Always enjoy your column, Mike. If everyone is so freaked out about CAW betting, why not just organize and beat them at their own game? Most of the algos used by syndicates are from the 2010-2012 era, while one forward thinking Korean operation uses algo 5-6 years old. That is the best I know. Forget all the Race Lens, etc. stuff - that is hobby stuff not professional wagering tools. Why not convince someone (like super quant Ariel Miller) to create cutting edge 2025 AI-ML algos that blow them out of the water. She'll never beat the sports books algos, but parimutuel tipping is like catching fish you put in the sink - easy.
Thanks for reading , Chuck. Not saying you’re wrong, but the easier thing would be for the tracks to just even the playing field. Rather than have weekend / Saratoga bettors like me building our own CAWs. But at some point, people with more money than me will do it!
Mike, in an ideal world, you’re absolutely right. But let’s look at it from a business standpoint.
The “hold”—or rake, if you prefer—is absurdly high in horse racing. And most tracks are mismanaged by C-level execs shackled by D-level regulation. To make matters worse, many tracks still act like it’s 1925, not 2025—trying to lure attendance with gimmicks like free hot dogs while the industry hemorrhages relevance.
States that allow terminal-based gaming, like Kentucky, can stabilize and grow. But in states that don’t, racing craters without handle. It’s just math. On some days, 70–90% of all betting volume comes from Computer Assisted Wagering (CAW).
Here’s the reality:
Let’s say the win pool for a race is $500,000 and the effective blended hold across all stakeholders is 14%. That’s $70,000 in revenue—split among track operators, purses, ADWs, and regulators. The remaining $430,000 is returned to winning tickets proportionally. In exotic pools, that hold can stretch to 25–30%.
Now imagine racing decides to eliminate CAW. Handle drops to $100,000 or less. Now the take is just $14,000. Multiply that by 8–10 races per card, every card, every week, and it becomes clear: there’s no way the industry can just "say no" to CAW. It’s not just dominant—it’s what’s keeping racing afloat.
It’s no different than equipment in sports. If you try to play golf today with clubs from 100 years ago—or even 15—you’re giving up an edge. But if you design a new club in 2025, with today's tech, and it's better than both, you don’t just compete—you redefine the game.
So, what’s the play?
Build the next-generation tool. Create a public-facing CAW-style model, community-powered, subscription-based, with adjustable factor weighting and open transparency. Let smart players crowdsource signal, beat the sharps at their own pivot points, and return the edge to the individual. Make the math accessible.
Mike / Chuck — I see AI and my name getting tossed around here. Love the topic.
Yes, building a proper toolset for racing using modern AI is absolutely achievable—especially from a machine learning standpoint. And no, I’m not talking about the legacy “AI” behind Equibase’s neural network dashboards or the Race Lens summary widgets. That’s hobby-level stuff—built on outdated frameworks and shallow inference.
The reality is simple: to disrupt CAW, you don’t need perfect predictions—you just need to outperform what’s currently being deployed. And from what I understand, the models powering most CAW operations today are built on computational architectures that are sluggish and outdated by modern standards. Just look at Ray Kurzweil’s analysis of the exponential jump in computational capacity, in his novel The Singularity is Nearer. We’re not playing with the same tools anymore.
Now, as for CAW: I agree—it should be banned. But it won’t be because of the economics: CAW provides liquidity, guarantees takeout, and props up failing tracks. No regulator or track operator is going to voluntarily walk away from that revenue stream.
I agree, we should level the playing field. Then we give players access to the most accurate possible data—wrapped in a flexible interface that allows for variable weightings, personal strategy overlays, and real-time adjustment. When you combine correct data with diversified decision-making, you get what racing desperately needs: a fairer, more dynamic market. Healthy markets need both accurate information and participant variance. Right now, CAW teams have all the signal and none of the diversity. Flip that—and we might just save the game. The good news is that the data feeds to drive this are CHEAP and it can occur outside centralized racing authority!
This has motivated me to drop a very detailed Substack article on how this is achieved for anyone interested.
Happy Father's Day my friend. I agree with you; Breslow is one to go between him and Cora, although I have been in the minority with my dislike of Cora. These call-ups and send-downs, all falls on the CBO. I do not think he will last past this year if we don't make the playoffs, or at least finish better than 81-81. Roman (which is my son's name lol) looks great, Marcelo makes it look easy. Bright spots. Not sure if those are Bloom's leftovers or not but. I hope they don't mess up their development like a pitcher or two we have lol.
ONE million percent agree on the statue and this Anthony kid has to be a STAR to get people interested in the sox. Enjoy Polar Park tonight. Go Wamps (my kid's hometown) and Rockets (my hometown).
I can’t believe Finn still gets paid to regurgitate press releases about new shows into actual ink-and-paper. But I guess if you’re reading the newspaper you don’t know any better.
Chad Finn is a thin skinned media suck up. Sorry to be harsh, mea culpa, but man that was the puffiest of puff pieces. I watched the game, KMS played with so much heart. I’d say start shooting some extra shots outside. I couldn’t make one trip up court myself, knee replacement, but shooting is contagious and I thought you passed up some open looks brother.
Always enjoy your column, Mike. If everyone is so freaked out about CAW betting, why not just organize and beat them at their own game? Most of the algos used by syndicates are from the 2010-2012 era, while one forward thinking Korean operation uses algo 5-6 years old. That is the best I know. Forget all the Race Lens, etc. stuff - that is hobby stuff not professional wagering tools. Why not convince someone (like super quant Ariel Miller) to create cutting edge 2025 AI-ML algos that blow them out of the water. She'll never beat the sports books algos, but parimutuel tipping is like catching fish you put in the sink - easy.
Thanks for reading , Chuck. Not saying you’re wrong, but the easier thing would be for the tracks to just even the playing field. Rather than have weekend / Saratoga bettors like me building our own CAWs. But at some point, people with more money than me will do it!
Mike, in an ideal world, you’re absolutely right. But let’s look at it from a business standpoint.
The “hold”—or rake, if you prefer—is absurdly high in horse racing. And most tracks are mismanaged by C-level execs shackled by D-level regulation. To make matters worse, many tracks still act like it’s 1925, not 2025—trying to lure attendance with gimmicks like free hot dogs while the industry hemorrhages relevance.
States that allow terminal-based gaming, like Kentucky, can stabilize and grow. But in states that don’t, racing craters without handle. It’s just math. On some days, 70–90% of all betting volume comes from Computer Assisted Wagering (CAW).
Here’s the reality:
Let’s say the win pool for a race is $500,000 and the effective blended hold across all stakeholders is 14%. That’s $70,000 in revenue—split among track operators, purses, ADWs, and regulators. The remaining $430,000 is returned to winning tickets proportionally. In exotic pools, that hold can stretch to 25–30%.
Now imagine racing decides to eliminate CAW. Handle drops to $100,000 or less. Now the take is just $14,000. Multiply that by 8–10 races per card, every card, every week, and it becomes clear: there’s no way the industry can just "say no" to CAW. It’s not just dominant—it’s what’s keeping racing afloat.
It’s no different than equipment in sports. If you try to play golf today with clubs from 100 years ago—or even 15—you’re giving up an edge. But if you design a new club in 2025, with today's tech, and it's better than both, you don’t just compete—you redefine the game.
So, what’s the play?
Build the next-generation tool. Create a public-facing CAW-style model, community-powered, subscription-based, with adjustable factor weighting and open transparency. Let smart players crowdsource signal, beat the sharps at their own pivot points, and return the edge to the individual. Make the math accessible.
Mike / Chuck — I see AI and my name getting tossed around here. Love the topic.
Yes, building a proper toolset for racing using modern AI is absolutely achievable—especially from a machine learning standpoint. And no, I’m not talking about the legacy “AI” behind Equibase’s neural network dashboards or the Race Lens summary widgets. That’s hobby-level stuff—built on outdated frameworks and shallow inference.
The reality is simple: to disrupt CAW, you don’t need perfect predictions—you just need to outperform what’s currently being deployed. And from what I understand, the models powering most CAW operations today are built on computational architectures that are sluggish and outdated by modern standards. Just look at Ray Kurzweil’s analysis of the exponential jump in computational capacity, in his novel The Singularity is Nearer. We’re not playing with the same tools anymore.
Now, as for CAW: I agree—it should be banned. But it won’t be because of the economics: CAW provides liquidity, guarantees takeout, and props up failing tracks. No regulator or track operator is going to voluntarily walk away from that revenue stream.
I agree, we should level the playing field. Then we give players access to the most accurate possible data—wrapped in a flexible interface that allows for variable weightings, personal strategy overlays, and real-time adjustment. When you combine correct data with diversified decision-making, you get what racing desperately needs: a fairer, more dynamic market. Healthy markets need both accurate information and participant variance. Right now, CAW teams have all the signal and none of the diversity. Flip that—and we might just save the game. The good news is that the data feeds to drive this are CHEAP and it can occur outside centralized racing authority!
This has motivated me to drop a very detailed Substack article on how this is achieved for anyone interested.
Happy Father's Day my friend. I agree with you; Breslow is one to go between him and Cora, although I have been in the minority with my dislike of Cora. These call-ups and send-downs, all falls on the CBO. I do not think he will last past this year if we don't make the playoffs, or at least finish better than 81-81. Roman (which is my son's name lol) looks great, Marcelo makes it look easy. Bright spots. Not sure if those are Bloom's leftovers or not but. I hope they don't mess up their development like a pitcher or two we have lol.
Good looks
ONE million percent agree on the statue and this Anthony kid has to be a STAR to get people interested in the sox. Enjoy Polar Park tonight. Go Wamps (my kid's hometown) and Rockets (my hometown).
As always DTH, thanks for jumping in the comments my man
And boooo Wamps
I can’t believe Finn still gets paid to regurgitate press releases about new shows into actual ink-and-paper. But I guess if you’re reading the newspaper you don’t know any better.
Yeah, I get features, part of the gig. But there's some bias in this one that felt worth mentioning, Dan.
Chad Finn is a thin skinned media suck up. Sorry to be harsh, mea culpa, but man that was the puffiest of puff pieces. I watched the game, KMS played with so much heart. I’d say start shooting some extra shots outside. I couldn’t make one trip up court myself, knee replacement, but shooting is contagious and I thought you passed up some open looks brother.
Have you applied for the UConn job yet?