person tell me how many synths I need for a reasonable sample size? Bryth, Septimus, Yugl maybe. Someone who understands the math regarding number of data points in a set and how many total synths I would need to start making conclusions, because I have no idea if 500 is enough, or 5000, or what.

edit: I'd also like a recipe I can get 51 over the cap on if anyone can think of anything. Craft doesn't matter, but it would be best if it was level 19 or below. I ruled out Black Ink because it doesn't have a HQ3 result and SE has fudged with that recipe in the bast (lowering the HQ rate). I'm having trouble finding a recipe that works. Tsuara doesn't have a HQ3 result either.

edit 2: I derp'd, I thought Tsurara had a 33/66/99 distribution, but it's 10/20/50/99, so it's a perfect recipe to try. I might skip Potion entirely.

Questions still stand though:

How many synths do I need to do total for 95% certainty rate or w/e is really good?

[Byrthnoth]

It depends what the effect strength is.

You can decompose it into a series of binomial distributions:
Success Rate - Did it succeed or not? (95/5% split expected)
HQ Rate - Did it HQ or not? (6.25/93.75% split expected)
HQ1 Rate - Was it HQ1 or HQ2/3? (75/25% split expected)
HQ2 Rate - Was it HQ2 or HQ3? (75/25% split expected)

This is a slightly different model than the one proposed here: http://www.bluegartr.com/threads/571...tion-read-this!
because it assumes that the HQ Level is split into two different comparisons, which is actually kind of consistent with what we know about the way SE does other things. Otherwise everything matches the model proposed by Aurik.

Assuming the above probabilities could hit +/-1% accuracy at the 95% confidence level for a Tier 1 synth would be:
Success rate - ~2,000 synths
HQ Rate - ~2,400 synths
HQ1 Rate - 120,000 synths
HQ2/3 Rate - 480,000 synths

Assuming the above probabilities could hit +/-1% accuracy at the 95% confidence level for a Tier 3 synth would be:
Success rate - ~2,000 synths
HQ Rate - ~10,000 synths
HQ1 Rate - 15,000 synths
HQ2/3 Rate - 60,000 synths

So if you're interested in investigating the HQ distribution, you're going to need a really really large sample size and should use a Tier 3 HQ synth.
However, if you're interested in the basic HQ rate then it is best to use a Tier 1 HQ synth because the lower probability of HQ makes small changes in HQ rate easier to determine.

Problems:
* Keep in mind that doing this number of synths does not necessarily mean you will have exact %s at the end of it. You could do 60,000 synths and end up with a 75.5%+/-1% HQ rate.
* The other problem is that determining synths to 1% accuracy would be far, far more accurate than any previous data, so you'd also need a control case if you wanted to go to 1% precision. I'm not confident that the food causes more than a 1% change in HQ rate, so this might be necessary.

[Buffy]

So, if I understand you correctly, I would need to do 480,000 synths WITH EACH FOOD to get to 1% certainty? So that's 1.5 million synths on the character using food (480,000 x each type of Rusk -Sugar/Chocolate/Coconut) and another 1.5 million on the control character?

Or did the big numbers make my head go boom somewhere?

[NynJa]

Why not just do 1 without food, and 1 with chocolate rusk..and see what the difference is, and work from there.

[Byrthnoth]

So, if I understand you correctly, I would need to do 480,000 synths WITH EACH FOOD to get to 1% certainty? So that's 1.5 million synths on the character using food (480,000 x each type of Rusk -Sugar/Chocolate/Coconut) and another 1.5 million on the control character?

Or did the big numbers make my head go boom somewhere?

If you're interested in teasing apart the HQ distribution, the best option would be to do 60,000 Tier 3 synths with the three foods and once without food (~250,000 synths).

[Buffy]

250,000 is still a big number, I could do it in like six months....assuming I did nothing else in game.....

Ok, we will start smaller....we'll do a stack of food each and go from there.

Bryth, my Tsurara recipe works? 4 clearly identifiable results, 51 over cap?

[Byrthnoth]

Yeah, you're actually 64 over cap and there are only 3 clearly identifiable results, I think.

Edit: Maybe there are four! I don't know. I guess I haven't made enough Tsurara.

[Hitou]

Probably been asked somewhere before, but does anyone know how much skill stalls give by themselves?

[NynJa]

Wiki says tsurara goes 10-20-50-99.

Stalls give 5

[bechyni]

So leveling Bonecraft atm (thinking of trying to get it 100+) although the guides that I'm finding seem to be pretty out of date and give me shitty recipes. Any links to a newer guide would be much appreciated.