keithcurtis said:
What specifically do you see as a problem, (i.e. that is not mathematically supportable) in Roll20's method of producing random die results? Anecdotal reports are not mathematically supportable.
A hypothetical case, not specific mine:
Let's say that you played 10 game sessions, with like 20 die rolls per session. At every session, at least one time you get 3 consecutives same results on die rolls, and that this only happens when the rolls are made short time after the one another. If you wait some seconds to roll again, the problem is not shown. Do you think this is normal? Do you believe that this might happen so often? Is this human perception? Would you try to get information from the developers?
Could you walk me through that? The smaller the sample size the larger the variance. This is basic to probability and statistics. If you feel that Roll20 produces too many duplicates in a row, then a large sample size absolutely will reveal that, given the proper analysis. Look for patterns.
I will try! Imagine another hypothetical case:
You are trying to generate groups of 10000 random numbers for a program in a universe of 1 to 100. As you studied the final result set statistically, everything seemed great. Than a user informs you that they a getting consecutive result, like 5 to 8 results constantly, everytime they executes the program.
And then, you don't get offended, and looks part of the result set, noting multiple times, from time to time, consecutive rolls like this: (1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1), ( 2, 2, 2, 2, 2, 2, 2, 2 ,2 ,2 ,2), (3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3)... (100, 100, 100, 100, 100, 100, 100). So you realize that this is not OK, because this is a "random" roll and is statistically and probability WRONG. Although your program is not wrong, the result is. You would never perceive the problem analyzing big chunks of data, because with (BASIC, as you mentioned) statistics and probability this (1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1), ( 2, 2, 2, 2, 2, 2, 2, 2 ,2 ,2 ,2), (3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3), (100, 100, 100, 100, 100, 100, 100) is similar to (1,1, 2, 2, 3, 3...100, 100). Finally, you analize your code and sees that, with your algorithm, when user does not adjust his computer time/clock, the problem happens. You curse the user and all his past, but at least learnt something new about Stats and Probs.
Be well, friend!