Bet the Process - Nate Silver

Podcast Release Date:
Podcast Link:
hagrinBPR Review Date: 2024-02-29 18:38:06
Rating: 7.25

Notes:
I know nothing about politics or political betting markets. I don't really know a ton about betting the NBA at a high level or playing poker at a high level. I am the wrong person to be reviewing this pod, I do the best I can.

5:45 - Interesting tidbit on Sloan - there isn't a lot of sports betting content panels. More mainstream business, AI, etc.

10:30 - Nate's origin story. I've heard it before, if you haven't it's quick.

12:45 - "You do not have 50 independent contests for president of the US you have highly correlated outcomes." Makes sense. 13:15 - economic components to the political models added, makes sense. Makes the argument not a lot of elections, not a lot of elections with good polling, sort of a small sample/data problem.

18:00 - "really smart people make really dumb points because of their political leanings". #shockSTUN. Goes on a 3 minute talk about how his forecasting is not aligned with the lack of truth seeking from the outside world.

25:45 - Nate posits that maybe his models lean to underconfidence vs the overconfidence you see almost everywhere else.

28:45 - "Differential Response Bias". Something about a certain segment of the population answering polls more often because a subsection of the populous was out more than another in 2020. No idea how valid this is, but interesting opinion.

31:30 - "Some finance and some sports betting people overrate how much they know about politics." Could have said "overrate how much they know about everything" Nate.

32:00 - "Building an election model has a high barrier of entry takes a couple months of work." This is really interesting because my super uninformed opinion would have been the opposite that there's enough to bootstrap your model pretty quickly. I've never built one, will never build one so I will remain willfully ignorant on this point.

34:45 - 538 criticism from "Gambling Twitter". 36:30 - the NBA model was using home court advantage data all the way back to 1977 (what?). He goes on this long bit about betting the NBA on a daily basis, injuries, effort, etc, but the most egregious of the model infographics was concerning ship probabilities, seems like we're not even addressing the worst of the concerns here.

45:00 - "Not saying GTO for live play is overrated but ..." - I'm out of my depth here, but it's always funny when statements use this phrasing/structure.