Tue Jun 10 01:33:22 2014 by Foxiekins 
I may be breaking Troll, but I was trying to use it to guage success probability of a roll versus a difficulty roll. Difficulty rolls are a number of dice ranging from zero to 5, which is a z5, of course. Ability rolls range from d2 up. I was using "if d2 > z5 then choose(1) else choose(0)" to force the results into two categories, success and fail. There may be a better way to do this, but I couldn't figure one out. My first test gave a 25% chance of success, which I expected, but when I changed d2 > z5 to d2 > 2z5, instead of the chance of success going down, it went up. I figure I must be misunderstanding something, but I am at a loss as to what. 

Tue Jun 10 08:57:28 2014 by Torben 
Your idea is fine, though you can shorten choose(1) to 1 and choose(0) to 0 . What goes wrong in the second case is that 2z5 means two individual z5 s rather than the sum of the two. So the condition d2 > 2z5 is true if the d2 is greater than any of the two z5 s. To get what you want, you should write if d2 > sum 2z5 then 1 else 0 . 

Wed Jun 11 07:49:27 2014 by Foxiekins 
Ah, that was the bit I didn't understand... sum 2z5 instead of 2z5... Thanks... 

Mon Jun 16 05:06:12 2014 by Foxiekins 
Ran into a bit more of a snag. If I understand correctly, d12+d10 should be a single value, like the roll of one die. But, when I have d12+d10 on one side of the condition, I get the error "Distribution error: Singleton expected at line 1, column 6." When I try sum (d12,d10), I get the error "Parseerror at line 1, column 11." And when I try sum (d12 U d10), I get the error "Distribution error: Nonsingleton used with > at line 1, column 19." I am at a loss for how to have more than one die in place of the d2 in the original comparison. 

Mon Jun 16 09:43:51 2014 by Torben 
Try writing (d12+d10) instead. You have probably run into a peculiarity of Troll: Comparison operators bind stronger than arithmetic so, for example, d6 < d8 + d10 reads as (d6 < d8) + d10 , which gives an error if the result on the d6 is not smaller than the result of the d8 (and, in any case, is probably not what you meant), 
