So, yeah, I work for Google, but I have no specific knowledge of the Lunar X Prize. I just took a look at their home page, saw the brief summary of the rules, but didn't find a complete draft. It looks like they are going to revise the rules a bit after some feedback.
Here's what I've been thinking: if you want to land on the moon, look around, and then get close to something else and take pictures of it, you don't really need wheels, because you've already got a rocket that knows how to land.
In the moon's soft gravity, it takes fairly small amounts of delta-v to jump a long way. In the moon's 1.62 m/s^2 gravity, you can get 50 seconds of flight time with a 82 m/s delta-v. Use some more delta-v to go sideways, and a bit more to manoever for the landing, and you could cover 500 meters with about 100 m/s of delta-v.
Landing from a lunar orbit takes 1600 m/s of delta-v, so adding a few hundred for a few hops is not a huge increase. Yes, it's exponential, but if done with LOX/kerosene or hypergolics, a 2000 m/s total delta-v budget for the lander implies a very reasonable mass ratio.
Why hasn't it been done before? Multiple rocket hops would have been stupid for the manned mission, because the landing was the highest-risk portion of the mission. It's still the highest-risk portion, and the lunar hopper idea stands a very good chance of crashing one of it's landings. But that's okay, because after a few hops the thing will run out of gas and be dead anyway.