# Using a weighted table?

I'm writing a simple computer program that has to use the following weighted table to pick a random "payment" where the relative probabilities are given by "weight".

How does one interpret the table above and use it to determine how much to "pay"?

EDIT: After speaking with some folks it seems this table is simple enough that I can just create a table with 41 entries where each pay amount appears ( Weight times ) and select a random number between 0 and 40 and use that to index into the table.

-
You should ask your boss. – martin Aug 23 '12 at 22:18
@kxx Thanks for the help. – Chimera Aug 23 '12 at 22:19
So, you want to pick a random "payment" where the relative probabilities are given by "weight"? If that's not what you meant, please edit the question to clarify – Mooing Duck Aug 23 '12 at 23:04

Assuming that the weight vector defines a finite probability distribution, you can sample payments as follows.

First pick a random integer $x \in [0, 41]$. Then the payment is as follows:

$0 \le x \le 20$ : $200 + 50 \lfloor x/3 \rfloor$

$21 \le x \le 38$ : $550 + 50 \lfloor (x-21)/2 \rfloor$

$39 \le x \le 41$ : $1000 + 500 (x - 39)$

-
How do you get 41? Along with this table came the statement, "Can be changed to a table of 41 objects of equal probability. The 200 would appear on it 3 times." – Chimera Aug 23 '12 at 22:40
I ask because I'm just trying to understand how to interpret the table so the next time I get something like this I can figure it out myself. Thanks.. – Chimera Aug 23 '12 at 22:41
You probably got the table from an excel sheet, and excel probably has a function that samples the payment automatically given these two columns. Maybe you should ask someone who knows something about excel. – martin Aug 23 '12 at 22:51
Thanks for all of the help. However, I'm not interested in learning about Excel, I'm trying to understand how to use such a probability table in a program. I'm looking to understand the math behind it. – Chimera Aug 23 '12 at 23:05