Philosophy Forums



Register | Forgot Password

Random or Not?
How do you decide random from non random?

PrintPrint


Page: 1 2

Random or Not?
123savethewhales
Assistant Professor
Avatar

Usergroup: Members
Joined: Mar 25, 2009

Total Topics: 6
Total Posts: 278
Posted 08/05/09 - 02:02 AM:
Subject: Random or Not?
quote post
#1
I was reading a gaming forum and I came across an argument like this

Person A: The item drop in this game is random

Person B: But I got 2 (insert item name here) in under 5 minute, then 10 minutes later......

Person A: That is just pure statistic, I have run thousands of test myself

Person B: No the system is rigged

Now in a game, we can always open up the source code and reverse engineer the program to find out how item drops are calculated every step of the way. Assuming that we don't open up the source code (since in most real life situations there are no source codes for you to open up), there seem to be no logical way of solving this problem. Yet most of us make these distinctions pretty much on a daily basis.

So take a real life example, say quantum mechanics, how do you distinguish rather or not it is a random vs non random phenomenon? Do you consider such a decision a rational one?

Keep it simple.
Aceedwin
Default Rank Rejecter
Avatar

Usergroup: Members
Joined: Jul 22, 2009
Location: The universe. Possibly.

Total Topics: 2
Total Posts: 108
Posted 08/05/09 - 02:11 AM:
quote post
#2
Hmmm, a nice little thought. Person B sounds a little paranoid but has a good point. There is no way to check whether a system is random without dissecting it. Perhaps, if we learned more about quantum mechanics then we'll find a link and quantum mechanics won't be random, but just another law, like gravity, albeit somehow less predictable.

Some people enjoy finding answers, but I dislike losing questions.
swstephe
Tenured Poster
Avatar

Usergroup: Moderators
Joined: Apr 20, 2006
Location: borneo island

Total Topics: 28
Total Posts: 3125
1 of 1 people found this post helpful
Posted 08/05/09 - 08:37 AM:
quote post
#3
123savethewhales wrote:
Now in a game, we can always open up the source code and reverse engineer the program to find out how item drops are calculated every step of the way. Assuming that we don't open up the source code (since in most real life situations there are no source codes for you to open up), there seem to be no logical way of solving this problem. Yet most of us make these distinctions pretty much on a daily basis.


This is actually an interesting problem in computer science. When people say "random", (like the person in that forum), they don't really mean "random", they mean "uniformly distributed". That means, over time, they should see each number produced a roughly equal number of times and with unpredictable regularity. There is no such thing as "random" in computer software because everything is deterministic inside the CPU. The majority of games and other programs would simply call their programming language's runtime library and access the "random" function. This function would produce numbers in a uniform distribution with a very difficult to predict regularity. If you started from the same point, you would get the same sequence of numbers. The initial point is called the "seed". If you wanted to make the program start at a random point, you usually access some function that is external to the CPU. Most programs I see will get some digits from the internal clock, meaning that the seed changes on the microsecond, (a thousand times a second). I've written programs which take the time between keystrokes as an additional modifier, meaning that two people hitting the "start" button at exactly the same time, (according to two system clocks), would still have different result, (I admit it was overkill).

You can test for distribution, (that's how they write the random function in the first place). If you run it a billion times, put the results in a histogram, it should be nearly a flat line, (uniform distribution). That satisfies most requirements for "random". As a mathematical function, "random' is really undefined. There was a Dilbert joke which showed a demon whose task was to produce random numbers. He sat there and said "3...3...3...3". Even if you had a *lot* of time, you couldn't say whether you were just incredibly lucky and happened to catch him during a particularly long sequence of 3's, or whether he was cheating and just returning 3. I've also seen a programmer joke which showed a function called "returnRandomNumber" whose code was "return 3".

There is also causal determinism. Something is "random" if it has no cause, meaning you can't predict the output of the function based on the input. Quantum indeterminancy simply says that you can determine velocity or position, but not both, meaning that its state is unknown until you measure it.

Ethics is the measuring of morality. Morality is the measuring of good. Good is the measuring of benefit. Benefit is the measure of values.
123savethewhales
Assistant Professor
Avatar

Usergroup: Members
Joined: Mar 25, 2009

Total Topics: 6
Total Posts: 278
Posted 08/05/09 - 11:05 AM:
quote post
#4
swstephe wrote:


This is actually an interesting problem in computer science. When people say "random", (like the person in that forum), they don't really mean "random", they mean "uniformly distributed". That means, over time, they should see each number produced a roughly equal number of times and with unpredictable regularity. There is no such thing as "random" in computer software because everything is deterministic inside the CPU. The majority of games and other programs would simply call their programming language's runtime library and access the "random" function. This function would produce numbers in a uniform distribution with a very difficult to predict regularity. If you started from the same point, you would get the same sequence of numbers. The initial point is called the "seed". If you wanted to make the program start at a random point, you usually access some function that is external to the CPU. Most programs I see will get some digits from the internal clock, meaning that the seed changes on the microsecond, (a thousand times a second). I've written programs which take the time between keystrokes as an additional modifier, meaning that two people hitting the "start" button at exactly the same time, (according to two system clocks), would still have different result, (I admit it was overkill).

Thanks for the description, while I do know about seeds and pseudo random generators in computer, and that it isn't random in a the level of no physical cause, I really don't know they go as far as to make keystrokes a factor as well. On the practical level, as you "overkill" your program, would it not match the random, if not more so, that of a roll of dice, or a flip of coin? (assuming you don't know how the number is generated every step of the way)

You can test for distribution, (that's how they write the random function in the first place). If you run it a billion times, put the results in a histogram, it should be nearly a flat line, (uniform distribution). That satisfies most requirements for "random". As a mathematical function, "random' is really undefined. There was a Dilbert joke which showed a demon whose task was to produce random numbers. He sat there and said "3...3...3...3". Even if you had a *lot* of time, you couldn't say whether you were just incredibly lucky and happened to catch him during a particularly long sequence of 3's, or whether he was cheating and just returning 3. I've also seen a programmer joke which showed a function called "returnRandomNumber" whose code was "return 3".

This is what I am referring to, in the sense that you cannot tell rather the program calls on it's pseudo random generator, or "return 3". You can't prove anything per say with statistics alone, beyond saying that it is highly unlikely that all those 3s are randomly generated. But I also notice how people tends to overestimate the pattern they see, especially when their data isn't recorded in a database. It seems that if you give different people the same set of random number that falls very close to the middle of the bell curve, he/she might consider his/herself as lucky/unlucky depending on their personalities.

There is also causal determinism. Something is "random" if it has no cause, meaning you can't predict the output of the function based on the input. Quantum indeterminancy simply says that you can determine velocity or position, but not both, meaning that its state is unknown until you measure it.

While I did read about quantum superposition, and all the talk about "chance to generate a hex boson" in the LHC, I must admit I get lost a lot when people talk about the uncertainty principle and somehow relate that with true randomness.

Keep it simple.
Death Monkey
Tenured Poster
Avatar

Usergroup: Members
Joined: Sep 18, 2003
Location: Eindhoven, The Netherlands

Total Topics: 7
Total Posts: 2598
Posted 08/05/09 - 11:36 AM:
quote post
#5
Moving away from the specific question of computers and pseudo-random generators, and towards more general issue of deciding whether something is random or not. I look at it like this. When we are confronted with a system that may, or may not, be random, the only thing to do is to try to reject the null-hypothesis that it is random. If we cannot do so, then we must treat it as being random until such time as we can do so.

The basic idea here is two-fold: First, if we can't reject the null-hypothesis that it is random, then there is no justification for abandoning that simpler hypothesis for a more complicated one that posits some rule or cause for what the system is doing. Second, and more importantly, without information about how the system differs from a random one, we have nothing to base a more complex hypothesis on. At most, we could speculate that it's not random, but not actually propose a non-random model that is anything more than a blind guess.

Put simply, as long as we cannot reject the null-hypothesis that the system is random, there is no way that a non-random model will be able to more accurately describe the system than the random one.


DM

Pseudoscience makes Baby Jesus cry.
Cadrache
Tenured Poster

Usergroup: Members
Joined: Dec 09, 2006
Location: AB, Canada

Total Topics: 98
Total Posts: 2125
Posted 08/05/09 - 02:02 PM:
quote post
#6
I always like how picking names from a hat is not random in that each pick has a different chance of being chosen.

grin

"...There was a writer who asked why it was that when we find positive experiences we say that only the physical facts are real, but in negative experiences we believe that reality is subjective. He made an example of those who say that in birth only the pain is real, the joy a subjective point of view, but that in death it is the emotional loss that is the reality." - Tony Ballantyne, Recursion.
swstephe
Tenured Poster
Avatar

Usergroup: Moderators
Joined: Apr 20, 2006
Location: borneo island

Total Topics: 28
Total Posts: 3125
Posted 08/06/09 - 02:08 AM:
quote post
#7
Maybe, rather than insisting that randomness is unprovable, you would be better of just saying they are suffering from Gambler's fallacy. The fallacy that probabilities will necessarily even out over time, that certain numbers are "due". On average, it is unlikely that I could roll 100 consecutive "snake eyes" on a pair of 6-sided die, but not impossible. The probability that I'm using some kind of weighted or flawed die is statistically more likely. In software, it might just mean that the algorithm doesn't produce a satisfactory distribution of numbers, (like "return 3", when you expect numbers 1-10 with equal probability). That would be a valid complaint, but not against randomness, just on a "fair" distribution, which would affect the playability of the game and enjoyment derived.

Ethics is the measuring of morality. Morality is the measuring of good. Good is the measuring of benefit. Benefit is the measure of values.
Phaedruswax
banned

Usergroup: Members
Joined: Aug 05, 2009

Total Topics: 6
Total Posts: 104
Posted 08/06/09 - 04:25 AM:
quote post
#8
Unless said number generator has access to infinity (impossible) any number generated lacks a degree of randomness.
123savethewhales
Assistant Professor
Avatar

Usergroup: Members
Joined: Mar 25, 2009

Total Topics: 6
Total Posts: 278
Posted 08/07/09 - 03:31 AM:
quote post
#9
Phaedruswax wrote:
Unless said number generator has access to infinity (impossible) any number generated lacks a degree of randomness.

I am not sure why infinite is necessary in a random event. A coin flip only generates 2 states (heads or tails), and aside from the physical element of the coin toss itself, heads or tails seems to be consider random.

Keep it simple.
deepthought
Initiate
Avatar

Usergroup: Members
Joined: Aug 08, 2009

Total Topics: 0
Total Posts: 4
Posted 08/08/09 - 09:05 PM:
quote post
#10
Randomness and predictability are flawed assumptions inherently limited by our perception of time.
Download thread as

Page: 1 2



Sorry, you don't have permission to post. Log in, or register if you haven't yet.