DS student here, trying to grasp DeepCFR, amazing video... just wanted to point out a thing... the "P(K|Qb)" you say that you use Bayes rule and you phrase it like "P(P1 would play bet when they have a king)/P(P1 would play bet overall)"... I might be mistaking, but i don't see how you are applying Bayes... instead, I got your same result using the definition of conditional probability P(a|b) = P(a,b)/P(b)... because I get "P(KQb)/P(Qb)", now Q and b/K are independent events, so "[P(Kb)P(Q)]/[P(Q)P(b)]"... from here, applying the definition of joint probability and total probability, I get your same final probability Apart from this, thank you so much
You're correct, he's not using baye's rule/thm. It's the definition of conditional probability in Bayesian stats that he's using. It's actually fairly common for people to mix them up. Nbd
