Paul GrahamVerified account

Today in Paul's implicit assumptions: that only the initial selection bias affects outcomes, not later biases. https://twitter.com/paulg/status/660534510379888640 …

@kevinmarks Did you actually read it? "And in particular it must not be invalidated by the bias you're trying to measure."

@paulg yes, you disclaimed that, but then went on to apply it outcomes of vc funding of women, implying that the disclaimer didn't apply.

@kevinmarks It doesn't apply when you're showing bias *exists*. It means FRC's 63% number is a lower bound on their bias.

@paulg if their female founder numbers look like YC's, I'd be surprised if there are enough to be statistically significant.

@kevinmarks Should I take this resort to (inaccurate) snark as a sign you've been refuted?

@paulg last time we discussed this you said the figures were inaccurate, but refused to give accurate ones https://docs.google.com/spreadsheets/d/1tZVHgKnNS5ghjkQou4QOyQtQWX9sURgSMVYgyC76z-Q/edit?usp=sharing …