How to read proof without being sold

A 70% win rate can lose money, a 20-trade record carries a t-statistic of 0.45, and 58 of any 1,000 coin-flippers will produce a publishable one. Selection, not dishonesty, is what stands between you and a track record.

What you will be able to do

After this lesson you can compute the expectancy behind any win rate, state how many samples a claimed edge would need before it separates from noise, and explain why selection makes an honest track record uninformative.

Inside this lesson

What a proof page is allowed to be
Win rate is half a statistic
How many samples would it take?
The pool you were never shown
Never screenshot a classroom blotter as a sales card

This lesson is on Scholar

Audit includes orientation and the two-rooms lesson. The rest of the path is paid course access — still education, still not advice.

Included on Scholar · $79/month tuition.