Square Enix's RNG

[Lafiele]

I don't get why in every single game there are these people who will makes these claims of incorrect RnG with completely no backing at all. I feel like people throw around this word without even knowing what it means.

[Tupsi]

I don't get why in every single game there are these people who will makes these claims of incorrect RnG with completely no backing at all. I feel like people throw around this word without even knowing what it means.

To be fair, algorithms can be skewed.

[Pikken]

No, a couple hundred is not a large enough sample size.

Last night, I helped my 5th grade daughter with math homework. She was to throw a die 100 times and make a frequency table with the results. Before that she should also guess how many 4s she would get.

Lo and behold, she got twice as many 3s as 4s! But she did not blame the die for having bad RNG, instead she learnt something.

[Wilksha]

RNG is just a RNG. Though putting random and computer together isn't really all that random. Look it up, the actual name is pseudo-RNG, meaning it attempts to be random like but really isn't because a computer can't just pick a number out of thin air like we can rolling a 100 sided die. The only thing going against you is likely the time seed they usually use for RNGs. So don't blame the RNG, blame the time! (Unless SE doesn't use time to seed the RNG, then blame whatever they seed)

[Ricky]

Understand and accept that anything under 100% is NOT 100% .. I did an HQ Rose Gold Ingot at 2% HQ chance, unexpected things happen

[Pseudopsia]

Random is not consistent. If you want to prove the percent is wrong you will need to post hard data with a very large sample size e.g. 1000 attempts.

[Rivienne]

I would like to throw my two cents in, overall the RNG is functional, but I would say it is broken to the point you have to adjust any percentage you see. I think everyone who has done a lot of crafting or gathering (or followed a lot of loot drops) knows, there are highly unlikely occurrences both directions, and "unlikely" strings of failures and success are common, when they should be statistically rare. Overall they tend to average down to a baseline, but that baseline is not the listed number in my experience.

Ignoring failures, look just at HQ for the moment. When I first started playing with using Unearth I did some tests to see how well it worked, and if it was worth the waiting for GP every few nodes, or paying for cordials.

Copper Ore @100% chance, 15% HQ. Using Unearth at every single node, so 25%, and 30% for +5% nodes.

I mined a total of 75 nodes(I was going to do 100 but I can only take mining for so long....) , so ~300 total ores mined. At the end I had 38 HQ.

That is obviously not 25%, more like 12%. While I haven't done another test to further verify this or do an even larger sample, this upholds my experience with RNG in this game: subtract ~10% to get the real world average percentage.

[DragonFlyy]

snip

RNG doesn't mean if it's 25% chance that you will get 25% HQ by the end of a run. It means when you click on that node you have a 25% chance to get 1 hq. The next time you click, you have the exact same chance. RNG means that even in 1000 gathers, you could get 1 hq.

The same thing happens with a dice roll. It's completely random. Try it sometime.

That is how random works. Yes computers have pseudo-RNG but in the grand scheme of things it works just fine for a video game.

[TomTom1968]

RNG doesn't mean if it's 25% chance that you will get 25% HQ by the end of a run. It means when you click on that node you have a 25% chance to get 1 hq. The next time you click, you have the exact same chance. RNG means that even in 1000 gathers, you could get 1 hq.

The same thing happens with a dice roll. It's completely random. Try it sometime.

That is how random works. Yes computers have pseudo-RNG but in the grand scheme of things it works just fine for a video game.

And dont't forget that the Human brain tends to forget about positive results more easily while negative results are remembered. I am also sometimes frustrated when I have a 50% HQ chance and don't get a single HQ item. But on the other hand, I sometimes have 3-4 HQ items although there is only a 15% chance. And everything below 100% can result in getting nothing even a few times in a row. Also, making 100 attempts and writing down the results doesn't lead to any useful insights because the next row of 100 attempts will surely reveal different results. That's what RNG really is. You might not like it but it is stochastics working as it's meant to be.

[Rivienne]

RNG doesn't mean if it's 25% chance that you will get 25% HQ by the end of a run. It means when you click on that node you have a 25% chance to get 1 hq. The next time you click, you have the exact same chance. RNG means that even in 1000 gathers, you could get 1 hq.

The same thing happens with a dice roll. It's completely random. Try it sometime.

That is how random works. Yes computers have pseudo-RNG but in the grand scheme of things it works just fine for a video game.

Distribution patterns. The farther away from that percentage you get the less likely it becomes that your results over a set of trials will deviate significantly from the probability of success.

So yeah, I do not think you understood me.. Distribution patterns and chances of success. I was using as an example the fact that the stated chances of success are flawed. The percentages listed do not accurately appear to reflect the distribution patterns, implying the distribution patterns themselves are flawed. Are you familiar with Bernoulli and Binomial Probability?

Chances of receiving my results of 38 (or less) over 300 tests at 0.25 probability is something around 1e-9. I could do the same test again, and I may when I have that much time, but based on my experience thus far every single time I mine, I anticipate similar results. These skewed results are normal. Not exceptions. And the fact I did 300 in a row to receive my tests should be at least somewhat representative of that. Obviously not definitive, but it does make things suspect since it does reinforce existing perceptions.