Circles Off - Ep 153 - Dennis Montoro (dennycaps1)

Podcast Release Date:
Podcast Link:
hagrinBPR Review Date: 2024-05-14 19:02:01
Rating:

Notes:
Bias - I know NOTHING about horses. Less than nothing. I'm not grading this one and as I type this not sure I'll even post my notes for this one, but we'll see.

4:00 - Backstory: Grew up on LI (I also grew up next to Belmont, but it didn't stick for me). Poker > betting MLB 10 years ago. Stint as a card counter (lived in Vegas for 15 years). Found (refound?) horses through horse racing tournaments (I think I remember Chris Andrews talking about this once that's my own knowledge of these tournaments) where you start with same bankroll & highest score wins. Success there leads to full time horse betting, some sports. 7:40 - 2020, 'demic happening, stumbles across on edge that ends up active until only just a few weeks ago was the Madden football simulations. 13:10 - Self-Taught Quant: How? "Always had a specific problem to figure out, just Googled / academic papers."

15:45 - Horse Racing Tourneys: NHC - National Horseplayers Championship. Different formats - "mythical bankroll format" and "live bankroll tournaments" - buy-in real money some base amount and bet real money into real parimutuel markets.

19:30 - Current Horse Landscape: Industry in decline. Market dominated by CAW - Computer Assisted Wagering teams (see - Bill Benter). Second tier smaller CAW teams, followed by pros/semi-pros. Industry doesn't do a great job marketing. No real track-to-track unity. "Wouldn't this be a great Gen Z betting activity?"

25:30 - Rebates: Takeout = just like vig. Rebates - incentivize volume to reduce the takeout to a more effective takeout. Offered to your large CAWs. Tied to some threshold of volume. Interesting point here where MLB betting and card counting were successful, it was the other stuff that caused the issues (keeping accts live, disguising play) meanwhile horse racing is saying "we'll pay you to bet" so the rebate system was attractive to make that switch. Takeout is ridiculously high vs say sports betting vig - typical win pool takeout 15-18%, exotics are 20+% taken right out of the pool so hard to beat without rebates. The rebates incentivize the volume which makes scaling this as an income stream possible. CAW teams negotiate the rebate % directly with the track based on xHandle. DCap gets his rebate offer after his first tourney win.

33:15 - Data: "Horse data might even be more data rich/robust than MLB data b/c so many things effect a horse race." Practice (commonly known as "workouts") times, pedigree, jockey, trainer, speed rating, class ratings, etc. Since there's so much data, how do you model this? DCap says 2 approaches - 1) your market makers . Ex - Speed rating - different types of speed ratings, one of more common is Beyer Speed Rating (Figure). Issue is horses with higher speed ratings might be overbet. Another approach - "a horse with these factors will typically be overbet".

39:00 - Historically Dec/Jan bad months for him (interesting tidbit, sort of like that dead month many MLB bettors have).

41:45 - Horse Bad Beats: Some of the worst are when the horse wins, but commits some sort of foul, subjective influence of the stewards. "Inquiry! Don't throw away your tickets!"

44:45 - Backtesting / CLV / Market Signal: Does backtesting and forward testing. "There's no CLV metric to know when you've got the best of it." Rob gets back to this idea of no CLV at 50:00, but putting those notes here. "You don't know your exact price due to parimutuel market." "But what you NEED to be able to do is PROJECT where the market is actually going to land." Ex - Horse is 5/1, you make your fair 3/1, but you project the market to close 4/1, so you bet the edge (ty for an easy to follow example, makes sense). Then, adds the complexity of now instead of closing near your projected, what if it closes 10/1? Good for you EV wise in traditional sports sense, but there's no "market signal to confirm that your model is correct". (Great line, probably as basic as basic can be for horse people, but a great explainer for someone like me). "Market signal in the parimutuel market 'in the last flash' is so important." (I'm guessing he means that the close is directionally correct to his projected/fair.)

47:50 - Measuring Luck: Gauge luck in finishes "not always true that the best horse wins the race". (So very similar to what Telemachus mentioned a few episodes ago or Rufus' xROI article I have referenced a few times now, but for horses.)

57:30 - "Your bet size is limited by how big/small the pool is." Bets through an ADW (Advance Deposit Wagering site) which sends that bet to the track, those bets get aggregated with the ontrack bets. Has to model how much money a typical race might see bet into it.

1:03:00 - Talks about "Tapeta" at Woodbine which is a surface type, which he says is a very misunderstood surface from traditional handicapping perspective. Rob asks about going to tracks to physically see the horses and I reminded myself I once ran in a race with horses (the Vermont 100) which was quite a weird experience especially during the night hours which was the last time I think I have been in the vicinity of a horse. 1:06:25 - Mentions they do have fixed odds in New Jersey (just started in Colorado).