heck no maybe the WHOLE summer prehaps for the exalted one and the 1990-2000 is a matter of luck even by pure mathmatical probability there is 4,000,000 different outcomes of AB.

There are not 4 million outcomes... it ends on multiples of ten and one team must have 2000. The actual number of endings is more along the lines of:
2(2000/10)=400

So, if AB was purely random, your chance every game for We Had It All Along is .25%. (1/400).

However, it's not purely random. There are influences and gameplay mechanics that make is very easy to end up with specific scores (such as 0-2000). A close game with two very well played teams will yield higher chances of getting the sweet spot. I've been in two games so far (out of 100) that ended with a score of 2000-1990. Unfortunately for me, the Horde won both.

@Physis... Actually, the number of possible outcomes is 40,000. Since the scores can range from 0 to 2000 on either end, and we're working with multiples of 10, you're looking at 200*200 instead of the formula you came up with. (Consider that for each Horde score there are 200 Alliance possibilities, and vice-versa.)

Molloch, how many games of AB have you played that didn't have one of the teams on 2000 points at the end? Think.

"Consider that for each Horde score there are 200 Alliance possibilities, and vice-versa."

You have the right idea, but you need to remember to isolate it for Victory conditions:
For a Horde win (2000 points) there are 200 alliance possible scores ranging from 0-2000. This means at the end of a Horde victory, there are two hundred possible scores the Alliance can have. If you ADD (not multiply) the combinations of Alliance wins, you get another 200. So 400 ENDING scores.

If you come into a game at a random moment, there are 40,000 possible scores during the course of the game (like you said 200*200). For the sake of "We Had It All Along" we only care about the ending scores though.

The probability remains 1/400. A horrifying prospect if you are striving for it, but much less worrying than 1/40,000 or 1/4,000,000.

Arrgghh - I take a break from my maths statistics class and read this! And the worse thing is I still can't determine which is the correct 'null hypothesis' and which is the alternative!

Whomever said 'lies, damned lies, and statistics' had the right idea ;)