Scott,

I think you're seeing the effects of innacurate orbit integration. If you take the applet and jump straight to January 2008, there is nothing like a close approach. Perhaps if you run it forwards from now to then it might get close, but these things are distinctly unreliable.

Nonetheless, the odds of an impact with ANY terrestrial planet must be over twice that of impact with the Earth, and perhaps out to about 3 times if you include Mercury (Kinda hard to see what happens there, though, due to the Sun glare). One day we'll get to image a collision in the inner Solar System.

Just hope it isn't us. Imagine an Atlantic tsunami as bad as the Indian Ocean one...