
Originally Posted by
Ultra Mackness
Interesting you've posted this - in Victoria we have a new study design for senior maths which includes this exact thing.
We actually have two things going on here that can't be mistaken for the same thing. First is the probabilities of success for the three outcomes you've mentioned. These exist as long term expected values or global aspects.
The second thing is the sample aspects. Each time you clicked out 100 of these virtual die rolls, you did a sample of 100, and there's absolutely no guarantee that in any sample of 100 that you'll get the sample aspects matching the global aspects, only a level of confidence that your sample aspects will match the global aspects.
In theory, you need to do about 33 of these samples (3300 clicks!) in order to be about 95% confident that your sample aspects will match your global aspects. Of course that also means that you're 5% confident that your sample aspects won't match. Downside is that this parameter can be tweaked by the game designers to crank up the number of samples you need to take, shifting 33 up to... well however high they want.
Either way, you'll still need to do a lot of clicking! Good luck!