true randomization

[GodForsakenSoul]

the only form of random number gen i can think of is.

dim $n
$n=Random(1,10,1)

and even that is not exactly random because most of the hits are 1-5. once in a while i get more.

is there anything more random then Random()? (see what i did there? ;D)

[99ojo]

Hi,

as you can see:

#include <array.au3>

Dim $arrandom [11] [2]
For $i = 1 To 10
    $arrandom [$i] [0] = $i
    $arrandom [$i] [1] = 0
Next
$count = 0
While $count <= 100000
    $n=Random(1,10,1)
    $arrandom [_ArraySearch ($arrandom, $n)] [1] = $arrandom [_ArraySearch ($arrandom, $n)] [1] + 1
    $count += 1
WEnd
_ArrayDisplay ($arrandom)

i will get at least about 10000 for each number.

;-))

Stefan

[Das Ami]

How random does it need to be? You could record the mouse movements of the user, I guess that's the closest to true randomization there is. (this method is also used by Putty)

[trancexx]

the only form of random number gen i can think of is.

dim $n
$n=Random(1,10,1)

and even that is not exactly random because most of the hits are 1-5. once in a while i get more.

is there anything more random then Random()? (see what i did there? ;D)

Random() is good. You are interpreting something wrong when you say that.

Run this to see how 'random' is Random():

#include <Array.au3>

Global $iNum
Global $aArray[11]


For $i = 1 To 200000
    $iNum = Random(0, 10, 1)
    $aArray[$iNum] += 1
Next

_ArrayDisplay($aArray)

Question is what do you see when that array is displayed to you.

Edited December 22, 2009 by trancexx

[GodForsakenSoul]

How random does it need to be? You could record the mouse movements of the user, I guess that's the closest to true randomization there is. (this method is also used by Putty)

as random as possible. there will be not much mouse moves... i think...

also, i don't know how to record mouse movements XD;;

[jchd]

as random as possible. there will be not much mouse moves... i think...

also, i don't know how to record mouse movements XD;;

Of course, replace "random" by pseudo-random everywhere.

AFAIK the algorithm implemented by AutoIt Random() is excellent and passes whatever test you need for layman applications.

OTOH if you feel you need cryptographic-level pseudo-random input, then rely on proven layers. But do you really need that?

I wouldn't bet my life (nor yours) on how "random" mice movements can be. That's in my view a very bad choice compared to a well know algorithm with proven track record and demonstrated properties.