Printable View
Wondered if anyone having a problem with finding fast copper mines.
My Jolly Geologist supposedly finds fast mines 45% of time... it seemed to me that I rarely had more than 1 out of 3 or 4 or 2 out of 6. Been thinking for a couple f months, this week (being sad) I've recorded of 35 searches 13 so far are fast (in same period with 45% fast coal 5 of 11 are). I shall keep recording wondered if anyone else thought there might be a bug.
Your sample is too small to conclude anything. Statistically it's a simple binomial distribution.
Let X be a random variable denoting the number of fast coppers in n number of searches:
Probability of success on a single trial: 0,45
Number of trials (n): 35
Number of successes (x): 13
P(X<=13)=0.2232
Which means there's a 22.32% chance of experiencing 13 fast coppers or less in 35 searches. All in all... 1 in 4 players would experience what you did or something worse, which makes it not all that uncommon.
This topic has been brought up countless of times btw. Let's say for the sake of argument, that the chance of fast copper was indeed less than 45%. What would that mean?
1) Some programmer botched writing 0,45 in the code, when writing the formula.
2) BB are lying to you about the chance of finding fast copper
How likely does either of those sound to you? ;)
Why are you people assuming that the 45% chance means that you have a 45% chance of finding a fast mine. What it means is that the whole server has a 45% chance. Ask yourself who has the best chance of getting the fast mines the player building level 1 mines or the player building level 5 mines who then has to dip into the pot more often.
The Mods! And BB_Ravel!Quote:
Originally Posted by Dorotheus
unless you wait every time until all your mines deplete before finding new ones most of the time you will have less than 45% of fast mines. fast mines deplete faster (doh) and when searching for a new one theres again 55% chance that fast will be replaced by normal..
I don't think you have actually looked at the text and tried to remember your high school statistics. The description reads:Quote:
Originally Posted by Dorotheus
"Gives a 45% chance of finding copper that can be mined faster"
That means 45% chance of finding fast copper in any given search
There are no dependencies on how many you already have or how many the rest of the server has found. It's a completely independent statistical calculation. If it isn't then the text is misleading.
It's just the same as flipping a coin except it's 45/55 and not 50/50.
I don't need high school statistics, I'm from the dark ages of computer science. I know the cunning maths computers use to give you the illusion that your PC knows how to carry out basic maths like subtraction.
Any usage of the pseudo random number generator be it for a mine search or loot drop is server wide because it is the same pseudo random number generator for the whole server.
I'm not sure what you are getting at, but ofcourse the same generator is used for the entire server. Why would my geologist work with a different algorithm than yours? The point is that the algorithm has a 45% chance of spitting out fast copper in any given search. Nothing to do with my neighbour just having found two fast coppers or having lvl 1 or 5 mines or whatever you are talking about.Quote:
Originally Posted by Dorotheus
If it's somethig else you mean then I'd love to hear your explanation, and if you're in computer science it should be easy enough for you to put up the math as well.
Computers can't generate true random numbers. What they do is take a seed number apply a formula to it and then use the result or part of it as a new seed number.
Here's a simple one you can try with your calculator called Cos56. Pick a number from 0 to 99, this is your seed. Do COS seed on your calculator. The 5th and 6th digits of the result are your new seed. Keep repeating the process for a series of seeds. Now if the seed is less than 45 you have a fast mine. 45 and above a slow mine. Now comes the important bit, have a friend say Geo every now and then. When they say Geo you look at the the seed number and say Fast or Slow.
On the server that seed number is being changed many times a second, using a formula more complex than the simple Cos56 above. Lets just say for the sake of argument 50 times a second. That's 4320000 times a day. If your sending your geo's out 6 times a day your taking a 1 in 720000 sample of the result, that's a very low sample rate to be using for statistical analysis.
Pointing the finger at pseudo-random generators in this case is utter nonsense. Sure, pseudo-random means that you could find patterns in the numbers spit out by the computer, but the numbers are spread over such a large number of players who build copper mines at such different times that the chance of this pattern lining up to consistently advantage certain players over others is astronomically improbable.Quote:
Originally Posted by Dorotheus
This
Makes me think that perhaps you do need some high school statistics after all. Because the fault in the logic is so easy to spot that we'll leave it as a trivial exercise to the reader.Quote:
Why are you people assuming that the 45% chance means that you have a 45% chance of finding a fast mine. What it means is that the whole server has a 45% chance. Ask yourself who has the best chance of getting the fast mines the player building level 1 mines or the player building level 5 mines who then has to dip into the pot more often.
To the OP: Cephalad has the right of it. What you've experienced is unfortunate, but actually more likely to happen than rolling the same number on a 6-sided die twice in a row.
Thanks for your replies Cephalad (and EctoRune), I have a rudimentary understanding of Statitsics having a Diploma in mathematics, seemingly others do not:) My post was out of interest as something I felt I had observed empirically but had no sound evidence for, and wondered if anyone else had looked into it substantively (to save me the effort). Anyway I continue to collect the data!
For conspiracy theorists, or anyone interested...
My sample is still too small... but now 71 searches...
now 26 out of 71 which as with the original post is still around 37%... (although last week it was a record 10/23=43% ... acceptable :)) again for the fewer coal searches it is 12 out of 27 (44%, was 45%).
Looking forward to a run of 8 fast to bring it to an even keel.
on the karma front my 2.5 (I added one midweek) explorers recently returned 3 nords in 8 days, and ages ago my first ever sftr yielded a white castle. I guess my skiving improved Jolly Geologist is just payback.
The term "you have X% chance for E(vent)" does not mean that X% of E is in between all thing what has happened. It says just E is among all thing what can happen, and its rate to all that thing is X%. In short, your 45% chance to find fast mines does not mean that 45% of mines found by you, must be fast. It means amongst the all of the mines you will find anytime in the near and far future, there will be 45% fast in sum. Because statistically not so big the chance for "very-very long" series by one element, the smaller samples will show something similar, but nowhere a parameter in above phrase, how big can be the differences... For example, when I got this skill, the first few days serches (perhaps 20-25?) found only 5-7 slow copper mines. Nowadays looks inversely, but I don't make such statistics.
I would think you would need a very large sample size to indicate whether its 45% or not. When I send mine out searching, I generally get a trend of 5-6 in a row of slow ones. Then I get 5-6 fast ones. I search and keep fast ones. Slow ones I destroy. In some cases I get more slow, but then it's usually followed by more fast ones.
At any rate, as far as I can tell it feels like its right, but I still think a large sample size is needed to actually determine if there is an issue or not.
^ This.Quote:
Originally Posted by Brankovics