Fix Rapid Synthesis so it actually hits at 50%

[GTK0HLK]

-cut-

I wasn't talking that he was talking about that. it's clear where my part of it is.
theirs is obviously in the OP.
<I was mostly asking for people to cut the back in forth in a thread where everyone can simply just say what they think without going to subjective. and then thread end.>

[dspguy]

"I don't understand probability and statistics."

If the OP ran Trial Synthesis (to save gil) and just spammed Rapid Synthesis and tracked success and failures - and did this 200+ times, they'd see it is closer to 50% than 25%. And the variance from 50% would work itself out if they kept going.

What the OP is seeing is what we all see. Confirmation bias. I'll never forget the 175 melds it took to put a 5% materia on an accessory years ago. However, after that, I took to tracking all my melds when I was pentamelding and compared them with that the game tells me the odds should be. The result over a long period of time? Their RNG is pretty spot on. Over a short period of time, it is hard to see it and we only see what we allow ourselves to see.

[Ayche]

You gotta remember to save up your luck so you can use it at crucial moments!

[Deceptus]

Welcome to RNG, over thousands of people.

It averages to 50/50.

Same as how a 7% meld can go on the first try, or 300 materia later.

At a 7% success rate, you'd have a 97.34% probability of getting a meld at 50 attempts, and a 99.38% probability at 70 attempts.

If the OP ran Trial Synthesis (to save gil) and just spammed Rapid Synthesis and tracked success and failures - and did this 200+ times, they'd see it is closer to 50% than 25%. And the variance from 50% would work itself out if they kept going.

What the OP is seeing is what we all see. Confirmation bias. I'll never forget the 175 melds it took to put a 5% materia on an accessory years ago. However, after that, I took to tracking all my melds when I was pentamelding and compared them with that the game tells me the odds should be. The result over a long period of time? Their RNG is pretty spot on. Over a short period of time, it is hard to see it and we only see what we allow ourselves to see.

Maybe, maybe not. If they did run that trial for 200 attempts and it came out to 30% success rate would you still discount it as not enough attempts? 500? 1000?

At what point does data become valid if it doesn't agree with what you think it should be? 100 attempts is a decent sample size and shouldn't be scoffed at, it should be considered.

[Limecat]

You exist to make up for that one guy who's nailed 99% of all his non-100% attempts at melding/HQing.

[dspguy]

At a 7% success rate, you'd have a 97.34% probability of getting a meld at 50 attempts, and a 99.38% probability at 70 attempts.



Maybe, maybe not. If they did run that trial for 200 attempts and it came out to 30% success rate would you still discount it as not enough attempts? 500? 1000?

At what point does data become valid if it doesn't agree with what you think it should be? 100 attempts is a decent sample size and shouldn't be scoffed at, it should be considered.

Do it yourself and find out. Not necessarily with Rapid Synthesis, it could be any RNG aspect of the game. I've done it.

https://drive.google.com/file/d/1WyS...usp=share_link

I ran /random 2 (which rolls a 1 or 2) and recorded the output. Figures are in the link, but for those not wanting to click the link, the gist is:

This simulates the 50% success/fail of Rapid Synthesis (RS).

First 189 rolls had the following results: 100 1s (53%) to 89 2s (47%)
Next 43 rolls had the following result: 15 1s (35%) 28s (65%).

Collectively, across all 232 rolls, it was:
49.5% 1s
50.5% 2s

There were also multiple instances of consecutive 1s or consecutive 2s which would have looked like lots of RS successes or failures.

It didn't take that many rolls to see the results begin to converge towards what we'd expect.

[Moonlite]

I always find SE percentages seem to be fairly accurate when you break the 1000 attempt of something.

[Midareyukki]

I know it's not the same, but didn't mathematicians realize that in a practical 50% probability, the actual application won't be 50% exactly? Something about how a professor told his students to flip a coin 100 times and register the results. The ones who didn't bother got closer to 50% than the ones who did.

So it's normal if it isn't 50% of the time. Because each instance is randomized. The game isn't aware of whether you have 50% exactly of all uses of Rapid Synth, and if you try to ensure that, you'll make it predictable.

[Liam_Harper]

Rapid Synth is definitely 50% and not bugged. That is how rng works. It's random. It could be 50/50, it could be 10/90, it could be 90/10. If it was always consistent and even, it would be a set pattern.

That said, I loathe Rapid Synth in Experts. I have a good success rate on Experts and I don't care if it's rng, that doesn't mean it's fun rng or good design. It's easily my least favorite part of Expert crafting. I hope they scrap it and replace it with something more consistent some day.

[Midareyukki]

tbh OP, try going on DNC and try to get a proc. Procs are 50% there, and you won't get them every other press. Because that's how RNG works.