Your WTF?! Gathering Probability Moments

[zdub303]

From some of the RNG i experience with this game, it is quite evident that the RNG algorithm they use in this game is rather poor.

Just FYI, you can never program a truly random random number generator. The program has to use seed values which, by their nature, affect the resulting generated number. Some RNGs are better than others, and with the amount of '1 in a million' events that occur daily to myself and other crafters in my FC and LSs, its clear that the RNG in FF14 isn't very random at all.

[HoroBoro]

From some of the RNG i experience with this game, it is quite evident that the RNG algorithm they use in this game is rather poor.

Just FYI, you can never program a truly random random number generator. The program has to use seed values which, by their nature, affect the resulting generated number. Some RNGs are better than others, and with the amount of '1 in a million' events that occur daily to myself and other crafters in my FC and LSs, its clear that the RNG in FF14 isn't very random at all.

Actually, you can have a pure RNG. It's call hardware: http://en.wikipedia.org/wiki/Hardwar...mber_generator

Given that you maintain it and have competent components and filters, it's completely random.

I hope SE uses some implementation of this. Otherwise, we could game the RNG. Remember FF12 cure list?

And I know this will probably suck the life out of this thread, but we're wired for patterns. We tend to remember the weird things rather then what we perceive to be normal. Thus, if we feel that RNG is acting weird - it's probably doing it's job.

[zdub303]

Actually, you can have a pure RNG. It's call hardware: http://en.wikipedia.org/wiki/Hardwar...mber_generator

Thanks for the link, interesting read. I highly doubt, considering the server uses a 0.3s server-side position check delay to conserve resources, that they would actually use a hardware based random number generator. Esp considering how slow it says those solutions are.

More likely they use the software based PRNG with seeds values taken from recycled output.

[HoroBoro]

Thanks for the link, interesting read. I highly doubt, considering the server uses a 0.3s server-side position check delay to conserve resources, that they would actually use a hardware based random number generator. Esp considering how slow it says those solutions are.

More likely they use the software based PRNG with seeds values taken from recycled output.

That's an interesting idea. Given my own experiences with crafting, I'm actually inclined to believe you (I sometimes feel like I can predict the next "good condition" or the next failure).

The next step would be to... I guess, pick up all the popular PRNG algorithms and put data into them I guess. Maybe find out what the formula is. Apparently the PRNG was a linear congruential generator: http://www.kotha.net/ff12random/rand...tml#a_overview

I doubt anyone has the motivation, and I doubt they would use the same type, (I certainly don't and wouldn't) but maybe we can take it from there? There's nothing in the TOS against using math.

[CJReece]

I was mining for a Leve, and I needed two more stones with 6 tries 80% chance, only got 1. Almost facepalmed myself into space.

[johnbeck1000]

same here. I kept track one night while mining just for S's and G's... an 88% success was 70% but that was only 100 chances, maybe it would equal out over more opportunities or another miner is getting the other side of that which is probably the deal. The negative mathematical long shot happens to me more often then not and I've never hit 4/4 on HQ mining items, ever

[Dorin]

7 HQs in a row at 15%. It was so ridiculous I even took a pic:

[TheTen]

The RNG is definitely a little wonky. I always have bad luck or good luck in a row, (missing 6 out of 8 times on an 80% chance, getting 3 of 4 HQ on a 12% chance etc.) which, sure, can happen sometimes... but all the time over two different characters? That's a pattern.

Luckily, I normally farm things for long enough that it all balances out in the end, anyway.

[darkvision]

i had a funny moment, with 15% HQ rate and i got 27 HQ gathers in a row. i did not use any ability or get any node bonus

i have had 50% HQ chance and 99% success only to fail all attempts

[Lstkaws]

It definitely has to do with the algorithm they used. I have noticed this a lot since gathering, so much so that I am now keeping a spreadsheet. I would guess that even with a huge sample size the percentages you see on the screen will not be valid. I plan on keeping the spreadsheet over a couple of months and see how it comes out.