testtest

asdfasdfasdf

4. ## Re: testtest

Dealing with random numbers is never as straightforward as people think. And when you want them from a skewed distribution, it gets even more complex.

5. ## Re: testtest

Therefore, if you're truly wanting what you suggest, the first order of business is to figure out from which skewed distribution you're pulling your random number: Poisson, Gauss with a skew value, logarithmic, etc. There are many to choose from. Possibly you even want to bracket part of one of those skewed curves. And, from there, you can start devising your algorithm to always sum to a specific value.

Also, you will need to decide if the Rnd function is good enough for you, or whether you wish to use the more robust CryptGenRandom (or other) API call (where you get more robust random numbers). And, again, in any of those cases, the results will need to be transformed such that they come from your skewed distribution.

-----------

Dealing with random numbers is never as straightforward as people think. And when you want them from a skewed distribution, it gets even more complex.

Good Luck,
Elroy

6. ## Re: testtest

Ok, just seeing this. First, and foremost, 30 numbers ranging from 1 to 6 will never be

7. ## Re: testtest

be Uniform Random (which is what I believe you're suggesting).

8. ## Re: testtest

. They can be Skewed Random, which pulls us

9. ## Re: testtest

us into an entirely different area. For a set number of integers (say 30) with a specific range (say 1 to 6) to have

10. ## Re: testtest

have a Uniform Random distribution, the mid-point of the range (3.5) times the number of integers (30) must equal the total your after ...

11. ## Re: testtest

3.5 * 30 = 105 (not 100).

(not 100).

13. ## Re: testtest

after ... and 3.5 times 30 = 105

14. ## Re: testtest

Ok, just seeing this. First, and foremost, 30 numbers ranging from 1 to 6 will never be Uniform Random (which is what I believe you're suggesting). They can be Skewed Random, which pulls us into an entirely different area. For a set number of integers (say 30) with a specific range (say 1 to 6) to have a Uniform Random distribution, the mid-point of the range (3.5) times the number of integers (30) must equal the total your after ... and 3.5 times 30 = 105 (not 100).

Therefore, if you're truly wanting what you suggest, the first order of business is to figure out from which skewed distribution you're pulling your random number: Poisson, Gauss with a skew value, logarithmic, etc. There are many to choose from. Possibly you even want to bracket part of one of those skewed curves. And, from there, you can start devising your algorithm to always sum to a specific value.

Also, you will need to decide if the Rnd function is good enough for you, or whether you wish to use the more robust CryptGenRandom (or other) API call (where you get more robust random numbers). And, again, in any of those cases, the results will need to be transformed such that they come from your skewed distribution.

-----------

Dealing with random numbers is never as straightforward as people think. And when you want them from a skewed distribution, it gets even more complex.

Good Luck,
Elroy

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•

Featured