EnumaElish said:
But, isn't ABDEFGHC a different outcome than ABDEFGCH? By your logic, these two would've counted as one.
If you consider all 8 people, then, yes, they are different outcomes- and the numerator of the probability fraction will be different than if you only consider A, B, C. But then you would have to consider
ACDEFGH
B (where they are not in ABC order) as different from
ACDEFG
BH. The denominator of the fraction is also different- the result is the same.
To take a simple example: suppose there are 4 people, A, B, C, D speaking. What is the probability that A, B, C speak in the order ABC?
There are, of course 4!= 24 different ways to order 4 things:
ABCD ABDC ACBD ACDB ADBC ADCB
BACD BADC BCAD BCDA BDAC BDCA
CBAD CBDA CABD CADB CDBA CDAB
DBCA DBAC DCBA DCAB DABC DACB
Of those, ABCD, ABDA, ADBC, DABC have A, B, C in that order: the probability that A, B, C will speak in that order is 4/24= 1/6.
If instead we count only A, B, C, we find 3!= 6 different orders and only 1 (ABC) has them in that order: Again 1/6.
Given any n people speaking in random order, the probability that a given 3 will speak in a given order is 1!/3!= 1/6. If you use all orders for n people as the denominator and all orders for n people that have the same 3 people in the given order, the numerator and denominator are both multiplied by n!/3!.