Greetings !
Purhaps at this partially "critical" stage
of the discussion it may be worth noting
that despite the apparent fact that math
is the best discriptive language we appear
to have so far for the type of observation
data we have - it's not that good at it's job !
Starting from the "simplest" problems of Newtonian
Mechanics and up to the most difficult problems of
modern physics, math is certainly not perfect.
The simplest example is the 3+ body problem that
appears to be unsolvable by mathematics. Also, every
time you use integrals, for example, you're making
approximations and these are often your only way
of adressing what you observe. I'm not even beginning
to discuss stuff like the Chaos theory according to
which we can only get multi-level approximations.
The point I'm trying to make, if that's not clear
enough yet, is that if math had a direct connection
to the Universe you'd expect it to be able to explain
it better, wouldn't you ?
Live long and prosper.