How to Model the Parents as a Conditional Probability Distribution

The do operator is the more intuitive thing, I think to humans. We think about intervening on things, and it's actually the conditional probability thing that's less intuitive. That's why confounding is so important, because it will, confounding will affect conditional probabilities. But then if you have a do behind the do operator, like do change this action to this action, it will do what you intuitively think it do. It'll be measuring what would happen if I were to take this action. Do calculus can produce many different algorithms for these causal mechanisms.