This game has seriously messed up RNG

Quote:

---

Originally Posted by Fawkes

The main problem with RNG in this game is that it's not made with human players in mind. I think the numbers are made to work out to the correct percentage if you do like a billion attempts, which is how SE tests it. However, we as players usually do more in the 10-100 range, and it doesn't work out as well in those smaller sample sizes. So pretty much everyone who has ever gathered has probably failed at least 5 in a row at 90% chance or higher, and that feels super broken. So even though the math might work out to be close in the long run it's a terrible system for the players of this game, because it's frustrating to us to see 90% and then not get 9 out of 10, because it actually means 9 billion out of 10 billion and all our attempts were in the wrong billion.

---

Truth is, if you don't get occasional streaks of good or bad luck, then that is evidence of a bad RNG.

ITT: People who don't know how probability works.

Quote:

---

Originally Posted by MistakeNot

Truth is, if you don't get occasional streaks of good or bad luck, then that is evidence of a bad RNG.

---

Yes, I know. I want "bad" RNG. Actual random is not good for everything.

So you don't want actual random in your RANDOM number generator. K.

Quote:

---

Originally Posted by Stormfur

Reclaim fails a lot more often than 10% of the time.

---

A person probably not that good of crafter anyways if they having to use reclaim a lot, lets not blame that on rng.

Quote:

---

Originally Posted by Aramina

So you don't want actual random in your RANDOM number generator. K.

---

Not sure why you can't wrap your head around it. I don't want a system where one person can get their atma from the first fate and another person has to do 200 fates for the same atma. I don't want it to be arbitrarily unfair for the sake of being "good" RNG.

RNG has its places. It is appropriate for lotting on loot. It is not appropriate for crafting to the extent that SE employs it. IMO your success rate for something like Hasty Hand should be influenced by the craft level and your level and your Craftsmanship/Control levels. The higher your personal numbers the greater success you should have, consistently. My non specialist numbers are 1005/955. In infuriates me when I blow 3 straight Hasty Hands on a level 20 craft. BTW if you notice RNG is not completely random but the pattern of success to failure varies at a noticeable rate.

Didn't read nine-page thread full of folks who don't know what random means, and folks trying to explain it to them. I'm going to toss in my two cents anyway.

The RNG in this game is working fine. It may seem like it's full of bad luck with all the bad luck posts on the forums - but that's because people rarely post their good luck, and NEVER their normal luck. In addition, you are much more likely to remember your bad luck streaks than your good luck streaks, so it can seem like that's all you get. That is perception bias.

I'm not against SE adjusting RNG with a "mercy" rule, stating that if you get a bad enough run of bad luck you get a guaranteed success. Some RNG-dependent games work that way. But I'm also perfectly fine with the "true" RNG system already in place. If you understand how probability works, than you start to realize that "nothing but bad luck" is significantly less likely to happen to you than being struck by a meteor.

Shortly after ARB's launch I collected the results of my mining grind (several hundred) and did some analysis to see if the results matched the stated probabilities for high quality finds. I got a 95% confidence level that the results DID NOT match (which is unfortunately not really that high).

With only 95% confidence I can't say too much, but based on that an my observations I think something is up with those numbers. I doubt their random number generator is at fault. If I had to guess I would say there are factors that affect success that are not represented in those numbers. Maybe some kind of 'biorhythm' affect or weather...

The fact that the game uses so much RNG to stretch out content is another topic altogether.

Quote:

---

Originally Posted by Fawkes

Not sure why you can't wrap your head around it. I don't want a system where one person can get their atma from the first fate and another person has to do 200 fates for the same atma. I don't want it to be arbitrarily unfair for the sake of being "good" RNG.

---

I understand fine, however, it seems you don't. Random is neither good, nor bad. It just... is. That's how math works.

edit: If you want a system other than RNG that is fine, but this talk of good RNG and bad RNG needs to stop :P

Quote:

---

Originally Posted by Rizon

RNG has its places. It is appropriate for lotting on loot. It is not appropriate for crafting to the extent that SE employs it. IMO your success rate for something like Hasty Hand should be influenced by the craft level and your level and your Craftsmanship/Control levels. The higher your personal numbers the greater success you should have, consistently. My non specialist numbers are 1005/955. In infuriates me when I blow 3 straight Hasty Hands on a level 20 craft. BTW if you notice RNG is not completely random but the pattern of success to failure varies at a noticeable rate.

---

You gear does influence your crafting ability, there are more rotation out there then just ones that use hasty touch, your craftmanship, control and cp all determine what rotation you can use.

Quote:

---

Originally Posted by Fawkes

I don't want a system where one person can get their atma from the first fate and another person has to do 200 fates for the same atma.

---

But you could be that first fate person! Maybe! :D

Quote:

---

Originally Posted by Aramina

I understand fine, however, it seems you don't. Random is neither good, nor bad. It just... is. That's how math works.

edit: If you want a system other than RNG that is fine, but this talk of good RNG and bad RNG needs to stop :P

---

No, you do not understand. Computers cannot be completely random, so "good" RNG and "bad" RNG is referring to how closely they resemble actual randomness. When I say I want "bad" RNG I'm not not asking for RNG that stays out late at night and doesn't eat its vegetables. I want the game to take failures into account. I want the displayed percentages to better match our expectations, rather than telling people they just have confirmation bias and RNG is RNG.

I used it a LOT when getting the Lumina stuff in 2.0 because the battlecraft or fieldcraft or whatever it was, was prohibitively expensive to just lose if RNG didn't favor you that run.

Quote:

---

Originally Posted by Bluejay

I doubt their random number generator is at fault. If I had to guess I would say there are factors that affect success that are not represented in those numbers. Maybe some kind of 'biorhythm' affect or weather...

---

No. Please, God no. Do NOT bring that into this game. We don't want superstitions like FFXI's infamous "crafting directions" to start to spread. (A superstition which, by the way, SE finally, FINALLY debunked, in spite of FFXI crafters swearing by it for more than a decade.) Please learn from our mistakes in FFXI. There, it was forgivable as the game DID hide a ton of mechanics from the player so it wasn't totally unreasonable to suspect that there might be hidden factors to crafting. This game, however, lays everything out on the table. All stats are accounted for and clearly explained by tooltips. There are no "secret tricks" that a crafter/gatherer can stumble upon that will improve their results.

Someone brought up the topic that, as no computer can truly be random, the game could be using a crappy random number generator that is prone to "bad streaks". If such a crappy generator exists (and if it does, I'd think it'd be equally prone to "good streaks"), why on earth would SE use it? There are plenty of random number generators out there that produce results which, as far as any human will be able to tell, are indistinguishable from pure randomness. There's no reason SE would choose a crappy one - not even price is a factor, as these random number generators are not expensive.

There's no hidden conspiracy, no secret tricks. 80% is 80%. 20% is 20%. To believe anything else is either paranoia or foolishness.

You tend to remember all the times where Hasty Touch failed 5 times in a row, and not all the ones where it succeeded every time. It's the same on AST, when people were arguing that Balance, Arrow, and Bole had a lower chance to draw than Spear, Ewer, or Spire. you don't remember the runs where all you did was draw the good cards.

My 4 star rotation gives me, with the worst luck, 7 chances to use a touch action. Usually I end up with 9 or 8. One of those can be a guaranteed Basic or Precise off the bat, with an extra for each ToT proc I get. So, yes, there's been plenty of times where Hasty Touch failed over and over and over and I had to Reclaim or did so poorly that it NQed anyways. But there's also been times where every single Hasty succeeded, and I'd say that those times happen about as often as the times where it fails a lot. Most of the time, I will only fail 1 or 2 Hasties per craft. 80% is 80%. You are just ignoring the times where the RNG was amazing.

I think the bad reputation associated with RNG in crafting stems from players remembering incidents of bad RNG (6 or more misses in a synth) and more likely than not, they're probably mostly being tripped up by synths in which there are 3-4 misses. Those are normal occurrences and a crafter should be able to handle them.

Let's take a look at an 11 touch rotation. In this case, without considering RNG reducers (such as BT swaps) you can expect to see 4 or more misses around 16% of the time and 3 or more misses close to 40% of the time. With 11 touches, to recover to 11 stacks of IQ from 4 misses, you'd need to use precise touch a whopping 3 times in the synth (and you'd still wouldn't get 100% from all NQ on a 4* with or without innovation). Simply put, a crafting strategy that consists of 11 touches is not robust enough. But since you do see 3-5 misses regularly, you should be developing strategies to deal with them.

The crafting system is based on probabilities. Conditions vary so your methods should as well. The difficulty level in HW crafts is very very low, so I really don't see why there should be many issues. Start with 2000/13187 quality and 6 misses should not be a problem. Start with 4000 quality and you won't even need any of the Byregot's abilities on the majority of your 4* synths.

The fact that your gathering rate on Hidden Nodes is capped at 95% and you STILL miss 1 or 2 or even 3 hit sin a row at 95% is completely unacceptable!!

Quote:

---

Originally Posted by Fawkes

Not sure why you can't wrap your head around it. I don't want a system where one person can get their atma from the first fate and another person has to do 200 fates for the same atma. I don't want it to be arbitrarily unfair for the sake of being "good" RNG.

---

That is not entirely unreasonable. What you want is something that isn't truely random, but more like "random-appearing, but not TOO random" that does take into account history so that long streaks of bad luck (or good luck) don't show up.
One MAJOR problem with implementing that is that you would need to track the state for every combination of <character> and <rng-based task>. That will be a LOT of data to keep track of, unlike today where the RNG just needs a single, global, state to do its work.

Quote:

---

Originally Posted by Decoy_Wolff

The fact that your gathering rate on Hidden Nodes is capped at 95% and you STILL miss 1 or 2 or even 3 hit sin a row at 95% is completely unacceptable!!

---

Thats not how RNG works, let me out it into perspective:

95% means that if you hit 100 times you should get 5 misses, but that also means that if you hit 1000 times you should get 50 misses, and if you hit 1,000,000,000 times you should get 50,000,000 misses, the game is taking into account an infinite number of hits, so your chance of getting a miss is "infinite / 20" which is far from 0, and the chances of you getting some of those hits in a row isn't out of the question.

Quote:

---

Originally Posted by Decoy_Wolff

The fact that your gathering rate on Hidden Nodes is capped at 95% and you STILL miss 1 or 2 or even 3 hit sin a row at 95% is completely unacceptable!!

---

Yeah, 3 misses at a 95% success rate is only a 1 in 8000 event. How many people across all the servers are making similar attempts in any given day? Its bound to happen to someone, it just sucks when its you.

Quote:

---

Originally Posted by Decoy_Wolff

The fact that your gathering rate on Hidden Nodes is capped at 95% and you STILL miss 1 or 2 or even 3 hit sin a row at 95% is completely unacceptable!!

---

Missing 3 in a row at 95% is uncommon (I've rarely seen it if ever; maybe once), but if you're swinging 5 times at 95%, you'll see single misses regularly.

Statistically, there is close to a 25% chance that you will see 1 or more misses out of 5 swings. So on average, you'd expect 1 out of 4 of your hidden nodes to have one or more missed swings.

This seems to be in line with my experiences when I did gather before. However, most of the times I gathered at 100% so my own sample size is limited.

I used to think these sort of threads were simply people not understanding how random numbers can create streaks or simply people remembering failures more easily than successes. But having recently made a new alt and going through the early levels of crafting several times, I'm now not quite so sure.

It seems like Basic Touch has a far greater success rate once you outlevel the recipe you're making than it does before. Five levels above the recipe and it'll only occasionally miss (about as you'd expect). Yet a level below the recipe, it fails more often than it succeeds. The tooltip doesn't acknowledge any such dependency on level. It claims you get a simple 70% success rate, but the change was pretty consistent and predictable. What's more, it occurred only with the Basic Touch, and not with Basic Synthesis, which actually does seem to maintain its stated 90% success rate throughout.

So while the game seems capable of a good random number generator, I'm no longer convinced that all the tooltips and other windows that specify a success rate are really telling the whole story. It seems more like "Success Rate: 70%" can in some cases mean that if you meet other (unstated) requirements, then it will use a 70% chance RNG.

(And after all, it's not like all the game's other tooltips and explanations are kept accurate. I'm pretty sure the Active Help description of the armory bonus still claims it depends on how far apart your levels are, which has never been the case, and people have been pointing out that discrepancy ever since the game started.)

Stat Trek's Random Number Generator is the closest thing you'll find to a TRUE RNG, what you see most of the time (especially in games) are called "pseudo random number generators" and, to put it simply, they work like this. Without looking at the faces, try to pull a "5" from a deck of 52 cards, and then try to pull the same "5"card from the same deck with all the court cards (face cards) removed and only the 40 pip (number) cards remaining. You're still pulling randomly for a specific outcome (the "5" card), but the ODDS have been manipulated by removing 12 of the cards. This is called a "manipulated variable" where you affect the RNG by changing the variables it works with. The game is just "fudging" the odds. Cuss it out and move on.

I once missed six 97% chance gathers on the same node. I really should have bought a lottery ticket lol

Quote:

---

Originally Posted by madolyn

i've seen this and from my personal experience since i moved from NA to JP server, I can't win a loot roll to save my life. constant roll losses when rolling 95+. If i don't roll a 99, i can just forget about it.

---

Did you set your language to JP (apparently it helps :p)?

Joking aside, I've also noticed a lot of streaks in the RNG system, but it does seem to be balanced over time (i.e. for each 99% failure, I have gotten a 1% success at some point, though usually not when I needed it). One positive side effect of these streaks, at least from the small experiment I have done so far, is that it may be easier to get melds on while you're on a lucky streak (e.g. after getting multiple HQ garthers in a row)... it's not perfect of course, but from the dozen or so 10-16% melds I have done so far while on a 'lucky streak', I am greatly exceeding the odds, and my expectations from previous melding experience.

no matter which way you want to play it, the fact stands. when you are failing 2/4 or 3/6 and so on with a 90% chance of success showing that is wrong. Period. If you are going to show a 90% chance of success I expect a 90% chance of success you either show it right or fix the system. there is no way around it.

You dont ignore the chance of success wtf. and you have to BOLD that. WTF. YOU ARE JUST SAYING BECAUSE YOU WANT PPL TO SHUT UP. THAT IS WRONG.

There is a larger randomness/variance to it so that even a very high chance of success still offer ample chance for failure. The system and the explanation/description does not add up. There is ample prove. STOP ARGUING ITS ACTUALLY CORRECT. IT IS WRONG IT IS WRONG IT IS WRONG. GET THAT INTO YOUR HEAD. YOU ARE NOT ALLOWED TO ARGUE. THERE IS NO DEFENCE. YOU DO NOT DISPUTE FACT.

Quote:

---

Originally Posted by hamond

There is a larger randomness/variance to it so that even a very high chance of success still offer ample chance for failure. The system and the explanation/description does not add up. There is ample prove. STOP ARGUING ITS ACTUALLY CORRECT. IT IS WRONG IT IS WRONG IT IS WRONG. GET THAT INTO YOUR HEAD. YOU ARE NOT ALLOWED TO ARGUE. THERE IS NO DEFENCE. YOU DO NOT DISPUTE FACT.

---

Actually, you are wrong in this case. This is RNG, not percentage based success. The 90% you see is the chance you have a in a single instance. otherwise, every 90 hits you would miss 10 in a row. Is this what you want? You want to miss 10 hits guaranteed every 90 hits? I sure don't. Personally, I don't miss much at above 90% and often see HQ/gathering successes at the 40-60% area. If you don't like RNG, I suggest you find something besides an MMO to play.

Quote:

---

Originally Posted by hamond

no matter which way you want to play it, the fact stands. when you are failing 2/4 or 3/6 and so on with a 90% chance of success showing that is wrong. Period. If you are going to show a 90% chance of success I expect a 90% chance of success you either show it right or fix the system. there is no way around it.

---

This is a very common logical fallacy that often shows up in gambling, called the Monte Carlo or gambler's fallacy. A good way to sum it up is this - roulette players who notice that it keeps landing on black will bet on red because "it's a 50/50 chance, red HAS to show up soon." But it's only a 50/50 chance for each individual spin, not overall, so it causes people to lose money. It's actually a pretty interesting subject: https://en.wikipedia.org/wiki/Gambler's_fallacy

Same logic applies here. Each hasty touch or whatever you take is a 90% chance, with a 10% chance to fail. So yes, it IS correct that you can fail that many times on a 90% chance of success. Again, you are ignoring times when the RNG was wonderful and amazing and you finished with every single hasty touch succeeding. Instead of screeching at people for being wrong when they're correct, consider rotations or abilities you can use to smooth out the RNG even further.

Quote:

---

Originally Posted by GunksFoy

Actually, you are wrong in this case. This is RNG, not percentage based success. The 90% you see is the chance you have a in a single instance. otherwise, every 90 hits you would miss 10 in a row. Is this what you want? You want to miss 10 hits guaranteed every 90 hits? I sure don't. Personally, I don't miss much at above 90% and often see HQ/gathering successes at the 40-60% area. If you don't like RNG, I suggest you find something besides an MMO to play.

---

The problem is not that its failing in a single instance. The problem is that its failing in alot more instances than it should. If you are suggesting a 90% chance it should be alot closer. You should not be failing this much. 90% is only an example, you do hell lot worse if you are under that percentage. The whole system is severely underperforming.

Quote:

---

Originally Posted by GunksFoy

Actually, you are wrong in this case. This is RNG, not percentage based success. The 90% you see is the chance you have a in a single instance. otherwise, every 90 hits you would miss 10 in a row. Is this what you want? You want to miss 10 hits guaranteed every 90 hits? I sure don't. Personally, I don't miss much at above 90% and often see HQ/gathering successes at the 40-60% area. If you don't like RNG, I suggest you find something besides an MMO to play.

---

If you get 10 misses for every 90 hits that is the perfect result based on theory. You guys are pulling the theory argument and now you ask what I want as the preferable result instead?

Stop playing around make up your mind. You are arguing the theory or arguing what people want as the result. I am been objective. Lay down all the facts then you can continue the debate. At the moment both sides are biased.

Elamys

I can tell you it doesnt happen. Unless I 100% it, I fail at least 50% of the time. So much that its not worth it for me to try the touch skill without steady hands. In gathering, the result and the displayed percentage on obtaining an item is alot worse. However, the percentage on whether HQ items is correct or not seems to be more inline. Occasionally you can severely over perform over a relatively small number of tries, say 30-40. In the case of HQ items its actually consistently over performing. That is result based.

For the theory, it is correct to bet on red if it lands on black. You know as well as anyone that it will not always land on black, otherwise the perceived percentage would be almost 90-100%. People are not stupid. While it is not reliable to land on red, it is definitely not false. Especially if you consider the chance to be 50% in which instance it doesnt matter which you bet, hence you cannot be false at all. You'll be right whatever you choose if the chance system is perfect. Secondly, probability changes when you are looking at things from different perspective.

We are not looking at an isolated chance of success here. If we are yes, it can fail. The problem here is not that yes it can fail, the problem is that its failing way more than it should over multiple tries. This more like conditional probability. Considering a single instance in isolation is completely wrong. Giving your reasoning a name doesnt make it right, particularly when you are applying it wrong.

And yet, when I look at my synths overall, when other crafters in this thread have looked at their synths overall, it is a 90% chance of success. Like I don't know what your issue is. I'm sorry you failed to HQ a craft?

This isn't even touching on the complexity of getting true RNG into a game, where "clumping" can happen. What do you want to do? What options are you proposing to solve this not issue, because it is in fact working as intended?

How often are you actually synthing and gathering to have this issue? Because I can guarantee you, it's not nearly as bad as you're saying, and if it is, you are doing something wrong, plain and simple.

By all accounts, the RNG method that the game uses is one of the highest quality ones available if your goal is true randomness(several people have argued for weighted probability - failing making you more likely to succeed in the future - but thats a separate argument and a core change to the way the game works that is highly unlikely to happen). You said that considering a single instance is wrong, but thats literally all the game looks at. When you press the button for synth(or whatever) it randomly determines success/failure based on the percentages. It doesnt consider whether you failed or not the last 5 times you tried, it only looks at the current attempt. I leveled all 8 crafters to 60 without help, as well as all 3 gatherers. Ive hand crafted roughly 10000-12000 items in that time, and gathered two or three times that many. Ive had 98% successes fail, ive had 1% successes work. I will tell you that in my experience the system is working as intended, as have several other people. This is really more of a psychological issue for many people - they remember their failures much more readily then their victories. The reason Elamys mentioned the gamblers fallacy(and linked to a description) is because it is directly related to this phenomenon. Many casinos put up a board of the most recent 10-20 spins on a roulette wheel because they know people will look at it and think that they can tell what will happen next based on what happened last. This causes them to bet more and lose more - the casinos goal. The wheel doesnt care if the last 200 spins were all black(as unlikely as that would be), the next one is just as likely to be black again as it is to be red - but people in general ignore or dont understand that.

Try the following: spend a couple of hours running /random and recording and plotting the results, I guarantee you will not see a pattern and that distribution is fairly even

Why aren't people using reclaim if they're failing that many hasty touches?

They're probably scared by the fact that Reclaim still has a chance to fail, but I've failed reclaim so rarely that, yes, I definitely think 90% is still 90%.

reclaim was when I stopped taking crafting seriously. It failed every third time.