Thought On: Crafting's "hidden formula" %

[Covenant]

This may be an old topic, but hear me out. There obviously must be a hidden formula for crafting so that when a skill say 50% it's not really 50%. It might be 50% - X% depending on the level of the craft. Because whenever I use a "50%" skill it works great at lower level, but is absolute trash at max level crafts.

I'm reminded of the old skill (don't remember the name) that raise the crafting 5 level to make things "easier" some remnant of that craft formulation must be lurking around.

I'm not asking SE to change to to be a true "50%"(though they should, since your capped at crafting depending on your level anyways), but be more "truthful" and say under description "50% depending may be modified depending on level of craft".

I'm griping because I'm doing glowy weapons an that final item is like failure streak after failure. There's NO WAY its just a flat 50% chance.

And don't tell me it's basic statistics, because if failure streak can happen then win streak can also happen.

[MistakeNot]

Win streaks can and do happen.

If you want us to believe your allegations, you will have to provide some hard data to show that you have not just been unlucky.

[AmiableApkallu]

It's basic statistics.

Seriously. If you have actual observations, you'd be able to calculate the likelihood of seeing what you saw if it was a true coin flip. You'd be able to work out whether some other probability better explains what you saw, and what that other probability actual is. Etc. Etc.

Without numbers and data, all you have is feelycraft.

[Jojoya]

Google "RNG seed".

While basic statistics are accurate in a "natural" environment, gaming RNG isn't natural. It's influenced by RNG seeds and the seed chosen can vary from player to player, character to character or even action to action. If there is something biased in the seed picked, results are going to fall outside of what basic statistics tell you to expect.

My personal "feelycraft" experience from completing Castle in the Sky is that most of my crafting classes fall in line with what basic statistics suggest should happen. I would get both the good luck and bad luck streaks but overall results were fairly close to the stated success rate. The one exception was Armorer, which had a terrible success rate on Rapid Synthesis and Hasty Touch even when the condition was Centered. It didn't surprise me since I had noticed that my Armorer RNG seemed off all the way back in Stormblood though back then I had assumed it was just "feelycraft".

Your best course is to avoid using actions that list a success rate. It shouldn't be necessary to use them for anything other than Expert crafting and the Expert crafting for the relic tools is easy enough in difficulty that you don't have to use them unless you're shooting for max rating every time instead of being satisfied with at least minimum rating.

[Covenant]

I was hoping someone had already ran this statistical calculations already (since I'm lazy and my math days are way behind me).

[Frizze]

This comes up on the forums... about once a year. If you go deep enough into the forums, you can find a lengthy dev post that explains with a little bit of detail the RNG generation method they use. At least one or two of the old timers on the forum have the link saved and can pull up the explanation when needed. The short version(as i remember and recall) is that while its a super long interval between instances of actual RNG repetition, any given short string of attempts is equally likely to be good bad or neutral(in other words, random). Most people who come here to ask or complain about it have a very limited sample size that isnt statistically significant enough to actually show anything. Then on top of that, the way most peoples minds work they are more likely to "remember" a string of failures then an equally unlikely string of successes(for many the successes feel earned while the failures feel cheated even though both are just RNG - people can be irrational).

With that said, anyone who has: done a significant number of trials, recorded all their results, and shared the details - every one of those people comes back saying that the percentages are essentially what the screen shows(essentially in this case being +/- like 1 or 2 percent - the sort of statistical noise that would go away with enough tests). And while i dont know all the formulas involved in crafting, im 95% sure that crafter level is factored in to the progress and quality calculations only. In other words, crafting something lower then your level and you make more progress/generate more quality. While making something above your level(and keep in mind, star recipes are considered to be higher level), you dont progress or generate quality as quickly. Those changes are enough to alter the difficulty of a craft without needing to adjust success percentages behind the scenes.