You just got four new horses in your barn and now you must spot them in races.
In the old days you would have to drag out a bound paper book to look up the conditions and then start scribbling.
Currently, most condition books are done “on paper” digitally by compiling information and presenting it to the users as a PDF file that must be downloaded to a device and then read thoroughly, whether relevant or not. Pages and pages of information (Aqueduct’s book is 31 pages. Even Fonner Park’s is 30 pages).
Have we advanced?
Adobe, Inc. developed the Portable Document File (known by its abbreviation PDF) in June of 1993. It revolutionized how to send documents.
Image from of vintage condition book on sale on Ebay.
We need to revolutionize the condition book.
What if all you had to do was type in your horse’s name and race options popped up for you? What if we matched the conditions to the horses and the horses to the conditions using algorithms? What if your eight-hour chore was reduced to one?
Let’s advance.
Horsemen are already using technology to track their stable, monitor and transfer veterinary records and keep safe watch over their barns.
The Jockey Club and HISA keep digital records on all horses registered.
The information is out there. We just need to integrate it.
The current condition book system is outdated, cumbersome for the Racing Secretary and users, and can take eight (8) hours to fill eight (8) races. You want software that creates better conditions based on available stock and trainer desires. We want to reverse the current backwards way it’s done.
Longacres condition book.
Trainers can set alerts to prioritize horses looking for a race. Racing office can create races off these priorities. They already send out text alerts for changes.
Assess the classes of horses and then tailor the conditions to suit the horse rather than trainers trying to figure out which conditions match their runners. We sort of have the cart before the horse if you will. We need to change that.
21st Century Technology Makes Sense for 21st Century Horsemen
Trainers at the highest level are already using technology far superior to the racetracks. Does that make any sense whatsoever? Not a perfect analogy, but can you imagine a company where the employees had access to better information than the employer? Is that logical whatsoever?
The only argument out there is the “Super Trainers” and “Super Owners” view their technology as a competitive advantage and want to keep it to themselves. That is ONLY self-serving to the powerful and the goal of growing a racetrack is competitive balance, not allowing “Super Trainers” to crush everyone else. Allow smaller owners and trainers to have better technology, and they do better.
Updating technology will also need to expand into tracking owners. We have owners who raced 150 times in 2023 and only 5 times this year. Who is tracking this? What are they doing to retain owners? What technology is there to support owners?
Microchip scanning is inadequate as the chip bands are too weak to be scanned as a van passes by and must be hand scanned one horse at a time. This technology needs to be improved and integrated with new technology.
Background on Existing Technology
AI and advanced technology are already in use in the racing and breeding industries.
The Jockey Club not only holds a vast wealth of digital information, but it has recently introduced new digital and AI services.
The Jockey Club today announced a new service for breeders that provides a complimentary race record for Thoroughbreds foaled in 2025 and later that is updated after each domestic or international start.
The Jockey Club also introduced Naming AI, a new tool designed to assist customers with naming their horses.
Equibase sends out free track information daily which includes Entries, Results, Workouts, Overnights, Condition Books, Stakes Nominations and Handicapping Weights.
The Jockey Club has their own technology company.
InCompass is a technology solutions company formed by The Jockey Club in 2001 to centralize the software applications and systems that serve North American racetracks and simulcast outlets, helping these facilities achieve operational efficiencies, reduce costs and increase revenue.
InCompass’s central database also serves as a platform for several industry initiatives, including the Equine Injury Database and the Jockey Health Information System.
• Track Manager Provides a unified approach to the management of data and technology at the racetrack.
Track Manager has streamlined operations in the racing and horsemen’s bookkeeper offices and provides additional services to horsemen and other industry stakeholders.
• Horsemen’s Bookkeeper Takes all race setup information from the racing department so no manual entry or transfer is required.
• InCompass Financial Services Allows horsemen to access Horsemen’s Bookkeeper accounts via any internet connection
• Pre-Race Veterinary Exam Examining vets can track the progress of a horse throughout the horse’s racing career.
Other companies have seen the benefit of developing software for these industries including stable and veterinary tracking and security.
Management platform designed for race-horse owners and trainers. (Photo courtesy of Backstretch)
Backstretch, Novak’s name for his management platform designed for race-horse owners and trainers, is up and running with much more to come, he said. The basic cost is $10 a month, with the option for additional ad-ons.
Features include delivering real-time notifications for entries, results and workouts and post times on a single account. The platform is designed for efficiency, including to provide convenient communication between owners and trainers.
Equine MediRecord (EMR) has partnered with most major American Thoroughbred Racing events such as the Breeders’ Cup, Pegasus Invitational and the Preakness Stakes.
Setting a new global standard in the management and oversight of equine health and regulatory compliance, The Ireland-based company has launched in Britain, Ireland and the UK and are branching out into Asia and Australia. Their innovative platform standardizes the recording of comprehensive horse records, including medical histories, drug usage, and induction times, accessible effortlessly via phone or laptop.
Stall Monitor is your security force that never needs a lunch break. (Photo courtesy of Horcery)
As the designated Software of Record, Equine MediRecord meets and exceeds the stringent requirements set by equine racehorse regulators across globally. By mandating registration on the platform for all racehorses, equine healthcare maintains integrity.
Horcery produces the Stall Monitor, a cutting-edge system that provides 24/7 monitoring with AI-enabled cameras and real-time alerts. Horcery bills the system as helping to protect equine investments, improve stable management and ensure horse safety while empowering horsemen to reduce risks and optimize performance.
Stall Monitor, which hangs 9 feet up in the back of the stall, is hard wired, making it more secure than operating on wifi. Horcery is a division of Tavistock Group, a $20 billion private-equity firm based near Orlando, DeAngelis said.
From the Oldest to the Newest …
Image courtesy of TVG
TVG, now FanDuel Racing, has been in existence for 25 years and created their interactive app in 2014. Their database includes past performances.
Launched on July 14, 1999, TVG (short for Television Games Network) was founded as a joint venture of TV Guide Inc. (which at the time was owned by both Liberty Media and News Corp.), the National Thoroughbred Racing Association, and AT&T Broadband.
TVG launched the first native horse racing wagering application for iOS devices, following suit to the sharp increase in mobile gambling.
The new TVG App is optimized for the iPad and features free handicapping information, live racing video and race alerts to help players wager on over 150 races. That has since expanded to races at over 300 tracks.
Live streaming for 300+ tracks
Exciting promotions
Free race replays and past performances
Free Race Alerts
Improved handicapping store
Seamless funding and withdrawals
Free HD Streaming of FanDuel TV Networks
TVG talent picks
HISA, the Horseracing Integrity and Safety Aughotiry was created by an act of Congress to unify and simplify horse racing.
Their website states “HISA was created to implement, for the first time, a national, uniform set of integrity and safety rules that are applied consistently to every Thoroughbred racing participant and racetrack facility.
The rules and regulations drafted by HISA’s Racetrack Safety and Anti-Doping and Medication Control Standing Committees are designed to enhance the safety and wellbeing of both horse and rider while ensuring the integrity of the sport for the benefit of the industry, fans and bettors. A safer, fairer sport will also be a more popular sport for generations to come.”
Their services include information and data available for horses, jockeys, trainers, veterinarians, racetrack personnel, stewards, farriers and Horsemen’s Associations.
Future Possibilities: Pennsylvania Leads the Way
Starting gate at Parx Racing in Bensalem, Pa. (Barbara Weidl/EQUI-PHOTO)
After publishing a briefer version of the above, someone from Pennsylvania reached out about a new AI condition book project.
John Julia is a Thoroughbred owner and currently sits on the board of the Pennsylvania Thoroughbred Horsemen’s Association (PTHA). Having spent the last 30 years in technology and with an organization that builds AI systems, John truly understands how backwards we are in this business and how taking advantage of these technological advancements can change the trajectory of our industry. Julia, Senior Vice President of Sales at Pinnacle AI, explains the process.
MbK:Has anyone written any code for this?
JJ: As mentioned, we would not be building a traditional application. Foundation models would be trained on specific data sets to yield the desired output. Model selection, source data, and training methodology are just some of the considerations required. We are more in the realm of Data Science then application development.
How exactly will it work and what is the end goal of developing this app?
The goal of applying machine learning, Natural Language Processing and Generative AI is to enable the industry to compete in the 21st century. AI-ML can be applied to many areas of this industry. As mentioned, AI-ML is quite different from building applications.
The difference between building an application and developing AI/ML systemslies in complexity, purpose, and underlying methodology. Traditional application development typically involves building software designed to perform predefined tasks or workflows based on rule-based logic. These applications follow explicit instructions coded by developers, and their behavior is largely predictable and static unless modified through further development. In contrast, building AI/ML systems involves creating software that can learn from data, identify patterns, and make predictions or decisions without being explicitly programmed for every scenario.
This process requires not just software engineering, but also data engineering, model selection, training, tuning, validation, and ongoing monitoring. While traditional applications are task-driven and deterministic, AI/ML systems are data-driven and probabilistic—requiring iterative training and optimization to improve performance over time. Furthermore, deploying an ML model is only part of the equation; building a complete AI-powered application involves integrating that model into a broader system that includes APIs, user interfaces, inference pipelines, and feedback loops for continuous learning. In essence, developing applications is about building functionality, while developing AI/ML systems is about building intelligence into that functionality.
What data resources do you have to access to set it up? JC/Equibase? Or is it based on the contributed data the horsemen input?
Equibase has well over 200 data points per horse entered in each race. The foundation of the Data Lake uses Equibase. The limitation of Equibase for a condition book is that it does not provide a source of truth for where a horse is stabled. In order to train the machine to generate a condition book, additional data source such as Trach Manager (for each participating race track) would be required to get real time backside inventories. No human in the loop is required as the process would be fully automated.
Are you planning to roll out a beta and make required adjustments?
Please refer to the AI-ML implementation checklist that I sent earlier. There is a specific process for training models and implementing Artificial Intelligence systems.
Would this be membership-based and for a fee?
We have budgetary developments costs figured out at a high level. The scope would change a little bit if looking at building a regional condition book.
What was the inspiration for this proect?
I have two passions in life …. Horse racing and technology.
What impact on Pa racing do you foresee?
Traditionally developed through manual analysis and historical precedent, condition books often fall short in adapting to the evolving dynamics of horse populations, trainer behavior, and race demand. The result: inconsistent field sizes, uncompetitive races, and lost wagering opportunities.
Applying machine learning (ML) algorithms to condition book creation represents a transformative opportunity for the sport. By leveraging real-time data and predictive modeling, ML can dynamically optimize race offerings to better align with actual horse availability and trainer entry patterns. This data-driven approach increases the likelihood of fuller, more competitive fields—directly driving up wagering handle, fan engagement, and trainer participation.
Key benefits include:
Larger Field Sizes: ML models analyze past race outcomes, entry trends, and horse populations to identify conditions most likely to draw full fields. This reduces the frequency of short, unbettable races.
More Competitive Racing: Predictive analytics help create balanced conditions that reduce lopsided matchups, resulting in closer finishes, more attractive odds, and stronger betting interest.
Real-Time Adaptability: Algorithms enable racing offices to respond to ongoing changes—such as weather, injuries, or shifts in horse availability—by recommending adjustments to upcoming races mid-meet.
Trainer-Centric Design: ML identifies behavioral patterns in trainer entries, allowing condition books to align with the real opportunities trainers are seeking for their horses.
Operational Efficiency: Automating aspects of condition book creation allows racing secretaries to make more informed, faster decisions based on vast datasets that would be impossible to process manually.
In a sport where field size and race quality directly impact financial performance, machine learning delivers a powerful tool to improve both. With enhanced forecasting, flexible race planning, and intelligent scheduling, ML-powered condition books will play a pivotal role in the modernization and sustainability of Thoroughbred racing.
Will you expand to other Mid-Atlantic areas?
We would provide to any race track or jurisdiction that would have interest.
What is your long-range goal?
My long-range goal – Save the Thoroughbred Racing Industry
In closing …
It is public knowledge that Racing Secretaries are facing a very real problem filling races, and many owners are becoming fed up with their options.
Prior to HISA, databasing a horse would have seemed like an impossible task to ask trainers. This is not the case. Many of the top trainers are ALREADY using sophisticated software well beyond the scope of our Racing Office. Future Hall of Fame trainer Kenny McPeek is one of the many who seem to be advocating for these changes. In fact, I think this is one of the first times I’ve ever seen a solution to a problem have seemingly UNANIMOUS support among horsemen, fans, and gamblers.
About the authors: Maribeth Kalinich is Senior Editor and contributing writer for Past The Wire. Chris Brown is an owner and breeder in Maryland who races under the nom de course Designated Hitters.
