Page 1 of 1
random number generator
Posted: Fri Feb 11, 2011 3:28 pm
by e4ans
Is it possible to upload a excel file on here? I have made a simple RNG to simulate FOBT roulette. In 200 presses only once did the same number come out twice in a row
Posted: Fri Feb 11, 2011 4:15 pm
by harry2
Rand function !
.xls files not allowed. Save it to JPEG or paint.
viewtopic.php?t=974&start=140
Posted: Fri Feb 11, 2011 4:21 pm
by quizard
It is actually some very high probability (about 99.5%) that the same number will come up consecutively at least once in 200 spins.
You can test it out here:
http://www.random.org/integers/
Posted: Fri Feb 11, 2011 4:44 pm
by e4ans
harry2 wrote:Rand function !
.xls files not allowed. Save it to JPEG or paint.
Spot on
=RANDBETWEEN(0,36)
Posted: Sun Feb 13, 2011 6:08 pm
by LordOoze
Generate a stream of random numbers then plot them as either a 2d or a 3d - to test for structures.
for 2d - plot n against n+1
for 3d plot n against n+1 against n+2
A good random number generator will show an equal distribution whereas a poor one will reveal clusters.
Oh and see if your generator will create a unique series each time it is called. Whether it called once each time the environment used to program it is used or sequentially.
An example is the basic rand function in matlab (c) will create the approx same series each time you initialize matlab (c)
Posted: Sun Feb 13, 2011 9:09 pm
by blackmogu
LordOoze wrote:Generate a stream of random numbers then plot them as either a 2d or a 3d - to test for structures.
for 2d - plot n against n+1
for 3d plot n against n+1 against n+2
A good random number generator will show an equal distribution whereas a poor one will reveal clusters.
Oh and see if your generator will create a unique series each time it is called. Whether it called once each time the environment used to program it is used or sequentially.
An example is the basic rand function in matlab (c) will create the approx same series each time you initialize matlab (c)
modern random functions can be seeded for a different sequence upon startup.
Posted: Sun Feb 13, 2011 9:15 pm
by Matt Vinyl
Further reading on how Excel generates random numbers: [url=
http://en.wikipedia.org/wiki/Mersenne_twister]Wiki Link: Mersenne Twister Algorithm[url]
Pseudo Random Algorithm. I'd still say these sort of randomiser algorithms are more than OK for most 'random' games?
Posted: Sun Feb 13, 2011 9:18 pm
by Matt Vinyl
Although this:
Observing a sufficient number of iterates (624 in the case of MT19937) allows one to predict all future iterates
is interesting...
Posted: Sun Feb 13, 2011 9:43 pm
by blackmogu
probably down to a flawed implementation of MT19937 in Excel.
Posted: Sun Feb 13, 2011 10:01 pm
by Matt Vinyl
Better get our notepads and a stack of those Ladbrokes pens and get noting down those numbers.
