III Probability

[shomurali]

• Skip HG since not really used in statistics, and if anything, do Poisson instead.
• Bayes' or no Bayes? Maybe commit and do base rate fallacy etc?
• Fix inconsistency in RV treatment (not sure what Andrew means, ask him).

[andrewpbray]

Hi @shomurali I'm resurrecting this thread to jot down some feedback about the notes from Fall 2023.

Intro to Probability

1. I'm afraid I still prefer a method that starts a discussion of probability by defining it. I think the long-run frequency definition is the one we should provide with an indication that it's not the only definition. It can also go directly into an example using simulation. Indeed, we focus on simulation/resampling methods of inference, where the mapping to long-run frequency is very direct. "Equally likely" doesn't really strike me as a definition since it has the word "likely" in it, which is another word for probable. I think of it as a tool for calculating probabilities when indeed the outcomes are equally likely.
2. How about defining empirical distribution and probability distribution instead of empirical histogram and probability histogram? The notion is broader than the choice of visualization and indeed the examples provided are bar charts, not histograms.

[shomurali]

1. I propose we discuss this in our next instructor chat. I know that "equally likely" has likely, but I still feel that it is a more natural definition. And then we can bring in the LRF definition.
2. Yes, I agree. I will make this change.

[shomurali]

For the next semester, move Conditional probability into the intro to chance, so we do both addition rule and multiplication rule.