Refining probabilities

magilla666

I've written about this before, and the situation seemed to improve for a short time. Now it seems to have not just gone back to the bad old days, but deteriorated beyond expectation.

Do NOT try to tell me I have a 30% (or whatever) chance of success at refining a runestone when I don't, please.
I just had 18 consecutive failures, with supposedly that chance at success.

Simple probability calculations tell us the odds of that happening are 1/(0.7^18) = 0.0016

You would expect that to happen sometimes, I know. Now combine that with the other recent attempts at refining: 8 failures, 11 failures, 9 failures (with either listed 30% or 40% success chances). It stretches credulity past the breaking point to think those could happen with such regularity.

(By the way, the same is true of collecting skill nodes- don't tell me I have a 75% success chance when I get 7 failures in a row.)

Now, I don't care if the odds aren't 30%, or 20%, or whatever.
I do care when I'm presented with obvious bullshyte. That's what this is, and it's really frustrating- and cheap and cheesy on your part.

looomis

Your sample size is too small and you are subjective.
I tell you, the success rates are correct.

This discussion exists in every game and everyone is arguing with 5th class statistics. Events with low probabilities happen very often because a lot of events happen!

adinosii

I did some extensive testing on the preview server a year and a half ago. I cannot find my post about it now, but the bottom line was that the average was just as expeced but there was an indication the RNG was "streaky" ... that is, you would get a higher number of consequtive successes or failueres than expected.

greywynd

First attempt: 30% chance of success, 70% chance of failure
Second attempt: 30% chance of success, 70% chance of failure
Third attempt: 30% chance of success, 70% chance of failure
Fourth attempt: ...
And so on.

oliboyph

adinosii said:

I did some extensive testing on the preview server a year and a half ago. I cannot find my post about it now, but the bottom line was that the average was just as expeced but there was an indication the RNG was "streaky" ... that is, you would get a higher number of consequtive successes or failueres than expected.

This exactly. It is all correct and the same problems exists in every game because there is no such thing as absolutely random in computers. It is called simulated randomness (https://en.wikipedia.org/wiki/Pseudorandomness) and it can result in "streaky" outcomes.