Silly idea: 1. take every email exchange you've had with past YC companies whose outcomes are now known. Good=bought, IPO, etc., Bad=no longer exists (w

Silly idea: 1. take every email exchange you've had with past YC companies whose outcomes are now known. Good=bought, IPO, etc., Bad=no longer exists (with exceptions for folks like Kiko?) 2. train a bayesian filter on the set 3. Apply to the rest with as yet unknown outcomes 4. Publish the results :D