Math, Materia, and How to Succeed at Melding

[AmyNeudaiz]

In response to the OP: "I believe that the numbers are accurate. I don't have any hard evidence"

-snip-

P.S. This is a repost from another materia thread but i felt it applied to here too.

And from that same thread in reply to that very post with the unrelated fat cut out.

Although I will say that on average spread out throughout the whole community, it's about right. But statistically you're gonna come across people such as yourself who get that unlucky draw where you just can't get anything to work most of the time and it's statistically probable that it's gonna happen to people.

[Madrox]

And from that same thread in reply to that very post with the unrelated fat cut out.

His post was related to what he was talking about and this thread.
Please don't start..

There hasn't yet been anybody from what I've seen test the success rates with large enough data of melds to truly say whether or not it is correct. Everybody gives up on recording the data. The theories are backed up by not very solid information. I personally believe that since it is an automated system within the game, it is very unlikely that it is even going to be wrong in the first place.

I think it would help knowing what the percentages are based upon (100 melds, 1000, 10,000) because I am 99% sure they are correct for the reason stated above. It would also be good to know if the that the melds count the percentages are based upon go for each person or serverwide. My guess would be the percentages are server wide and are based on a high count of melds which would open the possibility of a larger gap of players getting unlucky nonstop, and getting lucky a lot. If the percentages were based off smaller numbers, there's less likely the chance people would fail 20 times in a row.
Mix that with potentially multiple people reading around the same time. I don't have a background in statistics and could easily be wrong, that just makes sense to me as an explanation of what's going on for people like the guy above me. Hopefully I was able to explain that well enough.

This thread is definitely a step in the right direction to understanding it.

[Alzelia]

His post was related to what he was talking about.
Please don't start..

There hasn't yet been anybody from what I've seen test the success rates with large enough data of melds to truly say whether or not it is correct. Everybody gives up on recording the data. The theories are backed up by not very solid information. I personally believe that since it is an automated system within the game, it is very unlikely that it is even going to be wrong in the first place.

I think it would help knowing what the percentages are based upon (100 melds, 1000, 10,000) because I am 99% sure they are correct for the reason stated above. It would also be good to know if the that the melds count the percentages are based upon go for each person or serverwide. My guess would be the percentages are server wide and are based on a high count of melds which would open the possibility of a larger gap of players getting unlucky nonstop, and getting lucky a lot. If the percentages were based off smaller numbers, there's less likely the chance people would fail 20 times in a row.
Mix that with potentially multiple people reading around the same time. I don't have a background in statistics and could easily be wrong, that just makes sense to me as an explanation of what's going on for people like the guy above me. Hopefully I was able to explain that well enough.

This thread is definitely a step in the right direction to understanding it.

The percentages are exactly that, the chance that you will successfully meld. The larger the playerbase the MORE likely that someone will have a terrible run of bad or good luck.

[Madrox]

The percentages are exactly that, the chance that you will successfully meld. The larger the playerbase the MORE likely that someone will have a terrible run of bad or good luck.

But with a smaller pool lets say 100 melds the chances would be better generally of giving successful melds? That almost makes them sound misleading, not incorrect percentages lol.

[Alzelia]

But with a smaller pool lets say 100 melds the chances would be better generally of giving successful melds? That almost makes them sound misleading, not incorrect percentages lol.

Let me explain what I mean by the number of melds you need to perform to "expect" a successful meld. I'll keep the math real easy with a regular calculator. Take the percent chance you have of success, say 22%. Divide 100 by that percent. 100/22=4.54. Let's round that up to 5. 5 is the minimum number of attempts that will generally provide an successful meld.

The reason why I say you can "expect" one of the 5 melds is because each meld will have a 1 in 5 chance to succeed, and you have 5 chances.

Mathmatically you would have had a 72% overall chance of winning. (a 1 in 12 chance of 2 successful melds)

Another way to put it is that in the long run, 1 out of every 4.5 melds will be successful.

[Madrox]

Let me explain what I mean by the number of melds you need to perform to "expect" a successful meld. I'll keep the math real easy with a regular calculator. Take the percent chance you have of success, say 22%. Divide 100 by that percent. 100/22=4.54. Let's round that up to 5. 5 is the minimum number of attempts that will generally provide an successful meld.

The reason why I say you can "expect" one of the 5 melds is because each meld will have a 1 in 5 chance to succeed, and you have 5 chances.

Mathmatically you would have had a 72% overall chance of winning. (a 1 in 12 chance of 2 successful melds)

Another way to put it is that in the long run, 1 out of every 4.5 melds will be successful.

That's simple math, I know that. Your post wasn't hard to understand and neither are percentages. But it that how they really implemented it. I'm just bringing up another possibility of implementation. I'll start crafting more and experiment. If it was that simple I wouldn't think we would be hearing about people always failing over ten times in a row. Hopefully they're all just upset and full of it.

[Alzelia]

That's simple math, I know that. Your post wasn't hard to understand and neither are percentages. But it that how they really implemented it. I'm just bringing up another possibility of implementation. I'll start crafting more and experiment. If it was that simple I wouldn't think we would be hearing about people always failing over ten times in a row. Hopefully they're all just upset and full of it.

Chance you will fail 10 times in a row (@25% each): 5.6%
Chance you would bitch about it if you did: 100%

1 out of every 20 times someones goes for 10 melds in a row they will fail, and I'm sure they will talk about it and remember it. You don't really talk about the other 19 runs where you got 2 or 3 successful attempts.

[Madrox]

The percentages are exactly that, the chance that you will successfully meld. The larger the playerbase the MORE likely that someone will have a terrible run of bad or good luck.

I think people generally see 25% and generally assume that at most it probably won't take more than 4 melds attempts to succeed, and not to have issues like failing 10,15+ times in a row.

It is understandable too, you shouldn't have to read into the mechanics to try and understand the true success rate. I believe that is what is causing the coinfusion. Percentages always work the same, but the way they are used and integrated can always change the outcome.

[AmyNeudaiz]

Please don't start..

My reply was to his reply and thus it fits and is on topic.

[Madrox]

My reply was to his reply and thus it fits and is on topic.

I was talking about his post. Lol.