How AI technology is influencing the breeding and training of racehorses

Article by Virginia Lisco

Artificial intelligence (AI) is growing significantly across all industries, revolutionising the way we approach work. In many areas, it facilitates the storage and processing of big amounts of data that the human mind would not be able to handle alone, helping us make more informed decisions based on objective information. 

Although this development is progressing at a slightly slower pace in the racing industry compared to other sectors, the introduction of AI and the use of tech devices that collect data and turn it into analysis are gradually finding their place in this field as well. 

The world of racing, which has so far been characterised by an aura of unpredictability - part of the excitement and thrill that makes it so captivating - is actually an industry filled with data and variables that can be measured and analysed. 

Thanks to AI and new technological devices, which leverage generations of data, extremely innovative systems are emerging that provide objective answers to questions that once relied solely on human intuition and observation. 

However, this does not mean that the rise of these devices should be seen as a replacement for human expertise, but rather as a valuable support. These tools will never replace the role of bloodstock agents, trainers or breeders, but will instead assist them in the decision-making process by offering insights based on empirical data, making their choices more informed and objective.

From the initial selection of the right match between mare and stallion, through identifying the ideal yearling at sales, to measuring the horses' vital parameters during training and providing race recommendations, artificial intelligence and tech devices are offering 360-degree support.

BREEDING

Since the earliest days of breeding, the dream of every thoroughbred racehorse breeder has been to create perfect champions, endowed with the ideal combination of strength, stamina and temperament. 

Over the centuries, the quest to unlock the secret of breeding invincible horses has lingered between genetic theories, intuition and bold experimentation. Federico Tesio, one of the most visionary figures in the field, argued that success was not merely a matter of a fortunate combination of genes but was deeply tied to the intrinsic energy of the horse. This approach, combined with rigorous selection and careful evaluation of bloodlines, formed the foundation for pursuing those winning combinations that could produce extraordinary champions.

But what if today’s breeders could go beyond intuition and legacy knowledge? What if data science could predict the best match for your stallion and mare? 

With access to an immense database of nearly four million horses, AI-powered predictive tools are now transforming breeding and buying decisions, enabling breeders to make choices grounded in data. 

These tools, built upon established pedigree theories, offer insights backed by over 30 generations of bloodstock data, tracing back to the very beginning of the thoroughbred breed. By harnessing this wealth of information, breeders can optimise their strategies like never before, ensuring each pairing is a precise, informed step towards crafting the next racing legend. 

One company leading this transformative shift is Equine Match, leveraging AI-driven systems to revolutionise pedigree analysis and breeding decisions. Equine Match's advanced tools allow breeders to analyse not just the direct bloodline of a horse but also to assess its performance across three to five generations. 

By examining the racing and pedigree features of ancestors, including sires, dams, and their progeny, Equine Match can predict with up to 95% confidence how a horse will perform at various racing levels, from Black-Type performers to stakes winners or group winners. 

Equine Match’s predictive algorithms calculate these probabilities based on a unique combination of sire and dam ratings, which factor in expected versus actual progeny results and evaluate the broader distaff family. The pedigree pattern analysis further weighs the success of specific ancestral combinations using Z Statistics scores, comparing results against a standard score to identify statistically significant patterns. 

With access to a vast database of nearly four million horses, Equine Match can assess the quality of prospective matches, enabling breeders to search for their ideal sire and dam combinations, view real-time performance metrics, and even analyse hypothetical pedigrees for potential inbreeding coefficients and alignment with the ideal racing profile.

Another cutting-edge advancement in the breeding industry is the use of genomic tools to create the perfect match. These tools go beyond traditional pedigree analysis by examining the genetic profile of horses at the DNA level, revealing insights that were never accessible before. 

One leading example is Checkmate, a revolutionary online breeding tool powered by Equinome. Checkmate enables breeders to analyse their mare's genetic profile in high definition and identify stallions that best complement her DNA, all from a simple blood sample that uncovers details beyond the pedigree.

Checkmate empowers breeders and stallion owners to manage genetic inbreeding risks, increasing the chances of live foals and enhancing racing durability. Scientific studies have shown that higher levels of genomic inbreeding are linked to severe issues such as foal loss, stillbirths and decreased racing potential. 

Research from the Royal Veterinary College and University College Dublin has highlighted the significant impact of inbreeding on foal viability and racing success, underscoring the importance of managing these risks. Checkmate provides real-time predictions of genomic inbreeding in hypothetical foals by matching the DNA profile of a mare with that of up to ten chosen stallions, calculating risk scores and flagging high-risk matings.

Checkmate helps breeders avoid high-risk genetic combinations that could lead to foal loss, thereby allowing them to make more informed breeding decisions. By leveraging genomic data, breeders can refine their breeding strategies, enhancing the health and durability of their foals and pushing the limits of what is achievable in thoroughbred breeding.

The advent of these advanced tools represents a significant shift in the breeding landscape. As Luca Cumani, breeder at Fittocks Stud, notes, “This will be a progressive change to refine choices, and it will always need to be paired with human observation, but it's important to be open to understanding how these new technologies work.”

AT THE SALES 

During the Sales, the phrase “good luck” is often exchanged, reflecting the hope of finding and securing the best horse, a potential future champion, from the lengthy catalogues. 

But is success really just about luck, or is there a way to minimise this element and base our decisions on more objective criteria? 

Recently, a range of advanced tools have been developed to support bloodstock agents, trainers, and owners in their search, using AI-based systems that draw from extensive databases to inform better decision-making. There are various methods to approach these objective analyses, including pedigree evaluations, biomechanical assessments.

For example, a significant advancement in this field is HALO, featuring its AI assistant "Hailey," designed to guide buyers through the complexities of sales selection. Hailey analyses detailed data on each horse, from performance history and pedigree to biomechanical evaluations. 

What truly sets HALO apart is its partnership with the most advanced Vet-AI on the market, that contributes to providing a record of the health of the horses that will be purchased. This partnership enables Hailey to deliver a detailed health record of the horse being considered, interpreting complex veterinary reports such as X-rays with remarkable accuracy and predicting potential health issues.

Also, Equine Match helps users refine their shortlists by analysing pedigree quality, racing performance of progeny, and detailed sire ratings, to provide an objective assessment of each horse’s potential. This approach enables users to manage and compare a large volume of data in a short time, far beyond what the human mind could achieve, enhancing the overall selection process with precise, data-backed insights.

Another interesting tool is Tom Wilson’s biomechanical profiling technique, which evaluates a horse’s physical structure and movement patterns to gauge its racing potential. Wilson has extensively studied the link between biomechanics, kinematics and on-track performance. His approach involves collecting detailed data on every yearling sold at auctions in major markets like the UK, Ireland, France, the US, Australia and New Zealand.

Key measurements include joint angles, the speed and velocity of specific body parts, the coordination between joints, and the cycles and rotations of their gait. By analysing this information from a growing database of 6,000 yearlings, Wilson identifies the biomechanical traits that correlate most strongly with racing ability. This data-driven approach allows comparisons between a horse’s biomechanical profile and its subsequent performance on the track, highlighting the most critical joint movements that influence future success.

The use of AI and data-driven tools represents a different approach from the traditional methods of horse selection, striking a delicate balance between technology and intuition. As breeder Franca Vittadini notes, “AI will never help me feel the sensation a horse gives me when I look at it, or see the small flaws and qualities that make me think it’s the right one.” 

Yet, there is also recognition of the need for more objective strategies. As Spencer Chapman, co-founder of Equine Match, states, “we’ve got to find a more scientific way to select.” This blend of traditional expertise and modern technology is reshaping how the industry identifies and invests in future champions.

RACING 

The use of AI in racing recommendations is revolutionising how trainers, owners, and jockeys approach race strategies. Advanced AI tools analyse vast datasets, including past performance, track conditions, and competitor analysis, to provide insights that were previously difficult to obtain. 

HALO’s simulation video allows users to visualise a potential race scenario, predicting how a horse might perform under specific conditions such as pace, positioning, and track type. This technology helps trainers make informed decisions on race tactics, selecting the right race for their horse, and optimising performance on race day.

HORSE PERFORMANCE AND HEALTH 

The health and physical condition of horses in training play a crucial role, not only in achieving optimal performance but, more importantly, in ensuring a high level of welfare. 

In the racing landscape, new tech devices have emerged that assist in monitoring and preserving the health of horses. These tools are essential for analysing causes of underperformance, determining the ideal fitness level for racing, and, most critically, reducing the risk of injuries. 

The trainer’s job is extremely challenging, requiring them to understand, observe and analyse a horse’s physical and mental condition from all angles. They must assess the horse’s aptitudes, such as distance preference, track surface suitability, and the appropriate amount and type of work, all while trying to maximise the horse’s potential. This delicate task relies not only on the trainer’s personal observations, instincts and sensations but also on the feedback of riders.

The use of GPS devices opens a new frontier in modern training, providing real-time insights into a horse's performance and physical condition. These specialised tools work by attaching sensors to the horse, usually in the girth area, to capture data on various metrics such as heart rate, stride length, speed, and acceleration. 

This data is then transmitted to a connected device, like a smartphone or tablet, where it can be analysed by trainers and veterinarians. The GPS provides insights into the horse’s fitness level, recovery rate, and overall performance, allowing for precise adjustments in training programmes.

By continuously monitoring these parameters, GPS devices help detect early signs of fatigue, underperformance, or potential injuries, enabling preventive measures to be taken and ensuring the horse maintains peak condition while minimising health risks.

An example of this advanced technology is the Equimetre by Arioneo, a device used by multiple trainers worldwide and in collaboration with organisations like the Hong Kong Jockey Club, NYRA and Racing Victoria. 

Trainers like George Scott rely on Equimetre to gather crucial data that enhances their training strategies. According to trainer Scott, “We use Equimetre to collect data which I’m becoming increasingly more committed to. A basic starting point is the fractions. You can back up what you’ve seen with your eyes against the clock. Cadence is a great indicator on trip. Recovery taken at three stages after work and trot symmetry are also helpful.” 

This data allows trainers to corroborate their observations with objective measurements, providing a more comprehensive understanding of a horse’s condition. The influx of data from such devices is also giving rise to a new professional figure in the racing industry: the data analyst. 

To address this growing need, Arioneo has introduced the "Equiscience Programme," aimed at developing skills in data interpretation and analysis specific to equine performance. This program reflects a broader trend in the industry, where data-driven insights are becoming essential for optimising training and enhancing decision-making, complementing the traditional expertise of trainers and riders.

The primary goal of these companies remains the enhancement of horse welfare. As Coline Labadie de Faÿ, COO of Arioneo, states; “The vision has always been to collaborate with the industry and provide tools to improve horse welfare and prevent injuries on a global scale. Institutions worldwide are becoming more and more concerned by the social licence of our sport and are putting concrete measures in place to try to avoid all possible injuries.” 

This growing awareness is driving the adoption of advanced tech devices designed to monitor and protect the health of horses, even around the clock.

Among these technologies is the V-PRO by Equimetrics, which offers 24/7 monitoring of vital signs such as temperature, heart rate, and respiratory rate, providing real-time insights into the horse’s well-being through a user-friendly app. 

Another example is EQUISYM, a veterinary tool that helps diagnose locomotor asymmetries by analysing gait and movement patterns, aiding in the early detection of potential issues. These innovations reflect the industry's increasing focus on comprehensive health management and injury prevention, supporting the overall welfare of horses in training.

While these advanced devices offer valuable insights, they are complementary tools that work alongside the hard work and expertise of trainers, veterinarians, and riders. They enhance traditional methods without replacing the essential human touch in horse care and training. As Labadie de Faÿ suggests; “It will never replace your eyes; it will just help to not miss some little things.” 

These technologies are there to support, not substitute, the critical role of human observation, judgement and experience in ensuring the well-being and performance of horses.

The Gerald Leigh Memorial Lectures 2023

The Gerald Leigh Memorial Lectures, is an annual  gathering devoted to the racing industry and the health and wellbeing of the horses involved.  

This year, equine veterinarians, researchers, students and industry professionals from around the world attended the event, held June 8, 2023, at the historic Tattersalls Sales in Newmarket, England.  

There were insightful and informative lectures that educated the attendants but also instigated a healthy, lively debate on the health and welfare of the training and competing of horses. The underlying theme that was present during the whole event was all members of the conference had a deep passion and commitment to continuously progress and improve on managing the welfare and wellbeing of the horses in the industry, both on and off of the track.  

Two very special guest speakers, Sir Mark Prescott and Luca Cumani, wonderfully illustrated these sentiments as they described their reflections on the improvement and enhancement of horse safety.  

Horse racing may be regarded as an elite sport, and all activities involving horses have an element of risk. All stakeholders in the racing industry must continuously work to ensure that the risks are minimised in order to reduce the number of injuries and fatalities that may occur in training and on the racecourse.  There are now well-publicised concerns regarding the acceptability of exposing horses to risk in racing.  These lectures and all of the attendees embraced the values of the public will so that there can be continued acceptance of horse sports.   

Reducing the incidence of fractures in racing 

Christopher Riggs of The Hong Kong Jockey Club clearly outlined the various strategies to reduce the risk of fractures in racehorses. There are two principal strategies that may used to reduce the incidence of severe fractures in horses while racing and training:

1. Identifying extrinsic factors that increase risk and take action to minimise them. 

An example would be investigating different racing surfaces in order to determine which may provide the safest racing surface. However, studies have provided limited evidence and support for subtle extrinsic factors.

2. Identifying individuals that are at increased risk and prevent them from racing or minimise that risk until the risk has subsided.  

There are many research routes that are being undertaken to identify those horses that may be at a higher risk of fractures. There are investigations involving heritability and molecular studies that may provide evidence of genetic predisposition to fracture. However, Dr. Riggs explained that further understanding of the relationship between genetic, epigenetic and environmental factors is required before genetic screening is likely to be of practical use.  

Pre-race screening of horses by diligent clinical examination is poor at reducing the incidence of fracture. Dr. Riggs described another strategy that may assist with a clinical examination that is the use of biomarkers in blood and urine.  

Unfortunately, the precision to be of practical value has so far remained relatively unrewarding.  Wearable technology that records biometric parameters, including stride characteristics, has shown some promise in identifying horses that are at increased risk of fracture; although Dr. Riggs explained that this work requires further development.  

Finally, Dr. Riggs described both the use and current limitations of  diagnostic imaging in identifying pre-fracture pathology in order to identify a horse at imminent risk of fracture.  He conceded that further knowledge of the significance of the range of abnormalities that can be detected by imaging is incomplete.

Dr. Riggs concluded his lecture by expressing that the implementation of  diagnostic imaging to screen  “high-risk” horses identified through genetic, epidemiology, biomarkers and/or biometrics may be the best hope to reduce the incidence of racing fractures. This field can be advanced with further studies, especially of a longitudinal nature.

Professor Tim Parkin of Bristol Veterinary School discussed the need for further investment in welfare research and education. One avenue of investment that should be seriously considered is the analysis of data related to (fatal) injuries in Thoroughbred racing over the last 25 years.  

It was expressed, with the abundance of data that has been collected, that some risk factors would be relatively simple to identify. An encouraging example in the collection and use of data to develop models in predicting and potentially preventing injury has been conducted by the Hong Kong Jockey Club funded by the Hong Kong Jockey Club Equine Welfare Research Foundation. This may provide an opportunity to pilot the use of risk profiling to contribute to decision-making about race entries.  In addition, the results of the pilot study combined with other sources of data may encourage race authorities to mandate the collection of veterinary and training data in order to help in risk mitigation.

Horse racing is an international sport, and there are different governing bodies that ensure racing integrity. However, the concept of social licence equestrian sports and Thoroughbred horse racing continues to gain significant public attention.  Therefore, racing governing bodies are increasingly aiming to provide societal assurances on equine welfare. 

Dr. Ramzan of Rossdales Veterinary Surgeons provided an eloquent and clear message during his lecture that race yard veterinarians and trainers are instrumental in ensuring good horse health and welfare and reducing serious injury of the horse both while training or racing, which will provide sufficient trust and legitimacy from the public and society.  This feasible goal can be reached with good awareness of members involved in the care and training of each individual horse and conveying this information and any concerns to their veterinarian.  The veterinarian can also contribute by honing their knowledge and skills and working closely with yard staff in order to make appropriate and better targeted veterinary intervention.   

In the last two decades, there has been an incredible evolution and exciting developments in diagnostic imaging in the veterinary profession. It is believed that these technologies can provide a significant contribution to helping in mitigating fracture risks to racehorses on the course and in training.  

Professor Mathieu Spriet of University of California, Davis, described how these improvements in diagnostic imaging has led to the detection of early lesions as well as allowing the monitoring of the lesions’ evolution.  

He continued by explaining the strengths and limitations of different imaging modalities such as computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography (PET).  Being one of the leaders in the use of PET in equine veterinary medicine, he presented further insight on how this particular modality provides high-resolution 3-D bone scans while being very sensitive to the identification of bone turn-over prior to the development of structural changes and allowing one to distinguish between active and inactive processes when structural changes are present.  

He concluded his impressive lecture by providing evidence with amazing PET images that the role of imaging is not merely for diagnostic purposes to characterise clinical abnormalities, but can also be used as a screening tool in certain horse populations for fracture risk assessment or for the monitoring of lesions to provide clearance for racing. 

Fractures, due to bone overloading rather than direct trauma occur commonly in Thoroughbred racehorses and are the leading cause of euthanasia on the racecourse. Despite many changes to race conditions, the number of catastrophic fractures has remained relatively static, with approximately 60 horses a year having a fatal fracture during a race in the UK.  

Against this backdrop, there have been great developments in the diagnosis and treatment of fractures in the last 40 years. Prevention of racecourse and training fractures would be ideal so the development of efficacious techniques to screen horses at risk may reduce the incidence and preserve social licensing.  

One technique discussed by Dr. Ian Wright of Newmarket Equine Referrals was to help mitigate the impact of racecourse fractures, which would be acute immobilisation of racecourse fractures, thus, reducing associated pain and anxiety while optimising clinical outcome and reducing on course fatality rates. Because of our increased understanding of fracture pathogenesis and their associated biomechanics, effective fracture immobilisation has been made possible. The majority of fractures that occur in flat racing and between obstacles in jump racing, are a result of stress or fatigue failure of the bone and not associated with trauma.  

In addition, fractures seen on the racecourse are often found in the same specific sites (i.e., metacarpal/metatarsal condyles and the proximal sesamoid bones of the fetlock) and have repeatable configurations. With this understanding and knowledge, racecourse veterinarians can optimally immobilise a fracture in a logical and pre-planned manner.  

As Dr. Wright expressed, this allows the fracture patient to have reduced pain and anxiety and enable the horse to be moved from the course comfortably so that it can be further examined. Ultimately, this allows the veterinarian and all stakeholders to make effective and judicious decisions for the sake of the horse’s welfare and wellbeing. As Dr. Wright concluded, this benefits both horses and racing.

Dr. Debbie Guest of the Royal Veterinary College discussed a different approach in mitigating the risk of fractures during training and racing by developing novel tools to reduce catastrophic fractures Thoroughbreds. Because it has been found that some horses are more inherently predisposed to fractures than other horses, Dr. Guest and her team have developed a genome-wide polygenic risk score so that one can potentially calculate an individual horse’s risk of fracturing during training or racing compared to the population as a whole.  

This strategy may contribute in identifying genetically high-risk horses so that additional monitoring of the patients can be exercised during their careers and also leading to fracture risk, which are found to be the cause of approximately half of these incidents.  

The system of using DNA testing to identify biological processes that may or may not be present ultimately leading to fracture risk may be a powerful tool in lowering the risk of catastrophic fracture and requires further research and application.

Cardiac events & sudden cardiac death in training and racing

In racehorses, sudden death that is associated with exercise on the racetrack or during training is a serious risk to jockeys and adversely affects horse welfare and the public perception of the sport. It is believed 75% of race day fatalities result from euthanasia following a catastrophic injury. The other 25% of fatalities is due to sudden deaths and cardiac arrhythmias are found to be the cause of approximately half of these incidents. The lectures focused on this area of concern by providing three interesting lectures on cardiac issues in the racehorse industry.  

Dr. Laura Nath of the University of Adelaide, explained the difficulties in identifying horses that are at risk of sudden cardiac death. It is believed that part of the solution to this difficult issue is the further development and use of wearable devices including ECG and heart rate monitors.  

With the use of these technologies, the goal would be to recognise those horses that are not progressing appropriately through their training and screen these horses for further evaluation. This course of action has been seen in human athletes that develop irregular rhythms that are known to cause sudden cardiac death with the use of computational ECG analysis, even when the ECGs appear normal on initial visual inspection.  

Knowing that ECGs and particularly P-waves are used as a non-invasive electrocardiographic marker for atrial remodelling in humans, Dr. Nath recently completed a study on the analysis variations in the P-wave seen on ECGs in athletic horses and found that increases of P-waves in racehorses are associated with structural and electrical remodelling in the heart and may increase the risk of atrial fibrillation (cardiac event).

Dr. Celia Marr of Rossdales Veterinary Surgeons continued the discussion of cardiac disease in both the training and racing of horses. Unfortunately, cardiac disease knowledge does lag compared to musculoskeletal and respiratory diseases when considering the causes of poor performance in racehorses. Due to the fact that cardiac rhythm disturbances are fairly common, occurring in around 5–10% of training sessions in healthy horses in Newmarket and over 50% of horses investigated for poor performance, Dr. Marr expressed the need for further research and investigation in this area.  

In addition, this research needs to determine if there is indeed a link between heart rhythm disturbances and repeated episodes of poor performance and sudden cardiac arrest. ECGs and associated technologies are helpful, but there are limitations such as the fact that rhythm disturbances do not always occur every time the horse is exercised.  Therefore, it would be of great value that a robust criterion is established when evaluating ECGs in racehorses. The Horserace Betting Levy Board has provided funding for investigation by initially exploring the natural history of paroxysmal atrial fibrillation (self-correcting form) to understand risk factors and predict outcomes for affected horses.

Continuing the theme of the lectures on irregular heart rhythms and associated sudden cardiac death (SCD) in training and racing, Professor Kamalan Jeevaratnam described his exciting research in using artificial intelligence (AI) to identify horses at increased risk of developing irregular rhythms that may cause SCD.  

AI is an exciting and rapidly expanding field of computer science that is beginning to be implemented in veterinary medicine. With funding by the Horserace Betting Levy Board and the Grayson Jockey Club Research Foundation, Professor Jeevaratnam of the University of Surrey, has piloted three novel algorithms that help predict horses with rhythm abnormalities through the analysis of horses’ ECGs.  

It was acknowledged that further research is required to develop this technology by using data collected from multiple sources, but the initial results are promising in the development of an useful AI tool to identify horses at risk of SCD and prevent catastrophic events, thus, ensuring the welfare of the horse in racing.

Conclusion

The Gerald Leigh Memorial Lectures was a thoroughly successful and enjoyable event attended by a variety of different members of the horse racing industry. Not only did the lecturers provide interesting and valuable information but also excitement for the future of racing.  It was very clear that all the lecturers and attendees were passionate and committed to the racehorse welfare and wellbeing as well as retaining the social licence for an exciting sport.