Given P(she smiles at you | she likes you) is approximately 1

Assuming P(she likes you) > 0

Then we can sub into

P(she likes you | she smiles at u) = (P(she smiles at you | she likes you) * P(she likes you))/P(she smiles in general)

To get P(she likes you | she smiles at u) = x/P(she smiles in general)

Where x is some number between 0 and 1

Therefore we can conclude the more she smiles the less likely its cos she likes you. Therefore find the oens who dont smile and get em to smile. Therefore bigtiddygothgf.

unfortunately, P(she likes you) ≈ 0

Nice, we can also assume that P(she smiles at you | she likes you) is approximately 1, simplifying to only 2 variables, and also substitute P(you are likable) for P(she likes you) to remove all unknowns.

Damn you priors

So lowest denominator possible (rarely smiles)?

I'd be a lot less confused by a Venn diagram.