[jdlshore]

I'm not qualified to comment on Cam's math, but Evan Miller's tools [1] are my go-to source for calculating sample sizes and significance. How does this differ? (If it does.)

http://www.evanmiller.org/ab-testing/sample-size.html