In a thread below - Mike said:
I relish a good argument. I've recently provided several good arguments regarding the futility of deterministic prediction, and other, more specific issues, including the chaotic nature of the earthquake nucleation process, the human failings in pattern recognition, the Full Moon Effect, and much more. The responses have been only tangential.

I think Earthquake prediction studies rely on something like faith - it's too easy to say - It's Chaotic in nature an we'll never figure it out.. so why bother?

Well.. there have been some great debates on this argument - but my 2 cents worth is even though it appears impossible.. maybe with enough data and neural nets, we may someday be able to massage enough data to successfully predict.

We have to remember that our 'science lab' is the planet earth.. we have many quakes a day, but have really only been keeping reliable data for about 40 years.. given that it can be hundreds of years between failures on fault systems, this whole thing is in it's infancy.

My goal is merely to keep the conversation alive as to it being possible and to keep looking for the holy grail... just saying it's not possible now, does not mean it won't be possible in the future.. and then we can site things like the world being flat, the sun revolves around the earth, etc..

Canie