![]() |
|
probability and graphs |
Richard Phipps
Member #1,632
November 2001
![]() |
Just a quick question. Say we have a variable which we calculate by adding up two random numbers between 0-100 and then dividing the total by two. If we then count the number of times each final number (0-100) comes up after a few thousands iterations and plot it on a graph, would the output be a triangle shape or a bell shaped curve?
|
GullRaDriel
Member #3,861
September 2003
![]() |
It depend of your random number generator algorythm. "Code is like shit - it only smells if it is not yours" |
Richard Phipps
Member #1,632
November 2001
![]() |
Well, let's suppose for the sake of argument that it's a totally random output. |
HoHo
Member #4,534
April 2004
![]() |
see attachment. I took 2x512 random numbers, added them up and took the average as you said. Sorted the resulting numbers and created a chart out of it. If the line would be linear it would be a linear distribution. From what I can rad out of it it really is a bell distribution. __________ |
Richard Phipps
Member #1,632
November 2001
![]() |
Why thank you! I didn't expect someone to go to that much trouble. I just thought some of the math people here already knew the answer from their studies. Cheers! |
Johan Halmén
Member #1,550
September 2001
|
Hm, let's throw two dice. result probability 1 0 2 1/36 3 2/36 4 3/36 5 4/36 6 5/36 7 6/36 8 5/36 9 4/36 10 3/36 11 2/36 12 1/36 Now wouldn't that be a triangle? You need more dice to get the bell. One die gives a line, two dice give a triangle, more dice twist the triangle towards the bell. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Years of thorough research have revealed that what people find beautiful about the Mandelbrot set is not the set itself, but all the rest. |
Richard Phipps
Member #1,632
November 2001
![]() |
Johan: You are not dividing the result by 2 are you? |
Evert
Member #794
November 2000
![]() |
Quote: would the output be a triangle shape or a bell shaped curve?
It will approach a bell shaped if the initial distribution is flat. Punchline: if you sum two distributions, the result is a new and different distribution (a sum of flat distributions is no longer flat, nor is the a of normal distributions normal). |
Johan Halmén
Member #1,550
September 2001
|
Why should I divide it? Dividing is just scaling linearly, right? Gosh, now I don't believe in myself anymore. Got to code that myself. Hate you if I was right. Hate you if I was wrong, too. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Years of thorough research have revealed that what people find beautiful about the Mandelbrot set is not the set itself, but all the rest. |
Richard Phipps
Member #1,632
November 2001
![]() |
Now I'm confused too! |
miran
Member #2,407
June 2002
|
Compile this:
It will draw a triangle. -- |
Evert
Member #794
November 2000
![]() |
Quote: Why should I divide it? Dividing is just scaling linearly, right? Yes. You should divide it if you want to keep the same normalisation. |
Arthur Kalliokoski
Second in Command
February 2005
![]() |
To say it all a different way, let's throw the dice again. To throw a 2 both dice have to fall on one's, no exceptions. But to get a 7, you could have a "1+6", "2+5" "3+4" "4+3" "5+2" or "6+1". Well, I suppose listing the result to get 2 would take "1+1" and "1+1" for symmetry. They all watch too much MSNBC... they get ideas. |
Evert
Member #794
November 2000
![]() |
To clarify what I said in my first post: |
X-G
Member #856
December 2000
![]() |
Quote: Yes. You should divide it if you want to keep the same normalisation. But we're only interested in the distribution, which shouldn't change by scaling it. -- |
Evert
Member #794
November 2000
![]() |
True, but it's easier to compare the distributions if they're both normalised. |
Richard Phipps
Member #1,632
November 2001
![]() |
So how do you get a bell shaped curve with probabilites? (None of this is really relevant I guess.. Just curious) |
Matthew Leverton
Supreme Loser
January 1999
![]() |
X-G
Member #856
December 2000
![]() |
-- |
Richard Phipps
Member #1,632
November 2001
![]() |
Johan Halmén
Member #1,550
September 2001
|
miran said: It will draw a triangle. Thanks, miran. Now I don't have to hate anyone. [added] Arthur said: Well, I suppose listing the result to get 2 would take "1+1" and "1+1" for symmetry. No. There's only one 1+1 case. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Years of thorough research have revealed that what people find beautiful about the Mandelbrot set is not the set itself, but all the rest. |
|