Retroanalysis and Tic-Tac-Toe

Joseph Kisenwether, 18 Mar 1995

By Les Marvin
Originally published in the Journal of Recreational Mathematics
(Sorry, I don't have the vol or issue # )
           |   |
         X | O | O
           |   | X
           | X | O
           |   |
The above position was reached in a Tic-Tac-Toe game between two experts
(meaning that neither one ever missed a chance at a forced win, or gave one up)
What were the first and the last moves of the game?
This is a rather silly example (solution left as an exercise for the reader)
but it brings up the point.  Does anybody know of any problems involving
retrograde analysis applied to something other than chess? Please post them
or E-mail them to me if you do.
Also, regarding chess variants, how different do you need to make a game before
it stops being a chess variant and starts being an independednt game? I mean, after
a lot of rules modifications we could call the problem above a chess-variant 
problem but everybody knows Tic-Tac-Toe is a completely different game. 
Just wanted to hear your thoughts on the subject.