/sci/ - Science & Math

[Anonymous No.1346474]

if i have random number generator that generates any number from [0, 1, 2, ...] with equal probability, what would be the expectation value?

[Anonymous No.1346483]

-∏

[Anonymous No.1346482]

infinity

[Anonymous No.1346499]

sex.

Cause that's what we're all up to.

[Anonymous No.1346504]

>>1346482
ok, and for ... -3, -2, -1, 0, 1, 2, 3 it would be 0?

[Anonymous No.1346507]

>>1346504
i mean ... -3, -2, -1, 0, 1, 2, 3, ...

[Anonymous No.1346525]

>>1346482
i think its 0, because the probability of each number is zero. and an infinite sum over 0 is still 0...

[Anonymous No.1346531]

i would agree 0>>1346474

[Anonymous No.1346540]

a uniform unbounded distribution?
That would be maximum uncertainty!

[Anonymous No.1346553]

1/∞

[Anonymous No.1346587]

>sort of related

I use Math.random in javascript. Anyone has a quick way to make randomer(truer) numbers?

i have an algorithm that makes divides the upper limit then adds the random numbers. so if I wanted a random number between 0 - 100. i would add Math,random(25) + Math,random(25) + Math,random(25) + Math,random(25). It kinda works but as you can see its extremely rare to get upper and lower bounds.. so...

[Anonymous No.1346771]

>>1346587
i am not great at java but from what i understand math.random is only psuedo-random so use the largest divisor possible? i don't think dividing the upper limit is actually making the distribution less true. i think the opposite might be happening where the more trials (the blocks created by using the divisors instead of the upper limits) are showing the true nature of the distribution. or i could be completely wrong.

[Anonymous No.1346790]

There is no expected outcome in cases where the probability of all outcomes is equal.

[Anonymous No.1346806]

>>1346587
Try
(Math.random()+Math.random())/2
Averaging two random numbers together makes it twice as random.

[Anonymous No.1346812]

>>1346790
There is in finite cases. For example, the expected value of choosing at random {1,2,3} is 2... similarly E({1,2,3,4})=2.5

[Anonymous No.1346838]

>>1346474
Impossible anyway. If it generated every number with equal probability p, then the total probability would be

P(1)+P(2)+P(3)+P(4)+P(5)+...= p+p+p+p+p...

which is either 0 or infinity depending on whether p is or isn't zero. In any case, it can't be 1 like it has to be.

[Anonymous No.1346853]

>>1346806

[Anonymous No.1346857]

>>1346812
I don't know the actual meaning of expectation in this context, but I'd disagree with you. You're assuming that the usage of the number produced reflects the numerical value of that number.

Consider instead selecting at random from the set {yellow, crawdad, Thursday}. Would the expectation value be crawdad?

[Anonymous No.1346875]

>>1346790
http://en.wikipedia.org/wiki/Expected_value

[Anonymous No.1346878]

there is no random number generator that generates any number from [0,1,2,..] with equal probability.

[Anonymous No.1346896]

>>1346587
Sawtooth(x)=arctan(tan(x))/pi+1/2
Sawtooth(x) has period pi

Rand(x)=Sawtooth(Math.random()*1000*new Date().getTime())
True random arguably doesn't exist on deterministic computers. This function will give you a uniform distribution (just like Math.random()).

>>1346857
http://en.wikipedia.org/wiki/Expected_value

[Anonymous No.1346913]

but you could multiply a random number generator [0;1] with random sign - or + and multiply by a random constant

[Anonymous No.1346922]

>>1346896
>http://en.wikipedia.org/wiki/Expected_value

Gotcha, thanks.

[Anonymous No.1346940]

1, -1, 2, -2, 3, -3... <span class="math">\pm\infty[/spoiler]

[Anonymous No.1346968]

>>1346857
> S={yellow, crawdad, Thursday}
<div class="math">E(\mathbf{S})=\sum_i \mathbf{S}_i P(\mathbf{S}_i)</div>
E(S)=(yellow+crawdad+Thursday)/3

[Anonymous No.1346973]

Lmfao, enjoy your shitty 'random' numbers that aren't really random

[Indexer !0bdFZYyCZM No.1346991]

>>1346973
except "random" is just a concept

enjoy your theories.