I've always been fascinated with games as battles of the wits (Museum Caper Clue anyone?). I'm also fascinated with artificial intelligence, so it should probably come as no surprise that I would try to write a game where you match wits against an AI rival.

The premise of Kill Eustace is based on the scene in The Princess Bride where the protagonist and the Sicilian go head to head with two goblets of wine -one poisoned.

In Kill Eustace, you and Eustace are both trying to kill each other, but the objectives are actually a little deeper than that; you are trying to get the necessary items on the table to complete your plot while simultaneously preventing Eustace from doing the same. On top of this is a struggle for information; you are trying to determine which items Eustace needs while preventing him from knowing which items you need.

I wrote two forms of the AI to match a player learning how to play. The first AI has a very literal approach; it assumes the player will attempt to bring his needed items to the table directly and attacks them directly. The first AI does not think about preserving his own needed items and does not attempt to hide what he needs. The second AI assumes that the player is taking a more circuitous route and reacts accordingly; this AI is also focused on preserving his needed items. Because of these two considerations, the second AI takes greater risks and it is actually possible to kill him more quickly than the first AI, but it is also much more difficult to drag him to a draw.

As a game, Kill Eustace is certainly not for everybody. In fact, it may just be a game for me, but I'm ok with that.