Thad: Ok, that makes sense, although once you introduce the possibility of a draw, it seems that the proposition hinges on a certain definition of "optimal." For instance, in chess, would it be optimal to play a move that assures you of a draw, or would it be optimal to play a move that risks losing for a chance at winning, depending on what your opponent does? Obviously, if we assume both sides are playing optimally, then you take the draw because an optimal play against your move will make you lose, if there is a chance you can. But people don't play optimally.

But you're talking theory, and I'm talking practice. Point taken.

But wouldn't a drawable version of pente be harder to master (to memorize optimal play) than a non-drawable version? I imagine (but am only guessing, having not done the analysis myself) that this slight rule change would cause most of the major lines to be re-analyzed, as the optimal play of even the most basic openings (wedge, for example) leads to very close games, if both sides play them optimally.