Bet the Process - Mailbag Episode 2025 Dec 18

Podcast Release Date: 2025-12-28
Podcast Link: https://open.spotify.com/episode/4fOJnVFgdgn0MlmiI9ccRI
hagrinBPR Review Date: 2025-12-24
Rating: 5.00

Notes:
Bias - I generally hate mailbag episodes, but I'll try and grab a highlight or two.

8:30 - College Playoff Games - JMU +21.5-22 (which won). MIA +3.5 (which also won). Not sure Jeff liked either selection. Ole Miss -16 (which also won). Rufus has been on an absolute heater all year on these podcast picks. The most interesting point here was around 14:00 where Rufus goes through (briefly) the difference in the bowl games variance / error in score differential. "A point spread of 10 in a bowl game doesn't correspond to the same moneyline as a 10 point spread would for a regular season game." "The key metric there is what the standard deviation of the cover margin."

15:40 - Mailbag Starts - Starts with some Tony Kornheiser stuff where tony first met Jeff and said something like "You look like you don't have a lot of money".

22:30 - Prediction Markets Offering Insurance to Real People - Rufus tells a joke Tom said about betting on individual health outcomes and Jeff's reaction says it all. Jeff thinks that this is likely why the CFTC approved these that these should provide the ability to hedge outcomes.

27:20 - Modelers (Originators) Not Talking About Betting 2Hs - Rufus rightly says "actually, I think many people do this". (Groovin I know did a ton of 2H stuff). I know in our soccer chat we talk about live bets all the time. Jeff rightly brings up that betting 2Hs is a pretty tough grind at times. Jeff "don't think this is really a modeling issue". Rufus talk about the difficulty in fully automating this process and the pitfalls and I would completely agree.

31:20 - Mechanics of Filling a Position - Rufus rightfully talks about not really wanting to say much. Varies by sport. Sometimes will use movers who can get a lot down. Sometimes using their own network. Sometimes using exchanges. Sometimes you want to exploit books that aggressively move numbers. Talks about switching up every so often, can't do the same thing over and over. "Very labor intensive."

34:20 - Novig Parlay Issue - Rufus brings up Captain Jack's summary being pretty good. Jeff raises the question of "how much should a company be swayed by public opinion?". Nothing new here.

38:10 - Non-Quantitative Approach to Betting - Rufus: Can be done, hard to scale. 40:00 - Best Syndicates vs HFT/Quant Firms Infrastructure Wise - Rufus brings up Bloom, Jeff correctly answers the question that while they are impressive, they aren't close to the major HFT/Quant firms.

45:45 - Post Mortem Process - Team meeting every Monday, but not really diving into things on a weekly basis. Wishes they were better at collecting performance data in an automated way (assuming this is KPI stuff, this is a big focus of mine to).

56:45 - Interesting that they went to NoVig after getting approached by a bunch of other prediction markets for sponsorships, but they preferred to be aligned with NoVig.