Hey,

I was just wondering, how do you get a random value, with a normal distribution?

Ask Mr Google

https://github.com/torch/torch7/blob/master/doc/random.md

I’m still confused, how do you import the torch thing, the luarock et cetera didn’t work.

You don’t need a mersenne generator, keep looking

Like is it a simple command, or do I need to write a function to do it?

Function

Thought I’d give this a try.

```
displayMode(FULLSCREEN)
supportedOrientations(LANDSCAPE_ANY)
function setup()
lim=30000
create()
end
function create()
tab={}
for a=1,lim do
-- create a random value with a normal distribution
p=math.sqrt(-2*math.log(1-math.random()))
x=p*math.cos(2*math.pi*math.random()) -- x = random value
-- create a table showing the distribution of all the values
v=(x*120)//1
if tab[v]==nil then
tab[v]=1
else
tab[v]=tab[v]+1
end
end
end
function draw()
background(0)
stroke(255)
strokeWidth(2)
for a,b in pairs(tab) do
line(WIDTH/2+a,50,WIDTH/2+a,5*b+50)
end
stroke(255,0,0)
strokeWidth(3)
line(0,50,WIDTH,50)
line(WIDTH/2,50,WIDTH/2,HEIGHT)
fill(255)
text("Graph showing normal distribution of "..lim.." random values.",WIDTH/2,HEIGHT-20)
text("Tap restart icon for another distribution",WIDTH/2,HEIGHT-50)
end
```

Here’s another version showing the distribution of 30,000 points around the center of the screen.

EDIT: This shows a good example of what a globular star cluster looks like. See this link.

http://earthsky.org/clusters-nebulae-galaxies/omega-centauri-milky-ways-prize-star-cluster

```
displayMode(FULLSCREEN)
function setup()
tab={}
for a=1,30000 do
p=math.sqrt(-2*math.log(1-math.random()))
x=p*math.cos(2*math.pi*math.random())
p=math.sqrt(-2*math.log(1-math.random()))
y=p*math.cos(2*math.pi*math.random())
table.insert(tab,vec2(x,y))
end
end
function draw()
background(0)
fill(255)
for a,b in pairs(tab) do
ellipse(WIDTH/2+b.x*100,HEIGHT/2+b.y*100,3)
end
end
```

@dave1707 Thanks! Can you explain the maths behind it? I was trying to do it using only one random value, but supposedly you can’t integrate e^-x^2. Does Lua have the error function?

EDIT: I’m going to try an approximation of the error function later, and I’ll send what I end up with.

@MattthewLXXIII I can’t explain the math. I did a google search for normal distribution of random numbers and found various formulas. I tried to find the link to the one that was the best, but I wasn’t able to find it again. I don’t remember the exact wording I used in the search. I’m not sure what error function you’re referring to, but I doubt if Lua has it.