If EveryThingIsMath then every thing is predictable. 
A system has an input and an output. If we can predict all the inputs to a system, we can predict the output. 

However the precision of the output is dependant on the precision of the input. 

Also for most non-trivial systems, the resources (in time and hardware) needed to model the system completely may be greater than the resources that would be used in observing the original system.

To completely model the universe, we would need at least one universe. 

----

Not so. There's this thing called QuantumMechanics that says otherwise. You may have heard of it?

The mathematics behind quantum mechanics is entirely deterministic, right up until the point of observation.  Then you get non-determinism under the Copenhagen interpretation, but determinism survives intact under the multi-universe interpretation.  See QuantumPhysics.

''in any one universe, though, the result is that the experiment or whatever is not predictable. Which was my point. Or perhaps you'd like to predict when this radioactive atom I have here will decay?''.
----

Assumes that the mind (emotions, thought, plans) is predictable?  Or possibly irrelevant?  Gosh, I may have to dig up my CelestialBilliards example (which assumes a source entity with effectively unlimited extrapolative powers).

Is this the ultimate conclusion of ChaosTheory, or does it imply that ChaosTheory is incomplete?  -- GarryHamilton

----

Doesn't GoedelsIncompletenessTheorem come in here somewhere?

GregoryChaitin would be more relevant.