Overmelding needs a QoL change

[DrWho2010]

people always make suggestions to make things OP even though they don't "feel" they are. this is a fallacy. The success rate is based on the difficulty of the pentameld and intentionally designed this way. i feel the level is just right the way it is.

[Ryaz]

My only issue with overmelding is SE's streaky algorithm for RNG. They use one of the best RNG algorithms out there... and it actually is extremely accurate with a large sample size. However in the short team, it tends to hit the extremes. For example, yesterday I melded 3 rank 8 materia at 17% chance each right in a row. Then failed at a 22% chance like 20 times. It can get frustrating and it's why I tend to avoid using too many rank 8s when it comes to crafting gear. I'm not rich enough to blow 10 million gil on materia just for one set of gear.

[Plastics]

That's not how a 17% chance works. 6 tries does not equal 100% success.
Could be 3 tries could be 50.
Keep overmelding as it is but make it still faster.

That's not how probability works. See: https://math.stackexchange.com/a/165994

The expected number of attempts to attach a materia is indeed 1/p. What you're reacting to is the fact that sometimes you meld in 1 attempt, sometimes it takes multiples of 1/p. These are extremes of RNG, but it's a distribution, the probability of melding in 1 attempt is p, the probability of failing 6 17% chances is only 32.6%, and the probability of failing 12 is 10.6%.

The issue people have with nerfing melding is that it would mess with the economy, e.g. it takes X materia for people to meld their gear, and that demand keeps the prices reasonable so there's a value to them. Mathematically, giving a 100% chance at the ceiling of 1/p is using the same amount of materia. You shouldn't consider 1 single person, you should consider everyone in the game, and on that scale the actual usage per person on each of those meld slots very closely matches 1/p.

This change would just make melding materia much faster, which is exactly why it should be done.