EDG said:

For Ic are you remembering to do the -4 DM for atm 0 (and it applies to nothing else in the MGT rules)?

Not if it wasn't listed in the table at the top of the thread, so that would explain that one.

Really, what I did was this:

1. Calculated all combinations possible for the first roll (SZ or Pop) followed by a secondary roll (ATM,HYD and GT), (2d6x2d6=36^2 combinations.

2. Assigned each of the traveller codes 0-C (or A if capped at A) an outcome frequency

3. and then used those frequencies to compute the probability of any one trade code.

The frequency table is the first post in the stats thread.

I used the list posted by Grufty to assemble the trade code probs.

Most of the small differences are likely due to the fact that you're estimating, and I'm using exact counts....given the numbers of runs, your runs probably fall within a reasonable error around the exact count.

Not sure about the bigger ones. You're right about why. Your code, my process, who knows ? I'll check what I used for conditional odds in the basic permutations to code value part - thats where it'll show up.

I'm pretty sure about the permutations table-those is straightforward products.

edit: Also, Looking at the differences in our lists, even some of the larger discrepencies seem to be within margin for observed error - if you don't know the variance of a sample distribution (and we don't) , the best rule of thumb is to use the observed average +- about half.

so, a result of 18 has an error band including values from 9 to 27. Some of the more discrepant of your results do include the exact count value in such an error band, so your differences are likely it due simply to random variance.

Its not a great rule of thumb, but it works surprisingly well for % estimations.