Revert 148639 - Clean up Prerender field trials.

Revert 148639 - Clean up Prerender field trials.

Remove the second comparison group from the prerender field trials, and reorganize the code so that we can compile assert on the summation. The second group has already shown us our numbers are good. The compile time assertion should avoid fat finger errors that we've caught in other reviews (which would otherwise have been a _beta crasher_!).

I think we should also give consideration to renaming these histograms for ease of understanding. I know that makes a historic comparison cost, but it makes arriving new in this area so much easier.

I'll rebase the multiple prerender CL on top of this.

R=dominich@chromium.org
BUG=None

TEST=


Review URL: https://chromiumcodereview.appspot.com/10802062

TBR=gavinp@chromium.org
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=148643