I first produced this page in February 2001 following a Scientific American article by Ian Stewart. In 2002 Steve Schaefer contacted me to tell be he had calculated the number of solutions taking symmetry into account, together with many other calculations on a similar theme, such as the fact that (modulo symmetries) there are only 765 possible achievable positions if X goes first, and even fewer if the players are even moderately sensible. On this basis his equivalent table to that above would look like: