How Complex do Models Need to be in 2026 & PM News Breaks Contain | Ep 153

Nerd alert. The main topic is a discussion on deterministic v. probabilistic models in the current betting environment. News includes the solider who bet on himself to capture a foreign leader and made $440k before getting arrested and a rant on the people complaining about FandDuels injury insurance.

0:00 Deterministic v. Stochastic Sports Models

26:00 News

1:05:15 Q&A

Welcome to The Risk Takers Podcast, hosted by professional sports bettor John Shilling (GoldenPants13) and SportsProjections. This podcast is the best betting education available - PERIOD. And it's free - please share and subscribe if you like it.

Follow SportsProjections on Twitter: https://x.com/Sports__Proj
Follow GP on Twitter: https://x.com/goldenpants013

Transcript

SPEAKER_02 0:00

I would say the majority of people I know who I've tried to like model or whatnot, I think most problems, like probably 95% of the surface area on a sports book, I think could be attacked um deterministically, and it would be fine if you did like a good job of it and you could win doing that.

SPEAKER_01 0:37

GPSP. This is uh, you know, they say like one for you, one for me. This is for us, but it's for all of us. It's very nerdy. Um we're we're doing a main topic that was, I think, like best phrased in a question by Chon Figgins, which was Does deterministic modeling have any place in sports betting and its current landscape, especially for major sports, or is probabilistic slash stochastic modeling the money ball moment for this industry? So besides these being a lot of words that I didn't know like three years ago, what else you know makes this uh this uh interesting question, SP? Let's say if we have to hold people right now into the episode?

SPEAKER_02 1:29

Yeah, I mean, I think so so I I I guess let's get maybe some like definitions out of the web of like to make sure we're we're talking about similar things. Um so like when I when I read this question, basically, like in in plain English, the way I would I would you know rephrase this would be like, does a model that give you gives you a single number, is that still um useful, helpful, something you can win with versus something that's like more simulation based or the process, you know, if it's not simulation directly, it's the process giving you like the full range of outcomes and it's not um not range of outcomes sort of like slapped on after the fact. Because I think that's you know, when people think model, I think the first step they would go to would be like what is framed as a deterministic model here, meaning I want to project, I want to bet on rebounds. How many rebounds is this guy gonna have? I'd project him for 7.2 rebounds. Like that, that's what I would call like a deterministic model. Whereas um the probabilistic, stochastic, the output would be more like a whole, you know, like a you know, unif uh like the standard distribution, whether it's a bell curve or some other distribution, but you have sort of like the whole range of outcomes and and sort of like the model produces an output where you have probabilities at each, you know, six rebounds, seven rebounds, eight rebounds, and it's not something you're doing after the fact. So that's that's at least like how I'm interpreting the question. I don't know if you you you're interpreting it a similar way.

SPEAKER_01 3:04

Yeah, I think said like a different way. If you re-ran a deterministic model, you get the exact same thing. Um, but if you re-ran a stochastic or probabilistic or sim, like you know, I think when I we were talking about should we do this question, you were like, are you reading it as like sim vers non-sim? And I was like, yeah, basically. But if you rerun a sim, like you'll get a slightly different probability each time. Like you should at least. So yeah, I think like the find a median or find a mean, and then like have a distribution after would be deterministic because you would I think under most iterations of that framework, you would always end up with the same thing, like if you kept running it over and over. Um and yeah, so that that's that's how I read it.

SPEAKER_02 4:02

Okay, so I think we're we're aligned with with definitions. Hopefully that gives you know anybody listening like an idea of what we're actually talking about here. So, like going back to the question, does a deterministic model have any place in sports betting in the current landscape? Um, so the the lens I would think this through is like what what are you trying to accomplish with like a probabilistic output or stochastic model, if that's like the direction you want to go? Like, what are you actually gaining? I think this is this conversation is gonna have a decent amount of overlap. Um, to I can't even remember if it's the distribution episode or the simulation episode, one of those, you'll have a decent amount of overlap. But like to me, the big benefit, like before even deciding like what's better, can I win with one, all of that? It's like understanding some of the fundamentals of what you gain, or like what are the benefits and and of having maybe the more complicated model. So the more complicated would be like, you know, precise, you have the full distribution, that's what you're getting outputted. Um, the the obvious con of that is it's more challenging to do. So that's that's probably you know obvious for everybody. That's um, you know, there's more opportunity to screw up. That's probably like the pieces I talked about about like simulations in prior episodes. Um but what do you gain? So like that's the main main thing. So I think in my head, what you gain from these um these more range of outcome sort of models is you capture certain things that like a deterministic model wouldn't capture. But if you're if what you're trying to solve for or calculate, you don't need that, then it's in my opinion, it's a waste of time. So I guess like the short answer is does a deterministic modeling have a place? I absolutely think it does. And so like giving real basic examples, like if you're betting, and some of these, you know, people might get on me about, and there's nuance in it, but like if you're betting median projections in general, I don't think you're gonna gain a ton. Um, it depends on the sport, the problem, I get it that all. But in general, if you're betting on like median um, you know, strike prices or lines or whatever, like a sim in general is gonna be far less useful than if you're betting on something more complicated. So like I'll just do two compare and contrast, like at extremes. If I'm if I'm trying to model like win probability, like a money line, in general, like a a probabilistic model is going to be less important, in my opinion, than if you're trying to price like an SGP or something that's very, very complicated. So like that's that's the basic framework I would think it through, like first and foremost, is like, and you can extend that. So like alt lines are more susceptible to like getting the distributions really right. And that that's where a probabilistic output could help. Um, whereas like the median, if you're just betting like over under seven and a half rebounds minus 110, a deterministic model can work fine. If you're betting over under 18 and a half rebounds, like and you just are gonna get your mean dis mean output and like slap Poisson on there, like that's not gonna work. Um, you're not gonna get a good price on that. So that's that's where it's like the what you're actually trying to bet, I think, is like the first lens I would think it through.

SPEAKER_01 7:32

The old it depends. Um yeah, I mean, obviously it it depends, not even joking. Um, yeah, I okay, so I agree with everything you said. I think that maybe people may think I come may come on here and bang the drum super hard for sims or probabilistic um models, but I'm actually I'm I'm gonna throw a curveball. I'm gonna come up here, I'm banging the drum for deterministic models. Because I think it's it's obvious, like for reasons that SP said, like why a bit, you know, a sim is is really good, and you know, if you're gonna like take this all the way, eventually you'll probably want to have some simulation framework. Uh SGPs is a great you know use case of a sim. I'll get to kind of what I think is the other big uh benefit of the stochastic model for 2026 after I bang the drum a little bit for deterministic, but I just want to point out like I think like every model is deterministic in sports betting in some extent. So even if you have a sim, it like I was thinking a lot about what Louis said, and I should have DM'd him before this, but he wouldn't answer, as we all know, for a while. So but he was saying his sim, he was like, you know, the one thing that makes my sim a little different is I don't actually input minutes projections for players. So why that's like important is that every sim that I know of has some kind of deterministic input, power ratings, you know, a player's like for our golf sim, like a player's what X skill, skill in whatever. And like like let's say we rate a player, the simple, simple strokes gain thing, like they're ending strokes gain. Every time we run the pipeline, if we don't change anything, that is gonna be exactly the same. Now the sim output might be slightly different, but like I think it's really hard to escape any deterministic modeling. Like I think it underpins most sims. So first of all, there's that. And then second of all, you know, I think SP nailed it on if you don't need it, like it's there's no reason to to use a sim on something you don't need because it takes a long time to build. And then it also like could be it's so much easier to get it wrong, right? Like there's so many areas that it could blow up the sim, and it's like really, really sensitive. Which a deterministic model, you have a little more, you can put your hand in, you know, you can put your hands in there a little more, tweak, tweak stuff, and um it's more interpretable, and it's you know, probably closer, it's it's probably not gonna wander as far away from like true probability as a sim, or like its range. It's it's kind of like anchored to you know market data and stuff like that. So, you know, I think that that's one case. Now the other thing is because I'm a big back test guy, so deterministic models are a lot easier in backtesting. What I mean by that is like backtesting already takes a lot of compute, and if you're running a sim, basically if your back test is like, you know, go back, have these historic lines, run your sim, grade, grade against those lines, you know, go to the next thing, like that takes infinitely longer than some kind of like good enough version that uses like a deterministic version. So it could be like you go back, you have your two teams power ratings, one team's uh yeah, I actually this is I I don't do team sports, but let's say like you have a basketball team and a team's like a 10 and another team's a two, then you have some kind of like table that says like, you know, if there's a two power rating difference or you know, whatever, this is the expected win probability. That's so much easier to use in a in a sim or sorry, in a back test to make that run faster. So like to have some understanding of deterministic modeling, I think it helps throughout the whole process, like between creating your inputs and basically like back testing stuff, because you know, I think back testing a full sim framework's quite hard. Um, so yeah, I think there's like a lot of really important use cases for it.

SPEAKER_02 12:32

Yeah, so I mean, I agree with all that. I mean, I think for any of these models to like win, whatever type it is, um it's the quality of the input, really, like whether that's data or like how you're interpreting the data or normalizing it to go into the model or what you're including. Like if if you're doing a really good job of that, neither of these, like the decision here is like not gonna, it would it would shock me if there were situations where the situation the decision here was make or break if you were like doing the setup of the problem the right way. Um and and so like the way I I just I jotted down a couple of examples. So if someone came to me and said, Hey, I have this full sim of an NFL game. Play, you know, at each play, it's totally simmed out, every aspect of the game. And then another person came to me and said, I have this model, it's in Excel, all it does is try and predict. Uh I'm gonna use a super super bolt prop example, the first time, the the time of the first timeout in the game. Okay. And so I have this full huge play-by-play sim, and I have just an Excel model trying to predict the time of the first timeout. If you told me like nothing else about um both those situations, I would take the Excel model almost certainly, because the the problem that like the person in the Excel is like narrowly focused on the thing they are trying to bet on. Uh, if we're trying to bet on that. Whereas like the person with the full sim is not narrowly focused on that at all, and likely, you know, has a million things in their sim to try and capture everything, right? They're trying to bet on, see, and this is the benefit, right? So it's not to poo-poo it, but like the benefit is you can bet on a ton of stuff. Like if you're well calibrated, you can bet on like, you know, all the player props, all the game derivatives, all the alts, all the this is why it's like valuable. But if you are like narrowing your scope to something specific you want to bet on, I would say nine times out of ten, the like a focused model is gonna work just because there's gonna be things in you know that guy's Excel file that I know are not gonna be in that sim. Like there's gonna be like specific coach tendency timeout stuff, right? Like in the Excel model, presumably, that is not gonna be in like the guy's sim, because it probably doesn't like matter in the holistic what he's doing. And so that's where like I think the the scope of the problem really matters. I know the question was sort of like on on um major markets, or especially for major sports, which which was in the in the question. I mean, I think so so like on small focus problems, I think deterministic in general is actually gonna do a better job. Um I think the the thing with like you know true major sports, why sims are more and more important, um, of like is because like the non this is gonna be sort of hard to explain over a podcast of what I'm trying to say here, but like uh the game state is like not independent from what like time period n to time period n plus one. So basically, like if you have a if you're trying to come up with money lines, like percent this team is gonna win. I think you're actually going, I said at the beginning, like it's not that meaningful. To the extent it would be meaningful, would be you're capturing certain game state things that are not that wouldn't be captured just with like a standard deterministic model. So like a silly example of this, or maybe maybe it's actually a good example, is there's certain like games like so for like NFL team, for example, like let's say uh Chiefs are playing a team that is like very strong opening possession. They they always choose to receive. The this other team, uh, I think Atlanta chose to receive all this this past year, always choose to receive. And let's say like Atlanta is like super, super good at um, you know, for like first possession. Like they score well above expected rate, like based on spread and total and stuff. There's like a, you know, there's the the whole like narrative around like, you know, Mahomes always comes back and when the game matters, like all this stuff. There's like a uh a state dependency in the game, then that's going to like inform how the game is actually gonna play out. So like what happens in quarter one, impacts quarter two, impacts quarter three, that can yield a different actual like money line price or spread price or total price, if you actually sort of like play out the game path. And so that's that's like the one piece that I think um if we're being specific, like I think if you're trying to really model like um at the highest level, compete in the most liquid markets, like that stuff probably is important um in most sports. But I don't think most people are doing that. I know we have a question about major market modeling and stuff, so like I don't think most people are doing stuff like that. So, in general, like my advice would be would be towards deterministic and anything else, like unless you're trying to be like the true end boss of the sport. I I think for the most part, it's like not generally not worth the effort, in my opinion.

SPEAKER_01 17:51

I think it ties, yeah. Uh that's actually so it ties in with my like what I think right now is the the key use case of probabilistic, but it you're looking at that from a pre-game, but I think the obvious like extension is live or anything like aftertip. So in a sim, like SP said, game state is a part of a sim framework. So theoretically, if you have a sim, you should be able to like run the sim at halftime or after the first quarter, and then with the right data or data feeds, at least theoretically, we have a question about sport radar, but like you should be able to update your price as each thing happens in the game as time kind of runs down, um, as people score, as injury, you know, injuries or penalties or or whatever happen, right? So that's and I think the the other the reason why it's so important is because on prediction markets it's like 75-80% of volume trades live. So you know, in today's environment, I don't I don't think like I don't think there's ever been a time when a sim has been more valuable, but at the same time, like deterministic models are what underpin sims too. Like I you know, I think like a sim is almost like a a sim is almost like like this like machine and you feed it all you know, you feed it the fuel. Like it's okay, the sim's the car, and the deterministic models are like all the stuff that make the car go, basically. And you know, you need both. A lot of people know that gas makes cars go.

SPEAKER_02 19:52

It's very expensive these days, I hear. But yeah, yeah, no, I I think I think that's exactly right. I think with the the reason, like I said, the the the sims are helpful is like the scalability and being being able to plug into something very, very fast, like no matter what it is. If you have like a well-dialed sim, in theory, you can bet on anything, right? Um so yeah, I mean, it makes total sense to me live. Um so yeah, I mean with the prediction markets and how much emphasis is on live. Yeah, that I hadn't really even considered that, but that's that's a valid point. Um, certainly like because I I've dabbled before, I think I've talked about this like with um with certain um live models like way back back in the day. But even those were sort of deterministic in that they were sort of like feed in all the input to the game state and and return me a probability of you know this team winning at this point in time. Um I think that's challenging the challenging. The other thing that I think why it adds a ton of value for something like you, which we didn't really touch on, was like anytime you have interactions, it matters a lot more. So like when you're simming a golf tournament, there's all this interaction stuff going on with like if this guy does this and this guy does this, like how the course is playing matters and it interacts, it has impacts on multiple people. Um, who's ahead and like pressure impacts would have probably like impacts on multiple people. Um, and so it's like there's like a non-independent part. So, like that's all the, you know, I'm sort of glossing over it. This I'm basically using this question to just like, you know, promote uh the deterministic models because I feel like every maybe I in my head everybody understands the benefits of like the sims and whatnot, but like those are the benefits. Like it captures all these interactions well in in like types of of contests or sports where there's like like tournament setting type stuff, like

SPEAKER_01 22:00

You know, tennis, racing, golf, like all March madness, like like futures, whatever. Yeah.

SPEAKER_02 22:08

Right. Right. It it makes total sense to me to do that way. Um you know, like tennis, it makes sense to me to do that way. Um for a variety of reasons. Like I'm guessing like the, you know, I don't know a ton about tennis, and people sometimes yell at me when I talk about tennis. Um, but like, you know, the I'm guessing like people that the, you know, like the the percentage of outcome on a certain point probably matters a lot, whether it's like 40 zip or I don't know what it's even called, deuce or something, where it's like the high leverage. Yeah, yeah, yeah, yeah. Deuce. Yeah. So like I'm guessing there's like, you know, differences there where it's it's not just like you can't just like say this guy's gonna win, he wins like 60% of his points or whatever, and just go with that. Like that, I don't think that would that would work. So like these these like non-stationary uh processes, I think they're they're very good at because they will actually capture the range of outcomes well. Um much better than like, you know, like get the the mean outcome and then slap slap a distribution because almost always what that's gonna do is compress your distribution, like, you know, because you see, like okay, let's talk about like the BAM, the BAM at a bio game where he scored like 80 some points or whatever. Like, there's no way you could ever get that outcome in a deterministic model, right? There's no distribution you would ever get. Um, but if you had to sim it's like, okay, when the guy scores 50 points, his usage actually goes up to 100%, right? Then that would be like that could be then then it is in the range of outcomes, right? Right. Because there is this thing where if a guy scores, you know, 50 points, at that point he's probably shooting every single time, right? Which wouldn't be captured in like whatever distribution you're applying after the fact.

SPEAKER_01 23:59

Right, right, right. Yeah, that that that is a it's like what you said, like catching the tails, catching stuff that's not on the median. Um that that is what it's good for. And the golf thing is true because I was actually trying to think of like how I would even price an outright market without a sim. And it's like it's like a math problem that would be you know studied in graduate school.

SPEAKER_02 24:26

It just feels like some giant distributions, buddy.

SPEAKER_01 24:29

Yeah, exactly. Oh, and speaking of that, I was thinking of this like you could price SGP uh of like two legs, you know, kind of with a simple rule here or there, you know, you could have whatever their uh correlation is, maybe three legs, like sampling from historical data, but like so there are again like some spots where you could still get by with something a bit more deterministic for those types of things, but uh yeah, obviously once it becomes it's just like the golf thing or whatever, when it's it's just so many moving parts, like you just can't like a human just isn't gonna be able to kind of keep up with it.

SPEAKER_02 25:13

Yeah, no doubt. So there's sports and situations where I think it's it's almost unless you're getting very advanced like mathematics, it's it's almost it would basically be necessary. Um, but I think I would say the majority of people I know who I've tried to like model or whatnot, I think most problems, like probably 95% of the surface area on a sports book, I think could be attacked um deterministically, and it would be fine if you did like a good job of it and you could win doing that.

SPEAKER_01 25:47

I'm trying to save that as the intro in my mind 25 minutes. Can I write that down? That was quite good. Um should we hit the news? Yeah, do you have any questions? Because I feel like you kind of summed it up there nicely. Um okay, so FanDuel injury. Oh yeah, actually we have a we have a we have to make actually a uh what's that called where you like say you're sorry for misreporting the news.

SPEAKER_02 26:19

Yeah, yeah, I I don't know what it's called, but we we do need to make a correction. Correction, yeah, or something like that.

SPEAKER_01 26:25

So we said that uh fanatics invented that uh, you know, like pick your your lineup, you pick the players and you want them to combine for certain stats. Play it your way or something. Or is that how but anyway, FanDuel already has that and it's called your way. I think at FanDuel, I forget what it was called at Fanatics.

SPEAKER_02 26:45

I think that's right. We had a very angry FanDuel employee. No, just we had a we had a FanDuel employee inform us, you know. So I was over here, you know, celebrating fanatics. I mean, I I you know, this is what I said to him that FanDuel is clearly not marketing it well if if the you know the the beats on the ground don't don't even know it with between you and I. So that's right.

SPEAKER_01 27:05

That's right. The people who know so maybe people don't care as much as that.

SPEAKER_02 27:08

I I thought it was cool that.

SPEAKER_01 27:09

Yeah, I think it was cool too. I guess the the the takeaway was like maybe people, yeah, like maybe it wasn't as cool for Rex as we thought it was. But uh okay, so they are offering injury insurance that you opt into. Uh FanDuel. So basically how I understand this is that like you get to add injury insurance if you want to assert, you know, whatever to your to your bet, whatever your bet is, and it basically like lowers your odds of the bet, but if it you know so-and-so gets injured, you the bet voids. Is that how you read it?

SPEAKER_02 27:56

That's how I read it, and the the the key piece I was trying to do, you know, we had a lot of news items, so I only had a brief amount of time to to to do research on each one. But the one thing I thought I saw with this one is um it like applies on parlays, and so my understanding is it's three percent always, um, but it's three percent like no matter how many legs are in the parlay and everything, and it would apply to all the legs. So, like this was the interesting thing about what was that three percent on the payout. I wasn't clear on that, whether it was payout or the stake.

SPEAKER_01 28:38

We're not gonna research the news so we actually know it super well. I thought it was on what it's worth. Okay, oh, on stake. Okay, yeah, payouts. That could be more interesting. Okay.

SPEAKER_02 28:47

I mean, the thing that I thought was was most interesting about this is they, you know, FanDuel had later came out and said, like, uh, we we don't think we're gonna make any money on this. We we basically did this in a way to try and be like break even on it. I mean, I think in that type of analysis is really challenging to do just because you have no idea who and how often people are actually gonna opt into this, like how much anti-selection and like all of that. So it's it's really challenging. But I think the thing, the one thing people like don't realize or appreciating about this when they say, like, oh, FanDuel's just gonna be like making a killing on this, is first off, the parlay part. You know, like if it's 3% and like the average person who's doing this is probably also the same person who is making an eight-leg parlay every night with minus 300 legs.

SPEAKER_01 29:40

Like like LeBron's like LeBron to score 10 points.

SPEAKER_02 29:46

Yeah, that's that's I mean, like if you if you're on the exchanges, you see what people like to bet. Like they like to bet, you know, like the stuff that's like LeBron to score 10 points. Like that's that's what people want to bet, and they want to parlay that together. And so, like, I don't think it's the more there is of that, the less money you know, FanDuel's making on this. Because like a huge portion of the times bets like that lose, not maybe not a huge portion, but uh more way more of a portion that those bets lose than you know, like a standard 50-50 line or not a parlay, right? So I I don't think I honestly think they probably did like they're probably conservative in assumptions and everything, but I I honestly think they probably have numbers that are like not that far off um from like who's who's gonna actually buy this. Um, because there's like the Venn diagram from the people who are gonna buy this and who are betting like that is probably like just a single circle.

SPEAKER_01 30:45

Well, there's probably like three dots like outside that circle that are like Maple, Thomas, Stevens. Like it's just gonna be people who've like figured out why this is just the biggest free money glitch in a certain and like like you said, it would be something like that where like the probability of the bet losing has a large a large chunk of that probability is the player getting injured. I th I like I think that seems like the the first look as far as like how to game it a bit, since it's flat, like it's a flat um either stake or payout reduction. We we don't know. We'll get to you next episode on that. But uh, but yeah, if it's flat and it applies equally to all situations, like there's like in a case like that, there's almost always like something you can do.

SPEAKER_02 31:34

Yeah, right. Like, yeah, they try and be probably cons like reasonably conservative to make sure they cut it covers their bases. And you know, people like even casuals are are not you know dumb. Like if if their players like questionable, like they will be, I don't know if they'll do it optimally, but they'll be more likely to to use this, like when there's a chance like this guy gets you know ruled out. So I don't know, it'd be interesting to see. I mean, uh the people who complain about these types of things I really don't get at all.

SPEAKER_01 32:04

Um, like it's the mid, it's the middle of the curve.

SPEAKER_02 32:09

Like, God forbid, entertainment company does something to improve entertainment for entertainment customers. Like, like what are we complaining? Like, people want this, and they're like, it's like you know, me going to a restaurant and yelling and moaning about you know the restaurant upcharging the wine, you know, like what are we doing?

SPEAKER_01 32:30

Right, right. It's the wine, right? Exactly. The wine is the people want the expensive wine, yeah. Like that's yeah, obviously you could get it for cheaper at Trader Joe's.

SPEAKER_02 32:43

You're like imagine being that guy standing outside the wine bar being like, you know, holding up the sign and tweeting about you should be getting this wine at the at Trader Joe's.

SPEAKER_01 32:52

There are there are people like that. Um no, I yeah, I don't know. It's the same. I think we we've established an identity as like very, you know, against that specific group. And then they come up a lot uh in these debates. But like I said, I just feel like it's I don't know. It's there's just a certain amount of like pandering done by people who like learned how to win betting six months ago to like the recreational crowd that it just like drives me nuts. Like they're like try, like they see them as these like brainless children they need to protect. It's like maybe they have ulterior motives than you for why they're engaging with a gambling product.

SPEAKER_02 33:37

Well, that's exactly it, and that's the thing that that like I don't know if they don't get or like they do they can't they like are incapable of putting themselves in this brain, like the brain of those people, like those people do not care if they lose three percent more, they care when they get off work and like just want to enjoy the game, and you know, in the first quarter their bet is ruined and they have no entertainment. And like if they're willing to pay and they're happy to pay for that, like you know, it's I just I mean, like I could put myself in the sh in the head of that person, like it makes all the friends who just like to better own games and watch them for like five bucks.

SPEAKER_01 34:17

Like you know, just only twelve. What are you what are you losing EV? What's the what's the median better losing EV there? Three cents. Oh my god, send the cops to New York and arrest his fan duel executives.

SPEAKER_02 34:35

Right, and what does he gain? He gains like that, you know, one night. Yeah, he he doesn't have to like tot, you know, totally tilt his mind off when his guy gets injured, right? Like right. He feels smart when that happens, right? Like, yeah, he feels like a like, yeah, I don't, it's it's it's like clearly a plus EV utility thing for most casual betters, I think. Yes. Um, because like they're not gonna notice the 3%. They're they are gonna notice when they like, you know, whatever, once a month, once every couple months, where they just totally get dusted and can't even enjoy their their Saturday sweat or whatever.

SPEAKER_01 35:13

Yeah, I think I think we forget about this because it's not football season, but let's all remember in football season, like every week people are tweeting that they're gonna send like the FBI to FanDuel's door because they didn't like void some person who got hurt in the second. Like, maybe they just don't want to deal with that anymore. Like it's just maybe it's plus utility all around.

SPEAKER_02 35:36

Yeah, I don't know if though you know the people who are mad about this just think those people like the people who are tweeting those things are just gonna wake up one day and be like, yeah, actually, like I I can just bet the under. And that's like they're they're not gonna think that way. So like just give like this, this this is seems like a good middle ground for everybody.

SPEAKER_01 35:52

I think a good rule of thumb is like the person you're telling that you should they should bet on yours will never bet the under. Like, and then just remember that and and kind of move on with your life. Correct, correct. Um, okay. I realized that you were like, we should do this other one first, and I forgot. The one we're supposed to do first was, although that was good, I'm glad we did that one because I think that's more fun. But the uh I think the biggest news of the last couple weeks has been the soldier who got arrested for who was part of the the mission to I'm not gonna be able to pronounce the leader's name right, um, but basically to capture uh what was it, what's it called? It's Madura, right? Madura, Madura, okay. And he was part of the team that that captured him, and he placed some bet on Polly, um, Polly's offshore exchange. So the the blockchain won. And it was like, I forget the price, but basically it was like to win$440,000. Is that right?

SPEAKER_02 37:08

It was something like that, yeah. Something like that.

SPEAKER_01 37:10

Yeah, and then uh, you know, got caught, got arrested. Um, obviously, this is huge news. Um, you know, it obvious it completely broke contain. Uh as only the really bad stories do, unfortunately. But yeah, this this was obviously uh uh bigger news than just in our industry. Trump uh responded to it saying that he was not he said the prediction markets were like casinos and he's not uh a fan or not much of a fan of of them. Yeah, I don't know. It's it's a you know obviously this is a big story. I think I'll I'll just give my like quick take. My quick take is you know, this is like different. Like this is it's to me, this isn't a prediction market story. Like I understand that okay, actually I'm gonna I'm gonna slightly take that back. It it is because there's gonna be more opportunities for stuff like this to happen. But this is like just a a serious this is a crime, plus it's it's like I don't I don't know in the military like what this would exactly be considered, but like really bad, like like treason or or whatever, because basically you're like leaking in info um prior to an important operation, and like you're making that info public. So like I think when people make stuff public, albeit in different ways, like those people are treated like very harshly. And I I think like that's what this is more of. Um is just like more of a military kind of issue. But you know, prediction markets, there's no doubting that predict prediction markets create spots like this. Um, and you know, we're just a world of humans, and this will this will happen if you know the opportunities are there, and even with the right rules, like there'll still be stuff that that's you know, people will still try and you know sneak one by. You know, insider trading's never completely gone away, even though there's been a lot of like enforcement and and whatnot. It's just lessened. So, you know, not a good look. Um, not a good look.

SPEAKER_02 39:37

Yeah, I mean, honestly, you were going in it. I have two directions I wanted to go in this. You're going in one of the directions is like it makes me, I think like when prediction markets you talk about, like these are the types that people are like, this is what they're for. Like we're getting you know valuable information. But it hon it it it begs a question like this is giving information. Is it giving information that's valuable to like the people we want it to be valuable to? Because if I just think like in this case no, right? Well, yeah, so you know, just adversary, whatever. Just can I won't even use this example. Let's say there's an adversary, you know, to to our nation, and there's a market on will this person be in power in one year or something like that. Like, I don't think that like helps protect the American, like I don't know what that's contributing to the average American citizen. I think if anything, it's like intel to like foreign governments more than it is like to the United States. And it just is also, it obviously is like, you know, it's it I don't think it's fair to like I agree with basically everything you said about like I don't think it's like this is a crime, so it's like not fair to just like pin it on prediction markets. Um but it it begs the question like, are these actual markets that are like in our country's best interest, like this type of thing? Like I don't know what what information it's giving like the average American that is that is like beneficial, right?

SPEAKER_01 41:10

Um the other aspect that's neutral, right? Like absolute best case, like absolute best scenario, right?

SPEAKER_02 41:17

Yeah, yeah. Like I don't like I don't think Americans should really be keyed in on like what the probability of foreign adversaries are to be taken. Like maybe that's a hot take. Like I'm for like surveillance and government secrets. I don't really know, but like I don't actually think that's something that should be disclosed. Um so like if this is doing that, I'm not necessarily sure that's a good thing. Um, the other thing I wanted to say on this is like I I I it's great to see that they're like really um, I think in these first cases of this, it's really important for the prediction markets and like the branding of it and everything and sustainability to like throw the absolute book at these people um and like go over the top um to make it you know very clear that like this is bad and it will not be tolerated. Um, I think it's like I don't think because you know, certain people on the prediction market side are saying like this is the system working and stuff. I would push back a little bit on that because the first point I made, like, I don't think you can just like I don't think you could say it's it this is different than like uh it has different implications than a guy like throwing balls and strikes. Like when we catch someone who's like fixing baseball games, and after the fact we can be like, okay, the system's working, we'll get them out of the game. Like this has like real life implications. Like, obviously, like you could imagine an alternative scenario where like, you know, if if other you know, foreign governments are like tapped into this stuff, and you know, like getting information before the mission happens and people like are flying into their deaths, or like more I know this is somewhat extreme, but like you don't have to like jump that far to these things of like the mission not being successful or whatever, like that's not the same as throwing balls and strikes. So I don't think you can just be like, oh yeah, we caught them. This is fine. I think prediction markets should really um you know, work with the government of like, are these actually like beneficial? Who are they helping? What is the information we're getting from them? Do they pose more risk than they're worth? And like I it's not clear. I mean, I've only thought about this a little bit, it's not clear to me like this is like a slam dunk. Yes, we want these markets as a country.

SPEAKER_01 43:40

Would you say like these markets broadly is like foreign affairs markets? Like if you want to like when you're talking about this in general.

SPEAKER_02 43:50

Yeah. Yeah, I would say markets where uh we the the information that's gained is disproportionately helpful to non not our country's interests, right? Yeah, okay. Like I think that, you know, that's a decision, you know, like as a country, I feel like we would need to make is if if there's like if like why are we, you know, giving intel? Like we wouldn't just give intel to countries that are our adversaries, like through the government. So like I know this is like maybe not our government, but it's still our people like providing intel potentially to adversaries, potentially. So it's it's not clear to me whether that's something we want.

SPEAKER_01 44:36

Right. Right. Yeah that's a good point. Like you think about who's modeling this stuff or our uh three letter agencies and they're probably quite good at it. It's probably but it's there's probably not been a good public source until until now of something that could rival that and obviously surpass it if the people involved are smashing the buy button. Um yeah I don't know. I mean it it's just like this it's it's it's this weird thing where it's like sports are the best cleanest use of prediction markets. But prediction markets want to distance themselves from sports. And this is gonna be what happens in non-sports markets or a lot of them. A lot of the stuff that like is flashy and gets you press kind of also gets you press on the back end like this. That's like 10x bigger of a news story. So it's it's a really difficult walk. I mean you know obviously I'm biased I hope they lean into sports but uh yeah I think I mean hey it's the time to do it. We got the government we got the president's last name is the same last name as some guy on the board of both the uh both the exchanges so let's let's just have a sit down and you know you Donald Trump Donald Trump Jr. you guys sit down Luana and Shane and you just have this conversation.

SPEAKER_02 46:07

Say here we go we're gonna break it down would you attend that meeting I mean I would certainly attend that a meeting amazing content I was I mean just just even you know uh Trump talking about talking talking about this after the fact I thought was uh you know quite something it was like it was like he had no idea uh it just shows in my opinion like how disconnected you know like the different parts of the family are on this toxic yeah like it was very clear he knew like nothing about and it's it's so funny because so many people are like you know this administration's all in and meanwhile the president clearly has not gotten like the talking points at all like has not gotten a single one from uh you know commandeer sealag so it's it's uh it was oh yeah that that's the other thing it's like the DOJ brought the brought the hammer down not the CFTC which obviously like makes sense like this is a serious crime that's well outside the scope of just like securities but the CFTC is going to need to start doing something with insider trading that really needs to be the first step like not every situation is going to be like this where it's like a war crime and the DOJ is just going to like step in.

SPEAKER_01 47:31

You're gonna have so many like subtler versions of this that they're not gonna care about but somebody needs to do something about well the the funny one that I had on the news was did you see this senator who bet on him? Yes he tweeted it did you see his tweet being like I got caught like thank god or I did it on purpose.

SPEAKER_02 47:51

Yeah it was it was I mean that's that the the I had I think the takeaway there was um similar to what you said like it doesn't seem like the CFTC has any like actual I mean I know nothing about this but like it doesn't seem like they have any actual like because I look what happened to the guy supposedly he got suspended from Yalchi and had to pay a$500 fine um right like that I don't know like suspended from Calci is not enough right like right that that's clearly not enough for like actual I'm not saying for this case but like actual insider trading and um yeah I think they're gonna have to be like if I was in their shoes and like Calci's shoes, polymarket shoes, everybody involved like I would I would be trying to crack down on these people as hard and I think they're trying like they're clearly trying in the media to say like we we don't allow insider trading I'm glad to hear they've they've dropped the line of like yes what is insider trading maybe it's good like that has been dropped which is great it's a feature not a bug those people those people have you know have been abandoned. They lost all their money so thankfully they've dropped that but um yeah I think I think they just need to like really go over the top on this because I think I think we said this a show or two ago like this is going to be the thing that like when they're testifying in front of Congress and all this guard you know all that stuff it's gonna be like all about insider trading stuff is what they're gonna talk about for sure.

SPEAKER_01 49:23

And yeah clearly they realize that I feel like if anything they're probably like I mean you can't be mad at the CFTC if you're them because they've there's been no better thing for your whole business than the CFTC but they're probably like can you help us here with this like you know we like college what do they do college like you're fine$500 and the person's like I'm not going to send you$500 it's so I mean yeah I I don't know what that I so I don't know if they're fine with CFTC or Cal I know Cauchy's suspended but like yeah um I think it was Calci. But I mean that's the thing they don't what can they do? They the the the platforms only can do so much which is basically say you can't be here anymore. There's nothing else they can do right now.

SPEAKER_02 50:14

I mean Calci jail doesn't exist quite yet not like DraftKings jail not like DraftKings jail.

SPEAKER_01 50:21

Um okay let's talk about we'll just continue with Cauchi here so they there's been some I think the other like kind of big story it was more last week but there's a big uh class action lawsuit building where the states are kind of like coming together um and then I the the one of the items from the lawsuit was like a deleted Discord post from Tariq that said we also avoid anything that could be interpreted as gaming like sports as that is illegal under federal law. And then I they also had I I don't know if you saw this but Audi's post where he's like it's gambling it's obviously gambling or whatever which was like true and it made me like him I was like okay he's like a reasonable person and not completely like because at the early days like the people who were talking about this were so delusional and annoying like for Audi to be like it's obviously gambling I was like okay well at least I can like listen to this guy like like he's a reasonable person but like obviously that's now this is the the downside of that um but I think the Tariq one's quite interesting.

SPEAKER_02 51:35

Well what was your take what was your take do you so I had two two two questions I wanted to pose as related to this news topic. Okay first question do you think the way do you think the way that like this this these products have been marketed from sort of day one if they could go back and do it over would have any change on the like outcome or where we are today meaning like if they weren't so you know like the stuff like insider trading might be good they weren't so much like this is this isn't gambling or whatever they weren't so much like I think I think they're finding their footing on some of the messaging on some of these topics now. Some still is a work in progress. But do you think if they were like you know obviously it's easy to in retrospect go back but like do you think anything would change if we'd be in the same exact spot I don't think anything would change.

SPEAKER_01 52:30

I I actually think like if probably if I'm just being probabilistic about it like they're they've kind of gotten to a good spot like to get even here is pretty impressive so I probably wouldn't change anything. And the other thing is like yes there's obviously like the casino side is nitpicking these types of things but they were always gonna fight them no matter what because it's just like really it's it's just it's it's just money. Like like they're saying you're gonna take money out of our pockets we're gonna fight you. It's simple as that it's not anything deeper. So they would have found other things to talk about I think what do you think?

SPEAKER_02 53:13

I think I generally agree with you. I think you know the casinos like there's nothing they could have done or whatever like to to not have that be a fight I just I don't know how much it matters like the court of public opinion um but that's the part that I feel like could have been done differently. Yeah. You know just like the pop I know we don't even have it on the news like the John Oliver thing and whatnot like you know if you're if you're like a budding gambling industry like you know the first thing you should do is how do we stay off John Oliver's show for the next year? Like that's here's a question. Do you think that's even possible? No see that's that's the thing and it it's a good point you brought up like they've done an amazing job of marketing right so like you can't have one without the other I get that part. So I was just it's just curious a question. The other thing I had with this is what what is going on with the Kentucky Derby? You know they said I know this seems unrelated but they said we avoid anything that could be interpreted this was deleted Discord. Anything interpreted like gaming like sports okay so we've we've made the jump we now we don't think sports are like gaming where where is the Kentucky Derby on on Cal sheet well isn't that like horse racing has its own like crazy treatment by laws I think and like yeah yeah they there it's the only organization that that calci appears scared to anchor.

SPEAKER_01 54:35

Yeah I I I don't know yeah I I've always like horse racing in my mind has always been this like like kind of really mysterious like other thing there's like casino gaming there's like betting and sports betting and like horse racing is sports betting right it's like a race there's a human involved there's a horse involved the winner wins but like it's so clearly by our industry like defined as something else and we actually yeah we need to get a horse racing person on here that that that feels like like we got it I was really excited for for Kentucky Derby.

SPEAKER_02 55:15

I mean I'm guessing it's like a decision they're making that it's it's you know horse racing is in between sports and casino and it's they're saying it's too close to casino to do. Yeah um would be my guess of what's going on but I was really excited for you know Calchi Derby and I don't think they're doing it would have been great that would have been great.

SPEAKER_01 55:35

It would have yeah yeah yeah that would have been great we'll have to sell for the World Cup I guess. Let's do let's quickly go through kind of some of the rest here. Oh the sport radar thing. So Muddy Waters and actually another like short seller like at the same time put out like a short thesis on sport radar um because with the Muddy Waters thing I know they released like a documentary I didn't watch I don't know if you watched I did get a chance to watch okay okay so I'll I'll kick this over I'll I'll kick it over to you set us up with this one.

SPEAKER_02 56:11

Yeah I mean I I so to preface like I know nothing about this world of like short selling and like anything about it really um when I watched a documentary I mean like the the takeaway I had when I watched a little 15 minute video was like it felt like we're just investing with like our moral compass and like not our head is what it felt like to me. Like it felt like it was very like a judgment of like you know it's wrong like this company is like preying on gamblers and and um you know like in these gray markets and all this stuff and like saying they're like you know the FBI is sports bank but in these gray markets and stuff. And I was like okay all right this all makes sense. So now tell me like how that's gonna come crashing down on them. And then they're like video over and I'm like what is like like I don't in my opinion like you know the year 2026 is not the year to be like this company is um corrupt and they're gonna get taken down. Like I don't think this is the year to to be doing stuff like that. I don't there was no like there was no like as a as a naive person who knows nothing about it there was no clear like this is what's gonna happen for them to get cracked down. Like it was not clear to me at all like why they're actually going to start making less money.

SPEAKER_01 57:36

I mean the the best tweet about that then now that I like was big buck hunter he's like he's like wow this has made me really want to get long sport radar we had like all these like subcompanies they're in all these markets I'm like wow they're well diversified you know they're growing yeah yeah I mean I expected something to be like you know with the paradigm shift and you know comp countries cracking down on gambling like are this this happening in this country we think this is going to be like a you know a title wave to what's seen in other countries there's none of that it was just like yeah they're doing all this shady stuff we hope uh we hope it goes poorly for them from here on out right again I don't know anything about this so that could be bad read but that was like my read on it yeah that makes sense uh that whole like short selling like writing world is an interesting world but kind of off that like I only had one I had one take on it on Twitter which was I just posted a meme of like the Drake meme where it's like hey I'm a market maker on prediction markets and sport radar is like saying no and it's like hey I want to operate in illegal markets because that's what Muddy Water said they said we approached them and said we wanted to operate in illegal markets and they're like let's go I mean that's the other thing like I to me there was also like a lack of understanding of the industry a little bit and I obviously don't know about like you know they in the in the documentary they're talking about like you know markets in Asia and whatnot and I know very little about that.

SPEAKER_02 59:09

But like gambling is an industry that is largely in gray in the world and I don't necessarily view that as changing as changing for a long time. Like stuff will change in ebb and flow but like I don't see a hypothesis where you know gambling's existed in gray for a long time black or gray for a long time like it's not going to just like suddenly not be there in my opinion. Like the the people innovating in those markets are always going to be able to innovate faster than the people regulating. Just like look at the sweeps into the CFTC here and all this stuff like DFS like the the people who want to like offer betting products are just always going to move faster. So like there's always going to be customers for companies like this.

SPEAKER_01 59:59

The one thing that I've completely like become solidified in after doing this show and doing the news is like obviously the right if you are like trying to start a business and hit a home run like if I'm a VC like the thing you should do is basically like kind of break the law and then like have it become big and then have the law kind of embrace you a little bit. That's clearly the best path for any business. So if you're now this is not financial advice whatever whatever but to me that just seems clearly like what you have to do.

SPEAKER_02 1:00:33

That's like it's those are all the you gotta you gotta push the boundary in a way that you don't go to jail but like enough that like people are really angry with you. Yeah get the scale and then you know cash out and go from there.

SPEAKER_01 1:00:46

Yeah.

SPEAKER_02 1:00:47

Um there's the White House correspondence mention market was that the like people were upset with how it resolved because obviously there was a there was a uh shooting or attempted the assassination there um and did it all resolve no like was that why people are mad so I I uh I tried to check right before this I first off the the the Calci Discord was like I was having a um that place was something special on on Saturday night for any of you who who missed it. Uh sorry I shouldn't laugh but that's kind of funny no no we're just gonna focus on the Calci part of it part of it in the market. Okay so what I believe they did was they paused they paused it basically immediately they paused the market so no trading um to the best of my knowledge it is still pause and they're seeing if it's gonna get rescheduled within 14 days or something um and then they may resume it otherwise I think my understanding is it'll be um I think it'll be settled on last price the thing that I wanted to talk about here is like all the drama I saw was around like event does not so anybody who's not familiar with with every mention market there's a strike that is event does not qualify. So like all the strikes are will they say this thing and then there's one that's event does not qualify. And I actually think this was the you know often can be critical of Calci. I actually think this is the right decision. Like you're gonna make a bunch of customers angry by by pausing this market and whatnot. But you're very likely to stay out of the news if you do this and you just pause it and just make your customers angry. So I think it was the right decision. What I don't understand though is you just got to get rid of like the event does not qualify strike then. Because like this is like if it's not rescheduled like to me this is clearly the event does not qualify.

SPEAKER_01 1:02:50

I don't know what else what yeah how else would that happen? How else would it happen?

SPEAKER_02 1:02:54

I mean I I it wasn't they're not many times it's like the thing isn't televised. Yeah okay yeah yeah sometimes it's like a press thing and the thing's not televised or whatever but I I would just get rid of that strike. It seems like that's just gonna cause problems and and just go to last price if something doesn't doesn't actually air um because like I think it was the right decision to to pause this whatever all of that but like this I would be really angry if I traded a ton on event does not qualify and then they they go on last price or something. Because like clearly this to me is event does not qualify.

SPEAKER_01 1:03:31

Right. And like the when it doesn't qualify it's kind of like the injury protection it's like it's not qualifying for like some reason that's kind of fucked up a lot of the time.

SPEAKER_02 1:03:44

It's like the death clause like right don't like you just don't allow this because this is like how people would profit off of bad thing happening.

SPEAKER_01 1:03:54

Right. Right. Oh and Call She had its first block trade and good news for them. Um I think you know we should we should hit the cues but you know clearly they're trying to have some kind of OTC business.

SPEAKER_02 1:04:11

Uh we'll see I might be interested in it we'll see I don't know we'll more see the I just want to see the full you know 20 million whatever block trade just hit the order book.

SPEAKER_01 1:04:23

I don't want it to go you know through yeah through these private parties like yeah yeah yeah well you know an interesting there might be an interesting opportunity out there for you uh crushers in a single sport because if these financial institutions like I don't think they understand fully the how bad adverse selection is in in sports there could be some big block trade opportunities um potentially there uh in certain spots but maybe maybe maybe not I mean like there if there's there's a there's a way to get 20 million down on something like it's there so who knows be clever about it I guess um all right let's hit the cues yep jack Jack Heevey how would you rate the effort slash return ratio of expanding to different PMs versus expanding markets on just call she all right I'll give a a quick quick one I'll kick it over to you because I feel like you've been kicking this around um you know huge effort to go to a new one in my opinion like this is coming from a stance of like you're uh on you're trading in an automated way so you're using the API learning a new API like updating your software to interact with that all of is is is a huge huge hill to climb. Uh obviously if you're hand trading like you should be on as many as you can have money on as many as you can Um, because you're gonna be more like price shopping, opportunity shopping, you know, etc. But when you're trading in an automated way, I actually think it's the flip. Like you really need a good reason to hop over to a new platform because it's gonna take a ton of time, and there's probably if the platform is like Call Shi, where it's like kind of the place to be right now, there's probably a lot of um extra stuff that you can add on there using a lot of the code you've already written. So the reason has to be a pretty big one. And then to just cap this off, like the poly incentive program, people are like, oh, they're throwing all this money away. Like, I actually think it's really smart because what they did with that program, the poly US, like inside they have a lot of incentives for market makers right now. Like it's public on their website, uh, liquidity incentives. What they did was they basically like solved this pain point, they paid people to overcome this pain point and onboard onto their platform. So that was the the first time we kind of went to a different, like added a platform was was because of that. So I think it has to be for a really good reason. Um the poly incentive programs are pretty good reasons, so you know that could be something you might want to think about. But yeah, I'm curious to hear your thoughts on this one, SP.

SPEAKER_02 1:07:23

Yeah, that was the first thing I'd written down. Like, if someone's giving away money, that seems like a pretty uh pretty damn good reason to go, to go um do it. I I mean I think the way I'm thinking about it is just a balance of like, you know, I think with the prediction markets and the number of markets and like rec flow and everything, the way I try and think about it is each I want to be trying to like get my money through as many times as possible, get as much in play as possible in any given day, whatever week. And so if I can do that in a way that you know has similar ROIs or whatever on multiple sites, and I can just do it on one, like get get all the money into one, I think that's clearly better, less to manage. Um, I think you just need to do a risk sort of effort if you know, like how much it depends like how good you are at this, how much time, all these things. But like if you can double your ROI going from you know Cal sheet to polymarket, you have to sort of do like, okay, I'm doubling that, how much money is that over what amount of time and how much effort is that um to do? Um, I think the other thing is like there's benefits of having like a lot of money on one exchange to just like not to deal with like some of the stuff that's like capital intensive or whatever, and you just need a lot of money. And so from that perspective, there's also probably benefits of just being like solely focused beyond just like you know, you only have one set of code to worry about, things can only break in one place. There's also just like having enough money in one place is valuable. So I mean that's that's sort of how I'm I'm thinking about it, but um it does seem like it's getting increasingly attractive, like it's it's only moving in one direction, I would say, of like more and more attractive to go try on something like polymarket.

SPEAKER_01 1:09:21

Yeah, like it's kind of my favorite, my favorite quote right now is like having a lot of money is a big edge in prediction markets right now. It's like just full stop. If you just have a lot of money, that's like an edge. So, you know, that for reasons that SP said, like that means you're gonna be able to do the capital-intensive stuff uh on exchange because you don't have margin and everything is fully collateralized and all this stuff. So um the other thing which you said, which I think is really important, is like there has to be a reason, like a very specific reason that you're going over that isn't like, oh, I just want to be diversified, or like I want to have you know my money sp split up because like I don't want to get rolled by call she or whatnot. Like I think those are bad reasons. Good reasons are like this really specific thing that's going to make me more money than having my money on call sheet, I think. Kind of uh let's hit Thon. It feels like a ton of Alphon PMs is currently on the operational slash tech side. Do you see products coming to help bridge that gap? Or is the market big enough? And if those products come, what skill set would become most important to PM's success? Thoughts?

SPEAKER_02 1:10:47

Yeah, so uh and Thon DM'd me um to clarify a little bit. So I think, and this is where my head went with the question and what he clarified. I think a lot of what he's referring to in the products is like I don't know if you've seen this. I've seen I feel like I've seen like four or five products that like are basically helping you deal with RFQs on the prediction markets.

SPEAKER_01 1:11:09

Like I actually haven't seen that. Is that okay? I don't know if it's hard to be publicly available.

SPEAKER_02 1:11:17

I don't know if publicly available. It was it's more like uh the type of thing based on my read of it, is like inquire and we'll set up a call and see if like we're a good fit. Like, so I guess publicly available, but not just like the way odds jam is publicly available, and you just like I mean, I think that that's how I'm thinking about it. I mean, like a lot of the the alpha in like the RFQ stuff, if we're really just gonna focus on that to me, is like just being able to get it right like right now, is just being able to get it to work and like work fast enough and get enough, like it's just like there's a big burden there right now. So if that becomes like not a burden, um and like it's like the same progression as like something like odds jam or whatever. Like if you had your own odds screens and whatnot before odds jam, like you probably absolutely printed. Yeah.

SPEAKER_01 1:12:08

Oh my god, that probably was the biggest edge in like sports bagging history.

SPEAKER_02 1:12:12

Right. But like after odds jam, those things close faster, your accounts get limited like very quickly, it's just like way more competitive. I think that's true with like any of the products. I do think the market's big enough, like not enough. There's not obviously a lot of people that like want to book RFQs, but there's enough people that want to do it at enough scale that there will be products for this. Um, they'll probably be pretty high priced, and people will use them for sure. Um, so what skills become the most important? I think two things that I jotted down. One being able to punish people when you know where their vulnerabilities are. So this is true of like any product, but like if you know like there's eight bots that are hooked up to this software that's just let's say it's all it's doing is quoting like fanDuel SGPs or whatever or DraftKings or whatever, like then you you and you can punish that, right? Like it even if you have an SGP edge on some of those sports books, like you can definitely punish them, but like there's limits and all these things, and like just like you can really punish people on the exchanges if if like everybody's hooked up to the same bot. So I think that'll be one thing. There was this really interesting story, we don't probably have time to talk about it, but like um, I can't remember who published it, but like the the story of hockey, like hockey uh combo getting punished for like mispricing like nine and a half goals or something. Um so I think looking for areas like because the other thing that comes to mind with all this is like if everybody's like pooling from DraftKings prices or FanDuel, it's like those sites don't let you parlay everything. So trying to do that on Calchi and see what price you get could be interesting. Um, because that's like the second part of what I think the value is gonna be in is competing in the hard problems. So if you're if you're able, like if everybody has this like cookie-cutter solution that's just like grabbing the sports books and posting that price, like eventually that's like not gonna be good enough to get order flow, win, whatever. Um but if you can compete on the hard problems, like the things that those sites don't let you parlay together, or like the example that comes to mind is I don't remember if it was the Oscars or whatever, but they allowed combos on it. Right. They didn't announce it, but they allowed combos. Like whoever was was booking those combos was probably absolutely printing. Um for sure. Because like they were competing with nobody, they got like all the recreational flow. So like that that would be what I would think um would be like the actual like specific to RFQ skills is like competing at the hard problems in-game, live, stuff that you can't just sign up for a service and and post.

SPEAKER_01 1:15:02

Yeah. Yeah, I I I agree with I agree with the second part being like the thing that this I'm the most interested in if I'm like not already doing RFQs. I I don't think there's like a real market for somebody to be like, here's you know, here's how you hook up to like we're gonna hook you up to like DraftKings RFQ, you know, we're gonna hook you up to DraftKings SGP, and now you can quote RFQs, like you're just gonna get killed. And that's not to say you can't make money um quoting DraftKings, but you need to have full control over your software because all of it now comes down to like the risk management side and like all of these other problems that are super important. Um, and it's not the hard part isn't like quoting DraftKings. That's that's not the hard part. It's hard, you know, it's not like you just click a button, but like that that's not the thing. But the thing could be like having your own opinion, like SP said, on like a weather plus sports plus Oscar plus whatever, like and being able to actually like take your opinion and um quote accept, you know, whatever RFQs. Uh that to me seems like kind of the thing here. And yeah, I think I agree with everything you said on like what would then be important.

SPEAKER_02 1:16:32

So I can read the next one to you um from Calvin. Let's say I have a niche. Like I want to read this to you. This one's hard.

SPEAKER_01 1:16:39

Okay, go.

SPEAKER_02 1:16:40

Let's say I have a niche illiquid bet that I expect expect to close at minus 145. Right now, DraftKings has minus 110, uh, best price, second best price in market is exchange minus 120 with enough liquidity to get my entire stake down. I've previously tried listing this type of bet myself, offering the best price in market, but I never get filled on size because it's too niche. Do I click the DraftKings minus 110 on the minus 110 or the exchange at minus 120? How do I evaluate which bets are worth clicking on accounts with limited lifespans when worst prices are available on exchanges? And is there an actual formulaic way to answer this question for myself in the future?

SPEAKER_01 1:17:22

Um so I thought about this in terms of shippers. Uh when I was thinking about this, I was thinking about shippers like uh like each bet should increase the value of the account or something along those lines, right? So I was thinking, okay, what's the basically you have to have some like you have to have some like account value burn at clicking the minus 110 bet. Now that's like a very complicated problem that you probably like to be in the nature of this episode, like might need a sim for or something like that. But uh it's it's it's very complicated, like how much that you also have to have like insane domain knowledge about account value, which like I do not have. Uh and then you decide, okay, like is it worth it or not? I think the question of like which should I click first, I feel like you have to click DraftKings first. Um because I I think it doesn't really matter, like you're still gonna get CLV on the bet. And if you click the exchange first, like somebody could be set up to like hit DraftKings on the exchange moving, right? So like I think in terms of order, like if you are gonna click DraftKings, you might as well just click it first and then click the exchange. You could try clicking DraftKings and not clicking the exchange. If you that would be the only way you wouldn't necessarily get CLV, but it seems like if you expect it to close minus 145, you're sure it's gonna get CLV. So, like, if you are clicking DraftKings, like click it first, click the exchange. I feel like that's what Calvin would do here anyway. Um, but yeah, with the limited lifespan, I mean, you know, parlay it. I don't I don't know. The simple answer is always parlay it.

SPEAKER_02 1:19:16

Yeah, I don't know if I have like a clear formula, um, but like here, I think trying to put like an example to what you said. Um, I think like I'll do this in a very simple way, and you could obviously make this like more complex, but like just to get the the way I would sort of think through this is let's say you, you know, you have a brand new DraftKings account, and I'm just gonna use round numbers for this. Let's say you have DraftKings account, and on average, you think you can get 10K out of a DraftKings account just for a nice round number. Um, before like that's the threshold when you hit it, no matter like what your CLV, what you're betting, you're just like auto-limited. Um if you're gonna bet this bet and generate like$500 of EV, that's like on one side of the equation. So your EV on this bet. The other side of the equation is um what what you're doing to the value of that account. So in a perfect world, you're making a bet and increasing the value of the account, right? Like the 10,000 is actually going up. Like that would be doing something very square looking, but also generating EV. That's obviously hard to do. So let's say like a better case scenario is like break like break even to your account value, meaning this play does not look like particularly bad if it were to be reviewed. Obviously, most bets, most winning bets, unless you're doing you know something creative or you know, uh priming or whatever, like most bets are gonna like if someone looked at it, it would probably hurt the value of your account. So in general, you just want like the EV to be worth more than the cost to your like account, right? So like if I gave us and you could simplify this. So let's say, let's say I'm I I'm gonna make this bet,$500 worth of EV, but I think 10% of the time I make this bet, my account is just limited immediately. Like that would be a really bad decision to make that bet then, right? If you think, because you're effectively like 10% of the time, you're losing basically like your the EV of the account goes down to like$9,000 because you're losing the account 10% of the time. The bet's only generating you$500 an EV. So that's like a bad situation. Obviously, it's like a spectrum, like it's usually not one bet that takes you to zero, but that's like a a framework, you can think about it, is like percent this account or like this bet gets me limited and how much EV am I generating. And that's where like the parlay thing and all this stun uh stuff comes in, where when you cross like that 10k or whatever, you want to cross it a lot, like by a lot. So that's like the framework I would I would think through um for something like this.

SPEAKER_01 1:22:03

Yeah, I like that. I was like, it's like replacement value of the account. Like what like I kind of think of it like what would happen if you just went in, no cover and bet your stuff. Like how much would that account basically be worth? And then and then you take that and apply SP's framework of like what will this bet generate in EV and like what's the likelihood it nukes my account? Yeah, I like that.

SPEAKER_02 1:22:32

Yeah, the other part that you mentioned, which is really important there, is like the cost of acquiring new account. If you have like, you know, if you if you extend my formula framework to the cost to acquire account is zero, then like the decision is easy. You just bet into DraftKings because you can you have an infinite supply of DraftKings accounts. That's not the case for most people, and so that that's where that part would also come in. Like cost, and it's obviously like not necessarily a dollar thing, but like effort, time, resources, hassle, like the cost to actually get more accounts.

SPEAKER_01 1:23:09

Yeah, I agree. Uh Nick White. So I had a friend in high school named Nick White, who is still the smartest person I've ever met, and he's coming in here. It's probably him because this is a great question, a great point. Nick says, What can we do to get Callshi to move to a tiered tick-sized ladder? 50 ticks between five and ten cents is pure ridiculousness. Absolutely correct. Also, do you think there's a need for the top end to have a different ladder to the bottom end? 10 ticks between zero and one make sense, but 10 between 99 and 100 don't. Um yeah, that the second question I know I don't know if you could do that, like the 99 to 100 being different than the zero to one, but like absolutely, absolutely. I I mean nobody's gonna be surprised that I'm gonna bang the drum hard here. But the five, the the 50 ticks between five and ten cents is crazy. It's crazy. Um just flat out that just plain and simple. The penny jumping, I mean, I was gonna bring this up um at some point. I think maybe I talked about this, but I think one of the hilarious things that like I I said to you, I messaged you, I was like, penny jumping's hilarious because it was like, okay, once we kind of like modified some stuff to like deal with the nick tech environment, it's just like just be the most annoying person ever, like show small size and just penny jump, and like that's what you're supposed to do. Like, that's GTO. So, like, but it's funny, like I do kind of think it's hilarious when there's like a 50,000 share order, but like we have like our little thousand share order just always like jumping and removing and just you know like dancing around. Like this is what I said would happen, and it's clearly like what's happening. Um, so yeah, I think like it'll I think I actually think it'll probably change. Um I think there should be tiers between uh at different price levels. I don't know if like 99 to 100 and 0 to 1 being different. I don't know if that could like how how would that even work? Because you have to book the no side.

SPEAKER_02 1:25:35

Yeah, yeah, I think they have to be symmetrical. Unless unless I'm missing something, Nick. I think I think they'd yeah whoever's taking the yes on one side, it you know, it has to be a no. They they need to add up to 100, right?

SPEAKER_01 1:25:47

So right, but uh yeah, I don't know. Theoretically, like obviously zero to one should have the most fine ticks. We've kind of gone over this, and then it should probably should tier from there. But like if you can only have two tick sizes, I'll probably do something like 97 plus at the DEC and then go to one cent after. If you could only pick two, but yeah, it it clearly it needs to change.

SPEAKER_02 1:26:14

Yeah, I think we've we've covered this topic. If if you want to get a groundswell movement, Nick, I'll I'll be outside with uh you know picket same picket signs and campaigning.

SPEAKER_01 1:26:24

So um okay, I'll ask you this one. Uh Troy Cuban, in the big 2026, how much of an originator's edge comes from factoring in data no one else has versus out analyzing the data everyone has. So I feel like this was one of your best uh blog posts back in the day, or articles on your site was on this topic.

SPEAKER_02 1:26:50

Yeah, my uh my long uh, you know, we were talking will will I or fluff have a longer writing career? It's it's still to be seen.

SPEAKER_01 1:26:58

Fluff is drawing absolutely dead.

SPEAKER_02 1:27:01

I think I have maybe four to five articles ever written.

SPEAKER_01 1:27:04

Like I said, drawing dead.

SPEAKER_02 1:27:06

Yeah, I I I wrote those and then said this is way too much work. I need to get on a podcast where no work needs to be given and just talk. Um so I think in the the in parentheses of the question is respective of liquidity of the market you're trying to originate, I think that's an important point. So I think like in more liquid markets, it's gonna be um obviously more challenging to get a data advantage uh relative to like a less liquid sport um or prop or whatever. Um, just because like more people are gonna be competing for the data. Also, there's just like more interest even outside of betting. Like if you just think of something like you know, football, like NFL football, there's tons of data, charting data, all this stuff like that um doesn't exist. For like a sport that doesn't draw as much interest and is not as liquid. And so I think that that's a big um big difference between the liquidity. It's hard to really estimate this. I mean, obviously, like I might have opinions on the sports that like I know more about. I have no idea on the sports I don't know that much about. And even on the sports I know about know about, it's hard to know like what other people are doing and how much is is data advantage versus um you know like sort of analyzing and interpreting the data advantage. But I imagine it's it's just like very different in um in different sports, different like what types of bets you're making. So it's it's hard to have a definitive statement across the board. Um yeah, so I I don't really have a definitive answer.

SPEAKER_01 1:28:51

Yeah, yeah. I mean my gut reaction was like it will always it's gonna trend toward the edges will get smaller and they'll get smaller because the advantage will probably broadly trend towards analyzing over data. Like the data edge is the best edge, I I think. Like I think we both agree, or it's the biggest edge. Um and yeah, I don't know, as you know, more data is collected and publicly available, like obviously that edge gets a little smaller, and it's hard to it's hard to really grow that like organically because there's only so many, you know, you're just gathering data on the stuff that happens in the sport. There's not gonna be more things that happen in the sport. The rules can change, whatever, but like there's not gonna be like five times more things that happen in the sport. What's gonna happen is like the public data is just gonna cover a larger percentage of like the things that are happening. So then I just can't see like how it'll shift towards data. It just feels like way more likely to start shifting towards analysis.

SPEAKER_02 1:30:03

The only way I could see it's shifting towards data is I think with like the way AI is and stuff, um, there's far less of a barrier to entry uh to like modeling reasonably well. Um, like just the technical parts of it. So like someone who has like really good ideas about the sport and how a sport works and what's important and what's not, if they like had no technical ability at all, they would have struggled probably to like build a model. Whereas I don't think that's the case anymore. Um it's not as much the case anymore. So I agree with like the general trend of like there's gonna be it's it's every buzzing year, it's harder and harder to get a data advantage. But I think you could make an argument that like a modeling edge, um it like the the it compresses like the best modelers and the worst modelers. Um that that gap probably gets like narrower and narrower each year, whereas like the best data and the worst data, it probably doesn't change really.

SPEAKER_01 1:31:13

Yeah, this is a good like the there's probably not been much you know bigger jumps in then the closing of that gap after like Claude got Yeah.

SPEAKER_02 1:31:23

When Claude writes all our models, we'd all gonna there's gonna be no modeling edge, it's all written by the same person, you know. It's it's the same thing, right?

SPEAKER_01 1:31:30

It's all Claude. Oh yeah, yeah, that's a good point. I mean both are comp I mean everything compresses, sadly. But then new games come out and then we go play those. Um Mr. Predict. Do you personally know a lot of pro batters who beat lines in totals today at a professional level on non-niche sports? I feel like it's mainly propslash steam chasing these days, and feel like a lot have lost their edge. Is it even worth trying? I do feel like after we had uh the last episode, which was like kind of a rarity, because it was like, oh, Justin like is only betting mains and winning. And we were all like, wow, that's so crazy. Like what a what a unicorn. But like it's true, you know, you know, it took us, I'm sure we've we've had people on who beat mains, and I'm you know, but that's that was the thing about that episode. That was the unique sauce of that episode. It's it's definitely rare. It's obvious it has to obviously be rarer than it was like five or ten years ago. Like, there's no doubt, it's way harder. Part of it is maybe self-selection, because like there's now a lot of money to be made in props and stuff. So it if you can, if you're not restricted on what you can get down, if you can bet a large enough percent of your bankroll on props, then there's not really like a reason to do mains, um, like from a monetary standpoint. So I think that's a big part of it, and I think that kind of ties in with why Justin was saying, like, you can beat mains, um, as in like they're not completely solved, but it's just like a lot of people are choosing to solve different problems, a lot of smart people. I think that's that's that's part of it. But yeah, certainly like don't hear a lot about it.

SPEAKER_02 1:33:20

Yeah, I would echo that. I mean, I think you know, the last question, is it even worth trying? The one thing that came through, and appreciate Justin came coming on last week. The one thing came through like when talking to him is he was like obsessed with basketball in a way that like I think you probably need to be. Like, I say that as a compliment, like in a way you probably need to be um to compete in these types of things. So, like if you're gonna like sort of like half half-ass your way like into one of these, it's it's probably better to go to a less competitive market. Um so you so yeah, I mean, I think there's a lot of reasons, like you said, like there's more money in other things than there ever has been. Um, maybe that's you know, the the ratio has changed as we sort of talked about with prediction markets and whatnot. Now you can get like a ton of action on who's gonna win. Um so maybe that'll that'll change back and more people will try this again. Um, I think the recreational models haven't necessarily like um incentivized that skill. But um, yeah, I don't know a ton, and I think it's really, really hard unless you are like elite ball knower or like god tier modeler. Um, if you don't think you can be like one of those, probably better suited doing something else.

SPEAKER_01 1:34:40

Fully agree. Um actually, you have the new questions. I got because we we're basically cutting some, not cutting some questions, but we have a select group of questions that were really good around risk, managing your risk and getting blown up and all the stuff in the PM space. That's gonna be our next full main topic. Um, so if you ask one of those questions, we're not ignoring you. They were good enough.

SPEAKER_02 1:35:10

We're gonna do the full thing. Troy Soggy Sausage, Josh Royale, Richard Lee, you know, a good cast of characters. We're not ignoring your questions. We will get to them. I thought they were all great and we could combine them into uh a main topic. So we'll do that in two weeks. So let's do Ian's next. Um, his first one, I think we already hit it was at what volume level does a PM become attractive? I think we sort of talked about how we think about that, um, like going to a new uh production market. The next question, I'll I'll just rapid fire through. He had five questions. When will the CFT CFTC approve more applicants? My opinion, it's very slowly. I think the incumbents, the people who have the lice, you know, licenses, have every incentive to make this very challenging and say, hold on, you know, like we're doing all this stuff to you know protect consumers and we got to make sure like we don't have, you know, the floodgates aren't open. Um we actually had you know one of the news items was even um C-Lig was out there talking about potentially raining in parlays and combos. So like that could be like the the start of trying to sort of like keep out these sports only, you know, DraftKings fan do whatever the sports only companies. So I think very slowly I'd be surprised if like the floodgates are just open. Um I'll get through them on, then you could you could comment on any you want here. Any less oh wait, this one's for you. Any lessons learned from golf outcomes or markets recently? Um minus Scotty demand is perfectly.

SPEAKER_01 1:36:40

I was gonna say like it's like it's just uh the Bayesian update on golf is like I'm even firmer in my belief in that there's no price that you will stop getting Scottish Scheffler flow. And sadly, he's so damn good that like even if you're he you he he's scary, he's scary, man. So uh besides that, no, I think a lot of like what we talked about in the Masters is still my most recent takeaways, um, especially because the Zeke Classic didn't really move the needle much, the team event.

SPEAKER_02 1:37:17

So two more. Um, the last one was recent mixed legal outcomes for CalShi. I think we sort of talked about it. I think the courts have been not that great for them lately, honestly. The last one, any pain from this new whale on Calci, one million profit last week, fear jet. I don't really necessarily keep up with the leaderboards. I did see like all these uh yeah, all these uh new accounts. I don't I think it had something to do with fear. Yeah, all the adjective animal. I think it had something to do with them changing like their their social policy or giving everybody a social account. I don't really know, but um no, I didn't see this. I'm guessing if he if he made that much, I'm guessing he was making, but I I have no idea. Like making it's yeah, seems like you would have a lot of slippage if you I don't know his order size.

SPEAKER_01 1:38:06

Yeah, it could be like either you are doing some crazy good automated SGP taking or making. But yeah, no, I don't I think like one one thing that is pretty clear is like I think a lot of the biggest traders or whatever, like don't they make money for a reason by not crossing each other, basically. Not that you know who you are or not, but usually like there's not a lot of blood spilt.

SPEAKER_02 1:38:33

Uh this guy's positions are public. Oh, really? So so he's uh he had a a yes position, his most recent position, a yes position for 2.8 million contracts on Atlanta in basketball. So it looks like he's just betting enormous on what it's like a syndicate or something. Either that or a huge whale. It's hard to say.

SPEAKER_01 1:39:02

Like what was there no no position? There's no no position to balance that yes at.

SPEAKER_02 1:39:06

He has positions public, it looks like about 12 positions, so he's betting enormous.

SPEAKER_01 1:39:13

Not not like this is gonna be like the seal door one uh in poker. This is great. Okay, I don't know. I didn't realize. Okay, I'll we'll do a uh we'll do a health check on Fear Jet next. We'll do a Fear Jet section um going forward. I like this. Hopefully he stays on the leaderboard.

SPEAKER_02 1:39:33

Yeah, we need him on there two more weeks, or at least I would mean at least hopefully he stays public.

SPEAKER_01 1:39:37

Yeah, exactly. Public.

SPEAKER_02 1:39:38

Because he's he's the only one I've ever seen.

SPEAKER_01 1:39:40

I mean, it's like one week, so that's what makes me a little flop used to have his positions up, but he didn't know he was doing them, I think. Well, there's a difference there.

SPEAKER_02 1:39:50

So I mean, maybe this guy doesn't know yet.

SPEAKER_01 1:39:52

Yeah, I was gonna say, nobody tell FearJet if you know.

SPEAKER_02 1:39:55

Okay. Fear Jet, if you want to come on the pod. Okay, next one from Unhinged Sports. What has been the most difficult um since becoming a market maker on exchanges? Learn much yet from the likes of limiting certain customers who are Peer Sharp using their info ethan. Uh, even so obviously, like, I mean, we can joke about the RFQ limiting and stuff, whatever, but like on the exchanges, you can't really um limit. But what I thought was interesting, like the way I would frame it and like how I think about it, is you want to the most challenging thing is like how to think about how to limit people indirectly and like limit the right type of people. Um so you know, I think a lot of this we'll talk about in our next solo show, which we want to do on like risk and everything, but like certain things that matter, like your uh size you post, the width, uptime, when you're up, um, like all these things are indirect levers to limit certain people and not take bets from certain people, and um or at least take like uh the right percentage of bets from people you don't want to be taking from. So I think that's like all um the most challenging thing. The other thing I'd say is just like making sure everything works and like going to bed with stuff running and being able to sleep at night is some of the difficulties. Uh like not wanting to wake up, you know, I'm sure it's how 2 a.m. is yeah, like you wake up and your account's at you know zero, your entire balance is you know out and late, and you're like, I don't think that should be happening, and then trying to fix that. So that would be the other part, but I would say in general, the how to take bets from people you want to take bets from is the hard part.

SPEAKER_01 1:41:41

Yeah, I I agree with both those. I think if I had to add something, I would say like it's shown me the difference in customer, like so there's a part of the there's like part of this question I was like using their info, like quotes Rufus or whatnot. Like I actually don't think I would be happy to lose money to Rufus in golf. Um, whereas like I wouldn't be happy to lose money to somebody who like front runs injury news by two seconds. Because there's like nothing I can do with that information. Whereas like if I'm getting a winning originator betting into me in golf, like it's probably worth it. Like, whatever I'm losing an EV, like if it just like sparks a question in my brain about like, oh, why would they want that? Like that just having that question like generated is probably worth the EV. So like I think when I'm now thinking about like are all sharps equally valuable or what like valuable, like valuable to me as a market maker, like it's definitely not. But the people who have opinions, like I'm actually happy to like lose a decent chunk of money to them.

SPEAKER_02 1:43:06

Yeah, like if you lost, you know, if I if I spent you his flip of me winning a golf bet on you, you'd be furious because you know something, something is it'd be like you know, you and I playing poker and not when something has gone very wrong.

SPEAKER_01 1:43:18

That's Scotty uh for soul par or whatever.

SPEAKER_02 1:43:22

I gotta get back to even after the master. So I don't know when the next majors is, but I'm just I'm just going all yes. Okay. All right. We got one from DB to settle the back test or no back test debate. I mean, I know this is you know a fan favorite. I I think we've said everything that needs to be said on the the back test or no back test. I mean, one of the things that that I think I've been thinking about a little bit is like I think when we did the episode, the back test, no back test was very geared towards like, you know, the model and like not betting on exchanges, um, like like traditional sportsbooks at the time we did that episode. I think it's really interesting to think about like trying to back test strategies on exchanges. Yeah. Um, and it's much harder. I think it it highlights some of the points I originally brought up. And obviously it wasn't what we were talking about at the time, but like some of the challenges with back tests, because you could do stuff like if I decreased my order size to 300 shares instead of 500, like what would that do? But unless you're like being extremely meticulous on data collection of like what that would have done to the queue position and all these things at the time, it's just like very challenging to know. So that that's some of the challenges with it. But I think it is an interesting problem. There's probably a lot you can actually learn from like actual financial traders in this respect.

SPEAKER_01 1:44:48

Yeah. Yeah, I was gonna say, like, I think a lot like I think a lot of their game is um historic order book data. And yeah, you know, that's obviously a way harder back test than betting your model into like a line of DraftKings, which is you're kind of past posting there, but I'll allow it of the back testing getting uh less important or moving moving towards something else. But uh no, I I'm pro I'm pro back test. I saw this, obviously, you know, that was our sides in in that episode, but like yeah, I I was thinking about this question. Um and like you could do a few different things to give you some systematic way to evaluate your model, and one of them is backtest in a betting market, the other is in sample, out of sample, and I don't really know. And the third is what you we're all gonna end up doing anyway, which is like just bet it and see what happens. But you know, I don't think there's like a lot of other ways and like of those ways, back testing is actually I feel like gives you like what I don't like about in-sample out-of-sample for sports, and maybe it's a skill issue, but it's like sports kind of change year to year, and it kinda it doesn't create like what I feel like is a perfect like in-sample, out of sample environment, whereas like backtesting, you're just gathering for the last however many years that you have data for, and there's no split. So I don't know. That that's my last backtesting pitch. I'll put it to bed.

SPEAKER_02 1:46:33

It won't be put to bed. Um, it will not be we can go to James's next one or James's next one. If sports got banned from prediction markets, what's next for you? I've actually thought about this. Yeah, I think about it was just like the news. I mean, like I want to see the you know, I can't imagine for you the prospect of like texting that unresponsive partner. Like, I'm guessing that makes your like skin crawl like a lot. Yeah.

unknown 1:47:00

Yeah.

SPEAKER_01 1:47:00

So I'm gonna go back to that.

SPEAKER_02 1:47:02

I don't know. I I'm guessing for a lot of people, um, like in your shoes, it'd feel a similar way, like that would be a rough, rough go back. Um I mean for me, I'm just like always jump into whatever the new thing is. So like I'm sure I can entertain myself with something.

SPEAKER_01 1:47:17

Um yeah, I think we would probably like as crazy as it sounds, like I think we'd probably go to the stay on PMs and go to the non-sports markets.

SPEAKER_02 1:47:27

Like, like we still have I think I would do the same thing. Yeah, I think I would like get really into like the mentions or something, obviously.

SPEAKER_01 1:47:33

Yeah, right, right. Like, like, you know, I like I'm thinking like the two options are leverage the sports stuff that we have and go back to traditional like betting syndicate style. And then the other option is you know, leverage the PM tech we have and kind of operate it similarly as is. And like I I like you said, I just can't go back. I cannot go back. So it's I'm gonna, it's like time to learn mentions, buddy, or whatever.

SPEAKER_02 1:48:05

Yeah. Well, let's hope it doesn't happen. Let's hope it doesn't happen. Um I can read the next you want to read it?

SPEAKER_01 1:48:12

No, no, I actually don't know which order you put these in.

SPEAKER_02 1:48:14

Oh, I I got the order. That's right. Okay, so for next one from A Lotto Pro. Would you rather have Edge with or without CLV?

SPEAKER_01 1:48:24

It's a trick, trick question. Without.

SPEAKER_02 1:48:27

Yeah, I think it I think you know, the way I the simplest way I'd say it is, you know, if you know it's an edge.

SPEAKER_01 1:48:32

If you know it's an edge, it's obviously assuming you know, more confident. Yeah, yeah. If you know for sure it's an edge, like it's definitely without.

SPEAKER_02 1:48:42

Yeah, like the sort of um, you know, without COV means you have a unique edge. Generally, the downside of that is you have nobody to agree with you. It's no evidence that people agree with you. So that's like the pro and the con. Um, but a unique edge is gonna be more valuable than a non-unique edge. So yeah, definitely without COV, assuming you know it's an edge.

SPEAKER_01 1:49:02

Agree.

SPEAKER_02 1:49:03

We'll do JAGS SXC. This one's a pretty specific question. I I don't think probably either of us are in the best position to answer, at least parts of it. Um live ARBing, not just ROI and techniques, but delve into limits and how much a person can expect to make roughly per partner in a decent state, excluding PMs. Is priming worth it? And then the question about like moving to Ontario to avoid tax. So I definitely can't speak at all about like um like how much a P2 is worth on specific accounts or anything these days, given most of my boom dude is like on prediction markets. Um the one thing I want to say on this question my hot take is is you know, I I know a lot of people like you know, in this AP world have to optimize every single aspect of their life, and one of those extends to where they live. And it's like people literally will go live somewhere for. gambling or their job, they'll you know say it that way. My my hot take is nobody should choose where they live for a job, whether it's gambling or otherwise. Like you should you should choose where you want to live because you want to live there and then and then figure out the job elsewhere. Like unless you know the job has to be so good. You have to love the job so much.

SPEAKER_01 1:50:20

It has to be like you're a professional athlete and you got drafted by that team and you've been working your whole life for this.

SPEAKER_02 1:50:27

Right. Because otherwise like I I just think you know where you live and is more important than like like you could probably find happiness doing a lot of different things. I guess you could for a lot of people it's probably living similarly um so maybe those people like it doesn't matter either way but I would yeah I I don't the people who like move to like random states where they would have no interest in living yeah so dumb I can't imagine.

SPEAKER_01 1:50:54

So dumb I don't I I'm actually curious if that's a hot take or not because to me I'm in full agreement. And that being said we both live in the same area and we both moved here and it's a sick area. So I guess like we're self-selecting in that sense but like um yeah all right I'll read the next one from Larry Love.

SPEAKER_02 1:51:17

How do you go about adding a new feature or variable to a model? Do you just add it and see if it improves accuracy whatever accuracy metric you're using and then check to see if there's improvement in backtesting. I told you we weren't done with backtesting. Is there some kind of exploratory data analysis done before you even get to the last two questions? So I think in terms of like your first two questions adding a feature I think for the most part yes I just sort of add I I have a hypothesis I think something will be beneficial. I add it I see does it improve the model and then I start betting basically off it I generally will do less I'll do some type of like model fit stuff but I generally don't do like a ton of backtesting but that's more so because I'm betting like a bunch of different things and I'm always on to the next thing. So it's just like a time thing for me. In terms of the last question is there some kind of exploratory data analysis I think the one thing that I usually will do is if I have a hypothesis about some variable or metric or whatever being important one thing I will try and do is like I'll have so I'll have some model it'll predict outcomes and then what I'll try and do and maybe this can be considered like sort of sort of a pseudo back test is like look at the residuals um by whatever metric I'm interested in. So like I'll give a really extreme example if I was trying to predict um college football team passing yards what I would do is I would have some model predict it and then I would say okay output my predictions and compare that to actuals that's sort of like my residual and which teams are am I consistently overpredicting which are my consistently under predicting and there's always going to be just like randomness like you can't always parse a signal from this. There's always going to be just like cases where you overpredict like that's how a distribution works. But like if I did that and I saw I'm like the the top five teams I'm overpredicting all run the triple option. I would be like oh I should probably add a triple option variable to my model because I'm doing a poor job. So you can do similar types of things of like residual versus potential variable you're adding is like a way I would do it. And so you can also just get ideas of what to add as you look at like what is what what is the commonality of what I'm consistently over under predicting.

SPEAKER_01 1:53:50

Yeah. Great I mean great points. The one thing I would add for like the first question is I like to go one by one in my features um back testing. It's slow in back testing it's annoying but like you really I really want to know like I want to isolate impact of a feature. So like you can be looking at a back test know it's gonna take however many hours and be like you know what I there's like five things I want to learn. I kind of want to just throw them all in I used to do that but then I would always get nothing from it. I'd be like well it changed a little bit or it ended or it'd be like yeah like it performed just as well. But it's like okay well I still think these matter so now I have to go figure out if they do so like going one one by one especially in a back test like one one feature change change keep going keep going store your results be able to compare it um I think that's that's the most important thing I would say in in back testing when adding a feature is yeah isolate it store it compare it um the yeah the exploratory data is analysis I think is a really key part like that's the fun part you gotta figure out a way to like pseudo check like if you're on the right track. That's like kind of what I'm always thinking about on this stage like am I on the right track? Is there potentially something there? Like I'm just trying like I'm basically trying to answer the question is there potentially something here so a lot of times it's like form some kind of like proxy quick regression to test like if there's some effect in in this you know do some like for example like one thing a lot of people do is if they want to know if something's kind of important maybe they'll they'll say like they'll take the value from one year and see if it um is predictive at all for the next year. Like does is like a third down passing like in 2025 predictive like above expectation or something I don't know I sound like I don't know football but like is it predictive to 2026? That's like a very simple way to like dip your toe in the does this matter and it there's so many iterations of that but I do love that that stage of it. I think that's a really underrated part of of modeling and a good a way to like be creative in this where it like saves you a lot of time um and like makes you a lot better is like being creative in this part. Spend a little time here to save a lot of time down the road.

SPEAKER_02 1:56:35

Agree. I I think the more you can just cross things off and get rid of them quickly like do that quickly or find the valuable things quickly the the the just faster you can improve. So this one I'm gonna ask towards you um this one is kind of related to the sports radar post and court siders question. In your experience what is the latency you experience between the event happening and the API feed having the data would be interested if you're willing to share some feed providers you use um do you consider locating your servers in regions closer to Cal She basically a bunch of questions on like latency and and all of that stuff. I'll turn it to you because I you know I don't want to you know blow up your spot but I I do know you've had some boots on the ground at some of this and been doing some been testing some research.

SPEAKER_01 1:57:23

So I've done a little research I've done a little research um in in this I found I found like broadly somewhere in the five to eight second range of delay in some sports that I've gone and and tested um so that was the first question uh feed providers are used by sport feed providers are they might not know they're feed providers for our for us so anybody's a feed provider if you're uh yes forceful enough right that's right anybody's a everybody's a feed provider um so we'll kind of go there uh sport radar yeah we talked about like sport radar they are there right now like their party line is they're not um they're not selling to prediction market market makers so I don't really like I want to use them I would love to be able to answer this uh this question and co-location we considered it I don't think we're at the point where it really matters um was there another one solid bang for your buck data providers that still make live trading practical I don't know like you know obviously like no I don't know of any that are like providing game state data that are worth it that aren't like sport radar but I I didn't look too hard.

SPEAKER_02 1:59:08

Okay we got two two more um I don't know fun is the right word but you know broader questions so the first one from the artist formerly known as Mr. Peanut better now peanut um he's sticking to would you rather what crazy he's sticking with this I I don't I don't understand we need an explanation um would you rather never be able to drive again or ride again this applies to all transportation bikes planes trains etc so my opinion this is one of the clearest ones we've got um I feel like it's gotta be I I now I hope we have different I feel like I would rather like I don't need to be able to drive yeah for sure okay all right yes I mean like I hate driving I'll be honest I hate it I don't like it I mean like getting an Uber is would be like sort of a pain in the it it depends what city you live in to a little bit of an extent um yeah New York is awesome but like with self-driving cars potentially on the horizon like well well well well but does that count does that really count as being driven if it's you in the driver's seat of a self-driving car I I don't know I mean that seems like an important clarification.

SPEAKER_01 2:00:24

Yeah because how I understood it's like you do have to call an Uber to like go to the restaurant that's like six minutes away or four minutes away or whatever.

SPEAKER_02 2:00:35

You're gonna order all your groceries yeah but like would you call an Uber and it'd be a self-driving car like that seems fine.

SPEAKER_01 2:00:43

Yeah yeah that's fine that's fine. But I think if you own a self-driving car and you get in it and like program your destination and then go there like I do think that that's probably against the spirit of the question. I would say so probably too um and who knows maybe self-driving I I mean they they have them in some places so I feel like they can't be that far away maybe that's it does like slightly limit where you live like you need to be able to have like Instacart you need to obviously have Uber um but beyond that like if you're only driving if you're driving like I mean unless you're like Tom Cruise you're not flying anymore like you like you could I guess you could have a pilot's license and yeah the flying thing is like you know like the total non-starter like right where you go you're not you're never going overseas I mean yeah because then you'd have to drive the boat like the cruise liner you're on or whatever like there's no there's no way to really without like five years of training to be able to like go to a different country. Yeah so that that seems like a a straightforward one this one um is a little bit more complicated uh potentially so this one says from big papa if you're in up to your neck in a barrel of shit and I threw a baseball at your head would you duck so I clarified with Big Papa in Twitter if you saw that I asked what was his like last year of competitive baseball and he said like T ball in like age eight who played hockey and he would be standing 15 paces away so I paced off 15 paces it's pretty close like I feel pretty confident that I would hit somebody in the head a good amount from 15 paces um that being said I think for Big Papa I might just stand there and and and gamble with it. Because like going into the shit there's like obviously that's really like not something you want to do just from a comfort standpoint but there's like you could get disease or or whatever there's there's some safety concerns there. There's someone like a prof an MLB pitcher like I'm in. I'm ducking the second their arm like crosses like to where they're definitely throwing it I'm under I'll pop right back up they're throwing fast it'll be past me quickly like there's no chance because I feel like you could really really get injured but I think like if you get hit what somebody who doesn't play baseball they're throwing it like 50 you know at most and if they're if they're really trying to push 50 it's super inaccurate. So I don't think I'm ever going to like really get hurt. Um so I'm I'm taking it. What about you?

SPEAKER_02 2:03:41

Yeah you bring up an interest I saw I did not think of the illness uh component of this I mean because in my head the the the way I was thinking through this is like the marginal cost of putting your head in when you're already in the internet like it doesn't seem that much worse in my opinion. Like that's worse yeah yeah yeah like you know that's where your mouth your nose eyes you know I I didn't I didn't I hadn't considered that part I need I'll need to reevaluate that but you know my line of thinking was the marginal cost is is not high it'd be over relatively quickly you could take a shower but if you like break your nose or an orbital bone that is much worse. Yeah so yeah it's unclear it's it's this one is a much harder one.

SPEAKER_01 2:04:32

Yeah oh yeah definitely um I it gives me pause that his name is Big Papa because like if it was like just like if this was like uh soggy sausage and I knew they only played teat ball like I'll I'm I'm there I'm gambling like the problem with big papa we like your question it's just but the problem with big papa is like he played hockey and like I'm just worried that like yes he's not gonna be accurate but he's gonna be fast and I actually am like I'm more sensitive to speed than accuracy like if I know you're pretty accurate but you're slow fine whatever let's say yeah I mean I think it's 60 miles an hour and it hits you even 10% of the time I'm probably going down. I'm probably going down 60 is like it's from that close yeah yeah yeah yeah I'll I don't think we're ever gonna be in this I'm probably gonna duck I'm probably gonna duck but yeah you're right I changed my mind um okay I had a quick Bayesian update this is a true Bayesian update because it builds on my last one so new information um Apple TV is the number one streaming service in terms of original programming it's just like the one thing it doesn't have is a show like Traders or like the secret millionaire or whatever like a really good like game show but besides that like all their like fictional programming it's so it's all good. It's all good.

SPEAKER_02 2:06:22

I mean you know I'm not gonna sit here and name everything but like just being active right now like not even like shrinking um friends and neighbors there's this other like good uh like thriller show which now I'm forgetting the name of but that that's currently active like everything Apple TV puts out is like basically watchable that used to be HBO HBO feels weird now I don't know like there's a lot of like documentaries I don't want to watch and like their drama shows it's not like it used to be like HBO used to be like the spot like you know obviously the Game of Thrones time but you had the wire uh you had Sopranos you know these I think HBO is getting a lot of a lot of um a lot of support for shows that were made 20 years ago Apple TV right now is the place that's putting out the best original content okay so I so Apple TV would be like the in the game person we want in the on the show you know not posting on the prior accomplishment exactly correct okay I I'll take your word for it I I I you know I obviously can't confirm or deny my update is on um the current state of of Twitter so Twitter has gone through like what has like captured the algorithm has obviously changed over time and like the for a long time there's like the threads and all this stuff the the vague posting that's going on these days we need to we need to punish this in the algorithm it it's it's so painful like you know the the pictures and be like what do you notice like we need to it it's it's so painful. My brain's not involved enough to like get past the key there's there's nothing to notice the whole point is they post this and then get people like all angry in the comments so it gets a bunch of comments and it gets boosted like I don't understand what the I mean I guess I sort of do like if you're Twitter you're just trying to optimize for people on the website to like sell ad dollars um I guess but like man is it a brutal scene if you go to the the the for you tab it's just no matter what topic it could be politics it could be sports it's whatever like it's all just like like you know we're all like a picture a picture of what it's something and it's like we're all thinking it aren't we and it's like no I have no idea what you're talking about like right enough with the vague posts. So that's that's fine.

SPEAKER_01 2:08:59

That's like me getting old you know no it's terrible it's it's it's objectively terrible um yeah so we'll use our sweat to to put an end to it right here. Um cool well thanks everybody if you didn't get your question answered it's because it's come you're you're you're the main topic for the next solo show. Uh thanks everybody for for listening and we'll see you on the next episode.