A.I. (Artificial Intelligence) seems to be a term you can’t go two days without hearing right now, so what about in regard to betting the races? We talked to two very successful horseplayers who both believe that any significant A.I. in regard to helping one play the ponies is still many years away, but both have strong opinions on software programs that have already been affecting our pools for the past few decades already. Computer Assisted Wagering (CAW) programs are nothing new, but the programs do seem to be getting both more sophisticated and more plentiful. There’s two sides to every story, and we present them both here. By John Rallis.
A.I. (Artificial Intelligence) is a term you hear daily, and it’s already had an impact on your life, whether you realized it or not. But how has it, or how will it, affect horse racing? CAW programs (Computer Assisted Wagering) have had a big influence on our betting pools for decades already - whether you realized it or not - and here TROT looks at both A.I. and CAW programs, and speaks with successful horseplayers from both sides of the table, in an effort to help you see what’s already here, what might be to come, and what might be seen as good and bad about both. It’s an ever-changing world in which we live, and horse racing is smack-dab in the middle of it - whether you realized it or not.
The landscape of wagering as a whole has changed drastically in today’s world. From the days of placing bets through bookmakers via pen, paper and card sheets, to now having access to a multitude of betting outlets from the palm of your hands, things are far different than they used to be.
In society, we’re forced to adapt to new norms, and the same is applied to wagering; especially horse racing. The game and its objective remains the same, but just like almost anything in life, things will change over time. The same methods bettors used to apply in their handicapping years ago, might differ now.
The state of horse racing and its future is always a topic of discussion amongst various groups associated with the sport, and the betting handle is usually a major concern. Yet, despite all of the new competition we continue to face for the almighty dollar, the handle, which is arguably the most important component of our sport’s existence, has continued to remain stable, and even see an increase, over the last few years. A lot of this has to do with the involvement of computer assisted wagering (CAW), which are computer syndicate models that compile loads of data to formulate high-level gambling opinions. It’s so distinct in today’s game, that it’s hard not to notice its impact.
When we talk about evolution and the direction of where things are headed on this planet, we also hear a lot about Artificial Intelligence (A.I.). We already use A.I. on a daily basis, courtesy of virtual assistants such as Siri and Alexa, as a shortcut to learn a plethora of things. The question is, where does it fit in horse racing?
Of course, though it needs plenty of refining, we often hear companies talking about how they are going to incorporate A.I. into their future plans. Horse racing is no different. We have already seen it through Canada’s ADW [Advance Deposit Wagering], HPIbet/DarkHorse, giving out a horse’s projected win percentage via ‘Smart Picks’. Whether you agree with it or not, it’s likely just a sign of what’s to come.
Dana Parham, a self-made, multi-millionaire from betting horses (both breeds), and a variety of other business interests, is someone who has been able to have an incredible amount of success in a field that can be extremely difficult to turn a profit in.
In 1996, the Columbus, Ohio native, who had been betting for years prior and had aspirations of being a full-time gambler, created International Software Services, a model that was developed by a couple of associates that he met when he was betting overseas. In two years, his dream became a financially rewarding reality.
“There are a lot of quality handicappers out there,” states Dana. “The truth of the matter is, we are not smarter than computers when it comes to constructing wagers and calculating the expected value of those wagers. Given the data you feed the model, it will decipher the amount of bets, the combinations, and it will measure your edge. We had one rule: find positive expected value, and bet.
“The greatest component of a computer is that it does not operate on emotions, it is strictly data-driven. Bettors can talk themselves into a specific horse based on a variety of components, while the computers will provide you with wagers that they calculate are an edge. It’s a big difference. I mean, without a computer, how does one make a [projected] price on every exacta, trifecta and superfecta wager? Without a computer, you don’t have a clue about the price, or how to get it.
What started out as a team of two, grew to a team as big as twenty people. Dana always had a couple of his members dedicated to fine tuning the model, while the rest of his team were in charge of pushing upwards of $400 million (USD) per year through the windows.
After almost two decades of operating his model, Dana sold it back in 2016. When asked about the integration of artificial intelligence in horse racing, he thinks it’s a pipe dream… at least for now.
“I think these terms are thrown around without any definition or meaning,” he claims. “IBM worked on a computer program for 20 years before they could beat top chess champions using their machine learning, which is what A.I. used to be called. Google can’t even get the images right on what a Pope should look like using their own A.I., and nobody has more money than Google. Do we really think people are betting horse racing with Google funds?
“I don’t believe there’s any A.I., or anything close to it. If we look at all these great predictions of man, and where we’re going to be, then where’s our robo taxis that Elon Musk was preaching? We’re miles away from these promises that we’re told are right around the corner. Dick Tracy had a watch that he could make [phone] calls on; that was from a comic book from the 1940’s. Eighty years later, that talking watch became a reality. That’s how long the development of technology can take.”
What we are still seeing in horse racing today however, is a high volume of activity from various CAW models, which are similar to what Dana and his team created almost 30 years ago.
More and more, there has been an outcry amongst many horseplayers, about the odds of the horses that they’ve wagered on plummeting at the quarter pole - negatively impacting their return. Horses that are going to the gate at 3/1 or 5/2, are often closing at even money or less. Frustration levels are boiling, and understandably so, but Dana says in the grand scheme of things, that shouldn’t be the case.
“Personally, and I know this might not be a popular opinion, but I don’t understand why bettors get upset when they see their price points lessen by the time the gate rolls,” says Dana. “If you’re betting a horse at 3/1 and you see the price drop to 8/5, you should feel really good about your chances of winning that wager. Oftentimes, when that’s the case, it means you’re on the ‘right horse’, because that’s where the ‘smart money’ is going. So what’s the issue?”
Though Dana suggests that the confidence levels pertaining to a horse whose odds take a dip by the time the gate closes should be high, bettors are having a tough time wrapping their heads around the weight of it all, and most importantly, the accuracy of it.
“It’s very simple,” he states. “You have your model and you develop your factors. You know what your factors are worth, and then you have your parameters of what your Beyer rating is (for Thoroughbred purposes). Let’s say my model is willing to take $6 on a horse… and it’s currently paying $8. How much more do I have to bet to get it down to $6? Depending on the track and pool size, of course that number will vary.
“In addition to that, maybe my competition and their model will come up with the same price point as me. They also think the horse is worth $6… And they’re willing to bet it down as well. If we both bet it simultaneously moments before the gate, which is when most of the money pours in, that’s when you see the impact. With 60-70 percent of the money coming into the pools at the last minute, that’s what happens.”
Even so, there have been many prominent bettors who have vocally expressed their displeasure of CAW programs and the impact it’s had on their betting styles. Some of which, even, have proposed the idea that CAW bets should be prohibited from being placed past two minutes to post (MTP), which would allow horseplayers to have a better understanding what price point horses X,Y and Z will be as the gate rolls. Dana, however, doesn’t believe that should have to be the case.
“Look, in life we all have a choice,” says Dana. “If you’re constantly complaining about the late dip in odds, nobody is forcing you to play, honestly. It’s why computers take up the majority of the handle nowadays.
“As I mentioned earlier, it’s a positive if you happen to be landing on those horses, because your win expectancy will increase. In addition, there are so many other pools to tackle in order to try to maximize your profit margins. A lot of people don’t understand the object of the game. The object of the game is to find and construct bets that return more than they otherwise should. That’s it.”
One prominent horseplayer that we spoke to, who has been betting for a living since the early 80’s, has noticed his profit margins dwindle significantly over the last couple of years, as a result of a few different components. The timing and increments in which CAW programs send in their bets has been the biggest issue for him. He says that pari-mutuel wagering is no longer what he remembers it to be.
“In Canada, they created pari-mutuel wagering for a reason,” says the horseplayer. “The idea was that people wouldn’t have to go to bookmakers to get quotes on horses. The money goes into a pool, you see the odds, and then you make your decision. Right now, that isn’t really the case.
“Pari-mutuel wagering is about getting quotes, looking at quotes and then deciding which way you’re going to go with the data you have accumulated. In order to get a quote, you cannot have these computer syndicates sending in their batch-wagers seconds before post time, slashing the number we agreed upon betting. It really defeats the entire purpose of pari-mutuel wagering.”
Like many others who think that CAW programs should be limited from betting past 2 MTP, he says that the increments as to what they’re wagering is more of a concern. He says more needs to be done to protect horseplayers, especially here in Canada.
“Honestly, there’s a simple solution to it all,” he says. We just need to have them bet in increments and put a time-cap on it. The Canadian Pari-Mutuel Agency (CPMA) is the regulator of the pari-mutuel system. If it’s not going to be the CPMA, who’s going to look after Canadian horseplayers, then who will?
“In my opinion, I think the bets should be capped per transaction. I don’t care if they want to bet $1 million into the pool… go ahead, but you can’t do it all at once. You have to give the bettors who want to participate in pari-mutuel wagering a chance, so they can get an accurate quote. Otherwise, what exactly are we doing here? There’s nothing more sickening than betting $300 on a horse and having him flash to 1/9. You almost hope for them to lose in a way. It’s sickening.”
To put things into perspective, our horseplayer presents an analogy as to what it’s like being a punter in today’s game, and just how frustrating it can be.
“For example, say you’re driving around and your [gas] tank is nearing an end. So you’re looking for a gas station and you stop in at the ESSO around the corner from your house. You see the sign and it’s $1/litre. You pull up to the pump, and you put in 50 litres. After you’re done filling, you dig into your pocket to pull out a nice, crisp $50 bill. You walk in to pay your bill, and the cashier says, ‘That’ll be $75.’ You say ‘What?, the sign says it’s $1/litre, it should be $50’. The cashier says, ‘We haven’t changed the sign yet. It’s $1.50/litre now’.
“Well, the gas is in the car, so you really have no choice, you have to pay. The same with making a wager, you have no choice, your money is in. When you’re betting at 3/1 at the gate, and you see the odds of your horse flash to 3/5 after the quarter-pole, you can’t get your money back. When you’re making a wager, you’re buying something. You’re buying something at the wrong price because of this, which is why something needs to be done.”
What distinguishes horse racing from some other gambling opportunities, is that bettors have more of a choice - they’re not simply choosing between the Leafs and the Habs as to who might win. There is a list of horses to choose from in every race, with data to be presented for each of them, prices included, to some degree. Like any form of gambling, there are unknown variables that come with it, but at least horseplayers can make their own decisions in many ways.
“Having the right to make a decision on the information you’ve collected, that’s what this is all about,” says our horseplayer. “This is what distinguishes horse racing from casino betting and the slots. You have a past-performance program, you have patterns, data and price points to base your decisions on.
“Out of all the gambling you can do, horse racing is winnable, because you can create your own database and make your own decisions. When they offer you a casino bet, it’s completely out of control.”
Despite his quarrels over ‘batch-betting’ and the effect it’s had in horse racing’s gambling market, he shares the same sentiment with Dana regarding the integration of A.I. into the sport. He believes it’s still so far away.
“A.I. is just a buzzword - at least for now,” he states. “When you say A.I., people perk up nowadays.
“A.I. has not been perfected yet. The information that A.I. is using isn’t always concrete. A.I. scours the internet for information, a lot of which can be exaggerated. For example, if 10,000 people were putting out on the internet that a driver can’t win on a certain horse, A.I. will account for those opinions, which is the flaw in it. The accuracy pertaining to A.I. is statistics. Numbers are concrete - subject to change of course - but those will always be accurate, because the numbers are factual.
He’s adamant that in horse racing there is too much of a human element component that A.I. won’t be able to factor in. It’s because of this, he thinks that even if it makes its way into the sport on a more permanent basis, it will still consist of too many flaws.
“How will A.I. be able to know if a horse warms-up lame, what the weather conditions are like, or if a horse got a good training mile into them? How will they know that an essential equipment change has been made, or that a driver wants to take back to last or launch his horse off the gate for early positioning (especially if the lines indicate otherwise)… it just can’t. There is more of a human element to this game than any other betting game. A.I. won’t be able to account for that.”
He does, however, see the integration of A.I. as a shortcut for many players, with the data that is provided.
“Where it’s going to benefit the bettors is that it’ll speed things up from a handicapping standpoint,” he reasons. “It’ll speed things up when you’re looking at things such as class drops, prior success or failures versus specific horses, and other forms of data that are available. Look, I handicap races differently, but I do think it would provide a major shortcut for bettors.”
Though he has clearly expressed his gripe in regard to the way CAW program wagers are being put forth, he states that ‘unrealized takeout’ is something else that is also a big issue, and is another component that has negatively impacted horse players.
“The integration of cellphone use has hurt players,” he stresses.
It wasn’t so long ago that telephones - even public payphones - were illegal in both the paddock and the grandstand. Once again, times have changed.
“We already have takeout percentages on all wagers, some of which are higher than others. Now, because everyone has a phone in their pocket, there are people who are close to horses that can send out text chains to various groups who can really impact pools. This can negatively impact our price points even further.
“I call it ‘unrealized takeout’ because it’s something many horseplayers either don’t realize, or can’t account for. Now that our methods of communication are far more enhanced, anyone affiliated with a horse can instantly send out a message to small groups, about a horse they deem as live, or not, and so forth, which can affect the odds-board significantly. We as horseplayers are really up against it these days.”
At the end of the day, it’s important to acknowledge that CAW programs do, in fact, have a very positive impact on the sport’s handle. While some horseplayers may want some constraints in terms of how they push their money, we cannot downplay their contributions.
“Handle is imperative for the sport and the CAW programs have a positive impact in regard to that,” our horseplayer admits. “For the record, I’m not against them at all - the game needs them - I just would like for them to operate differently, that way we are all on a level playing field. This is a great sport and I’ve had success betting on it for many years. We need more people to see it, to understand why our sport, as well as pari-mutuel wagering, is so enticing. With the rise of sports betting, I’m just hoping our game doesn’t get lost in the shuffle. Our game is good enough to attract new customers and keep the ones that already exist, we just need a few things to change.”
