Understanding Probability and Sample Sizes

For anyone having trouble understanding the Random Number Generator in the crafting system, please take 8 minutes of your life to watch this very basic video about the concept of probability.



Note that the "sample size" of your probability experiments need to be "extremely large" to find the true probability of an event occurring. In other words, you can't just take a handful (400 for example) of events and suggest that this will show you the probability of the RNG is "broken".

It doesn't work that way. Especially when each event is mutually exclusive. Meaning that it doesn't matter AT ALL what the last RNG number was, because if you have for example an 83% chance to achieve success, you could still potentially have hundreds or thousands of failed events in a row. Because that's how probability works.

If you want a GREAT quick and easy primer on the subject, please listen to this:

EDIT: Here, please listen to this episode of Radio Lab at the 8 minute 25 mark.
http://www.radiolab.org/story/91684-stochasticity/


Thank you for your time.

 

I don't need a very large sample size to know that the existing system is broken. i.e. It's not fun.

 

It's perfectly acceptable to say...

It's not fun.
You don't like it.
You want it to change.

It's not acceptable to say the RNG isn't working correctly by using misleading or unmeaningful statistics in an effort to create the change you want. People are entitled to their own opinions, not their own facts.

For example, I like the RNG system and I actually think that's way too easy to create good gear. I think the percentages are too far in the favor of the player.

 

I don't mind for gear to be hard to make, but the high percentage showing on the UI right now is just poor user experience.

"Hey.. this item gives you a 83% successful chance..oh wait, there is one more thing. It's only 83% after 100,000 attempts by the way."

 

You still don't understand. It's an 83% chance. It doesn't mean you're not going to fail. If you do fail, it doesn't mean the next time won't be a failure.

It means that if you do something 1 million times, the average success rate will be about 83%. Which also means that there will be pockets of time in that 1 million sample size where you fail over and over again.

 

It doesn't matter what's the logic behind the number. From user experience perspective, I expect to see the success chance specifically to the current instance of what I am trying to craft.

If there is a situation where people need to make decision based on probability of success chance. It's most likely won't be based from 1 millions theoretical instances. For example, let's say a group of leaders try to weight on the decision on the success chance of running certain operation against the enemies. When their advisors said there is 80% chance of successful strike.. They can't mean that out of all 1 millions time, you will win 80% of the time because there is only one shot and no second chance. Just like enhancement and masterwork. You only get one shot, your item will be gone if you fail.

 

Except that the pseudo-random number generator that is being used could be broken or have some odd quirks and that would be a completely legitimate query. Given how you prefer people to stick to the facts, I assume you know all the details regarding SotA's number generation, right? =P

 

The way I see it, 80 % success still leaves a significant chance of failure. If you can't afford the negative outcome with those odds, then don't roll the dice. People might have unrealistic expectations when it comes to what they should be able to craft. Always prepare for the worst when you make your decisions.

Personally, I would like if armor broke down faster and there were no COTO:s. I like the idea of gear as something temporary and disposable, rather than aiming for some kind of end game gear that you will use forever.

 

I am pretty sure it is still near impossible to achieve true random number generation anyway. I think the thing that bothers people is that so much relies purely on the RNG, especially crafting and combat. The range of damage barely tightens with combat experience meaning that no matter how skilled you become with that sword, you are still going to hit things for 2 damage even if it is a stunned target well inside your ability to defeat. Even grandmaster craftsmen are completely destroying dozens of attempts in a row at making the simplest of objects... i mean, how can anybody claim to be a grandmaster when they turn a two ton pile of resources into a handful of scrap in an attempt to make a dagger? At worst there should have been some sub-par daggers forged, how hard is it to turn an ingot into something with a blunt end and a pointy end? A drooling idiot in prison can do better with a toothbrush and a lighter.

 

Not really related to the topic but I recall reading an article ages ago about using cosmic rays to produce truly random numbers. It's pretty interesting and worth looking up if you like that sort of thing.

 

I didn't say the RNG couldn't be broken. I said that no one using meaningless statistics will be able to prove it so stop using meaningless statistics.

I don't know that the RNG is working as designed, I can't know that. But you don't see me making threads with a sample size of 12 or 400 trying to defend it because I understand how probability works and I don't just make stuff up because I don't like a system.

It's fine if people want to say that they don't like the RNG system, great wonderful. Just stop acting like it's broken because you got a few failures in a row. That's not how probability works.

 

The value, in my honest opinion, should not be displayed in the first place, as without an appropriate mathematical background (education) it can easily be (and keeps being) misinterpreted.
The number does not and can not say 'please expect this number of successes in ten attempts'. It more says: doing a set of attempts expect less attempts to achieve a successful one when the number is higher.
The random number generator generates (unless changed since) as expected according my own measuring and it is not broken.
As pointed out here, there would be a good reason to point at using a randomness in a process of crafting, based on subjective feelings, as something that does not feel right, and might be replaced by another system. I could agree with that. But it makes absolutely no sense to me keep putting on table 'a broken RNG' - if I may say it in a politically correct way.

 

People thought critical hit chance was due to issues with RNG, but it turned out to be an error in the formula I believe.

People also blamed RNG for an issue with failing when it was displayed as a 100% chance, which is apparently now fixed. I assume this problem was an error with 100% being displayed as a rounded up number, which was actually >99.5% but <100% (I can't confirm this it's just my assumption)

Good on you @Baron Drocis Fondorlatos for pointing out the truth about sample sizes though. I doubt anyone has had the resources needed to accurately test anything probability-wise yet.

 

Except that probability should not have much if anything to do with the fun. And if people are going around sputtering numbers based on a low sample size and it's incorrect, then it's just Chicken Little.

This just kind of irks me because you get these self-declared experts. But you cannot be an expert unless you are involved in the coding/algorithms/spreadsheets/databases or whatever over at Portalarium. And let's remember they acknowledged having an incorrect baseline number that affected calculations for other numbers. Doh!

Maybe people are thinking that when the number was borked things were fine (sure, give me more gold and xp and skills and respawn rates), but now the number has been corrected and the friggin sky is falling.

 

The issue with critical rate was to high was due to the equation did not properly express the crit chance design feature. Once it was correct, it became too low for most of the people who relied on crits to deal damage. However, that is a game balancing issue.

For crafting, while the success may be accurate based on the concept of statistics. However, what good does it really do to have this number to be shown at this point? Consider that nobody gonna craft 1 millions items, so that they get can successfully create 800k items. People want to craft the item at that instance. Perhaps they may have spare resources to create up to 10, 100, or 1000? If the % success chance does not represent the actual chance they have within the scope of the resources, then what good does it do to be shown on the screen since it is not reliable according to the resources people can realistically use for crafting the items they want.

Whether items should be easy to craft or not, that is out of what I want to point out for the moment. The issue here is that we do have a number display on the UI that makes many people upset. Should that really be ignored?

 

Maybe the number is there to alert people to the real possibility of failure, and so they're not too upset if it happens.

Next time you visit a casino, look at all the poor schmucks pulling the handles of the pokies, and hoping desperately for a win so they can pay their rent or at least eat that week. On some level they know the odds are against them (the house always wins overall) but there is always that chance of a big win.

There's something about human nature - if there's a random chance that the next event will be infinitely rewarding people tend to keep going for a long time.

FWIW, just be mindful that crafting can fail, and be mentally and emotionally prepared.

As has been elegantly and correctly stated, in the long term (1,000,000 or so events) the casino will always win, and your critical crafting success WILL be x% as indicated. However you can have a lucky streak along the way, or an unlucky streak.

 

I didn't say anything about numbers. Crafting itself is not really fun and is downright depressing in the amount of materials required and items destroyed, whether or not it's 'working as intended.'

 

So what you're really saying is, that you never had any intention of staying on topic, and only came to troll the OP and derail the thread?

 

No, what I came in to say was that complaining about the RNG or complaining about the people complaining about the RNG isn't really the issue.

 

You're right. I was trying to make some rope, so I could make a small bag, so I can name the bag [Kobold Kush]. What a chore just to do that. It's not being able to do simple things like this in crafting that dissuade me from even leveling it up.