About RNG..

Like. Okay. I guess I already know people will flame me or get mad but I just feel like pointing out a thing...

People complain about the RNG and the usual response seems to be "random is random" etc.

Here's my point: I understand, and I bet several other people understand, that randomness is random. That means it's very..very random..and even stuff with low odds can happen. Whatever can happen. Okay, sure.

But can anyone be sure that the RNG is proper? Like.. no programming error? No purposeful rigging? You just have to accept that it is correct? Programmers don't make mistakes? Computers don't make mistakes?

I don't think anyone has really conducted an extensive test, so it's hard to say. But that's just my point. We know a perfectly working RNG works perfectly. Now we just need to know it works perfectly.

Now I'm not saying this is the case in FFXIV. It's just a comment.

The issue with coil is that for whatever reason they did not split the available loot evenly between the chests in Turn 1. All of the primals however are one chest, now there is a higher % chance of seeing a book drop but they are two different classes at 50 so that does not count. You would need to have someone gather the information from at least 1000 or more kills to determine if there is truly a skewed pattern for drops. Obviously this could be the collective information from several players.

Personally I've seen one bow drop in around 120 Garuda kills which would illustrate for me personally a .0083(Repeating of course) drop rate on the bow. I imagine another player has a much different drop rate out of their set.

I played through FF14 and 88% of the game was great but it's still no quality trash. Sorry about that SqEnix. You know how it goes.

Serious though, it isn't random, that's the thing. I've made hundreds of Mythril Ingots in the last week. No good con, I get 78% and my HQ is around 30% so it's not random. It's stable fail and it's not just FF14. No game I have ever played had legitimate numbers in their so called RnG. I don't know if it's a Dev cheat or just not efficient but it never works out to what it's supposed to.

Some cases like Gathering it never can. You can't just chop a third of the strikes off the nodes and expect things to ride along tickity boo. It just don't work that way.

As for the Garuda book mobile, I'm still convinced they did that on purpose expecting people to be leveling the latest class over all others.

No computer is truly random, there will always be a certain pattern.
Basic mathematics.

Just wanted to add my thought to this as well. When you are dealing with a computer generated random drop rate, you really aren't completely random, because the program has to have some sort of logic to create the seeming randomness... unknown to us but known to the programmers.

Perhaps one of the factors it takes into account is how many of a given class were created server or game wide as opposed to present on a raid. Therefore, multiple u-nplayed or under level character's could be affecting the programmed drop rate.

Like Shioban said, technically it's not true randomness since it's not mathematically possible for a computer to generate truly random numbers without a source of entropy which I seriously doubt Square Enix (or any MMO for that matter) bothers with. But computers can come up with good enough pseudo-randomness for things like RPG rolls, so that's not really relevant for our particular case.

For all intents and purposes the RNG is sufficiently random. Unless SE is picking a bad seed for their random number generation (which wouldn't affect the pseudo-randomness of a particular sequence of numbers, merely the rate at which the sequence recurs), then it is not "broken". You'll have lots of people claiming different things like Ksenia up there, but the anecdotal report of one person (or even a hundred people) is not particularly useful when you have so many people playing the game. Yeah, some people might be really, *really* lucky/unlucky and are landing what seem like one-in-a-million odds, but... there are over a million people playing this game, so that's not particularly surprising.

The randomness, or even lack there of, of a system does not matter because the RNG does not serve just one function or just one guy. Let's suppose I decide that in my loot dropping algorithm I will randomly inject a string of 100 Dragoon loot drops in a row in any loot drop that could have dragoon loot, and beyond this anomaly everything else is random. This anomaly will only happen at most once per day. So say some guy got Dragoon loot running AK 5 times in a row, and a common thing to say would be this is clearly the example of faulty loot drop. Except that my 100 Dragoon loot drop algorithm applies to every loot dropped on the server. It's pretty hard to imagine less than 99 loot dropped between 2 AK runs in a particular world, let alone 5 AK runs in a row. Therefore, even when you know there's clearly a 'Dragoon drop time' thing on, it is still impossible for any individual to see this. The only way you can see this "Dragoon drop time" is if you somehow contacted every guy who was also playing the game at the same time as you and see what they dropped, and only then would you be able to see that indeed 100 Dragoon loot were dropped in a row somewhere in this world.

The RNG does not even remember what you got last time. It is just a very long sequence of numbers that serves virtually any function in the game. Even if the sequence itself is not random, there's still no way for any individual to see it because what appears to you as two consecutive results is actually separated by thousands if not more RNG calls required by other people.

If you're that certain the RNG is broken, you need only prove it.

Burden of proof is on you.

"What if"s without evidence mean nada.

Love these threads, and yes no RNG on a computer can ever be truly random. Hence, (see below) ...

Quote:

---

Originally Posted by Borfin

If you're that certain the RNG is broken, you need only prove it.

Burden of proof is on you.

"What if"s without evidence mean nada.

---

I don't think there's anything wrong with the RNG but how do you expect people to prove something regarding droprates? I went 1/55 on my ifrit's sword which is... a stretch but is the way of randomness. In the 20 Garudas I had to do for scholar book, I saw SMN book 9 times. Again, odd.

How would you want a person to prove it other than providing personal experience that demonstrates it?

Why can't the proof be on the defenders. You claim something is fact, so prove it. Do 100 garudas and log the drops. It's reasonable to assume you'd see at least 8 of each drop, right, or even 6? Except that, according to a lot of people, you won't. You'll see a disproportionate amount of summoner books.

Isn't it odd that summoner books, the only class 2.0 added, *feel like* the most common drop? One could argue that smn is more common than sch because SE felt that a summoner is more likely to have competition for the drop (not true, but they might have felt it)

Quote:

---

Originally Posted by Borfin

If you're that certain the RNG is broken, you need only prove it.

Burden of proof is on you.

"What if"s without evidence mean nada.

---

It would be actually harder to prove that it is truly "random". And seeing how no one has been able to prove this yet... Aren't we in a draw?

Not to mention what has already been said about RNG and entropy, where you can never assure "true" entropy (aka randomness), you can only try to achieve a higher amount. It's a crappy thing really, and very complicated or as simple as you wish to make it.

If you win the lottery twice in a row you can be sure someone will be doing an official investigation as to whether you cheated or not. Even though you can't prove if something is truly random or not, you can be reasonably confident if you see odds that are on the order of lottery level luck. That said most of the 'bad luck' people report are nowhere close to lottery level odds, and the ones that are tends to be so improbable that there's no reason to believe they're true (usually they'd be events that's equivalent of winning 2 lotteries in a row).

RNG is usually based on current time, coming from the year all the way down to a microth of a second. It creates enough variance that it might as well be completely random.

Problem is people tend to focus more on the failures and ignore successes. For example, I can complain all I want about Hasty Touch repeatedly failing and never working, but if I look at the big picture over a long period of time, it usually comes out to around 70% success with Steady Hands.

Another issue is the size of the samples. Even 30 Garuda kills is extremely small and even then there's about a 2.9% of not seeing your weapon. You may say you've seen 9 summoner books and no scholar books and I can claim it's the opposite for me. This is also assuming that every drop has an equal chance of even dropping, which could very likely not be the case. Unless someone has the database from the game to confirm it, I honestly wouldn't be too surprised that the weapons that don't seem to drop that often from Garuda end up appearing more often from Titan.

Don't mistake the term pseudo random number generator for meaning not-even-close-to random number generator. For all purposes aside from incredibly demanding monte carlo simulations and cryptography PRNGs can be thought of as perfectly random.

The most common PRNG has a period length (number of runs before it will repeat the sequence) on the order of 10^6000.



Nothing in an MMO will ever even come close to pushing the limits of 'randomness' for any recent PRNG.

It's all in the odds, the odds are that you will get a weapon or an item from X chest 100% of the time. The odds of a specific item never change. In terms of primals the odds are 1/9, of any specific weapon dropping. And the odds never change, no matter win 9, 90, 900. So it is very possible with the odds never changing to in fact never see an item drop ever. It's like a container filled with colored tags, you draw a tag, and the odd is 1/9 if it being the color you want. Then you put the tag back in, shake, and draw again. The odds are again 1/9. I have done garuda and ifrit over 100 times each and i have not see all the weapons drop yet. And i may never see them, the odd of 1/9, is not the same a 1/9 percentile.

This is one of the game models i hate, because you are not guaranteed your item after the 10-100 or 1000th run. I have seen people get item X 1/1 kills as hey hit gold on the odds, i have seen people 0/100 with the same identical odds.

Turning RnG in to a science is fine for the Forum but game play wise it's another matter.
Run Garuda for 3 weeks a hundred times, get books every single time. Not 8 classes, miss the one you want but ALL books and the RnG becomes rigged.
I farm Mahogany, base 82% and get 1 in for on a node, you obtain nothing. Hit Field Mastery II and boost it to 97% and still get a you obtain nothing on every node? I just stop hitting Field Mastery II and mash insert.

We stop trying and that's the threat of funky RnG. People stop bothering.