I can see that. "Potential companion A is x% likely to put out, PC B y% likely, and PC C z% likely. Therefore, if I ask all three out at once, I am (1-((1-x)*(1-y)*(1-z)))% likely to get laid, and (1-((1-x)*(1-y)*(1-z))) is obviously a much higher number than x, y, or z individually, with the added bonus of a certain percentage chance of a threesome-or-better."

I totally don't see why that didn't work out for him.