In tennis, line calls were made by human judges with human eyes, human biases, and human error. Players disputed calls. Crowds argued. Reputation influenced decisions in ways nobody could quite prove but everyone suspected.
Hawkeye didn't eliminate the human element. It didn't replace the umpire. It measured the one thing that was objectively measurable — did the ball land in or out — and removed that specific question from human judgment entirely. The umpire still calls lets. Still manages the players. But the ball's landing position is physics. Measure it.
Dressage has exactly the same problem and exactly the same solution available.
The Division of Labour
AI scores the horse. The judge scores the horsemanship. These are not the same thing and should never have been treated as the same thing. The confusion of the two is precisely how the current system became corrupted.
The horse's gaits are biomechanics. Rhythm, regularity, purity, stride ratio, symmetry, suspension. These are measurable physical phenomena. They do not have opinions. They do not care who is riding or what the horse cost or which country the judge comes from.
A four-beat canter is a four-beat canter in an empty arena at 2am and it is a four-beat canter at the Olympic Games with twenty thousand people watching and a judge who has watched this combination win for a decade. The AI does not know the difference. It counts the beats.
What AI Measures
Rhythm and Regularity
The purity of the four-beat walk, the two-beat trot, the three-beat canter. Accelerometers, pressure plates, and high-speed video analysis measure footfall timing with precision no human judge can match. A four-beat canter — one of the most common faults at upper levels, the direct biomechanical consequence of over-collecting a horse whose physical development cannot support genuine collection — can be identified objectively every time. Currently it is frequently missed or forgiven because the overall impression is spectacular.
Stride Ratio
The ratio between collected and extended work demonstrates the horse's genuine range. A horse showing a ratio of 1.1 to 1 between collected and extended trot is not genuinely collecting and extending — it is changing tempo and perhaps elevation, but the physical range is not there. A horse with a ratio of 1.6 to 1 is demonstrating something real about years of correct development. Technology measures this exactly. Currently it is visually estimated, and spectacle inflates the estimate.
Symmetry
Left-right symmetry of movement is a fundamental indicator of straightness and soundness. A horse consistently loading its left hind more than its right is telling the judge something important — about soundness, about crookedness the rider has not corrected, about the asymmetry that training is supposed to resolve. Currently this asymmetry is missed in the spectacle of a big-moving horse. Technology flags it every time.
Gait Purity
The lateral walk — where tension causes the natural four-beat sequence to collapse toward pairs of legs moving almost simultaneously — is one of the most welfare-significant gait faults in competition dressage. It is directly caused by the tension that rollkur and tight nosebands produce. It is frequently scored too generously because the walk looks active. Sensors distinguish active from lateral with precision that no human eye can match in real time.
What the Judge Is Left to Do
Once AI has scored the gaits, the judge's task becomes both simpler and more demanding.
Simpler because they are no longer trying to simultaneously assess rhythm, regularity, stride length, gait purity, expression, relationship, and accuracy — doing all of it imperfectly, with the whole assessment skewed by the horse's natural talent and the rider's reputation.
More demanding because what remains is the hardest thing to assess and the most important: is the horse a willing participant? Is this partnership genuinely harmonious? Is the collection offered or extracted? Is the expression real or is it tension masquerading as brilliance?
These questions require genuine expertise, genuine feel, and genuine understanding of horses. They cannot be outsourced to a sensor. They require a human being who has felt what correct looks like and can recognise it at a glance. The judge becomes, finally, what the judge was always supposed to be. Not a measurement device. A witness.
The Scoring Structure
The score is a composite. Sixty points from AI for objective gait quality — rhythm, regularity, stride ratio, symmetry, suspension. Forty points from the judge for willingness, expression, harmony, and the quality of the relationship visible between horse and rider.
A horse with a magnificent gait ridden with tension and visible resistance might score 52 from the AI and 24 from the judge. Total: 76. Good horse, compromised relationship.
A horse with a more modest gait but extraordinary harmony and genuine lightness might score 44 from the AI and 38 from the judge. Total: 82. Correct horse, exceptional partnership.
Under the current system the first horse wins. Under this system the second horse wins.
Which is right? Which is dressage?
The FEI Contradiction This System Would Expose
The FEI rules require the horse to go with its nose slightly in front of the vertical at all times. The same rules describe the Give and Retake of the Reins as a movement where the nose may come slightly in front of the vertical — a statement that only makes sense if the normal working position is behind the vertical.
AI measurement makes this contradiction impossible to maintain. When sensors measure where the nose actually is, stride by stride, throughout every test at every competition, the gap between the rule and the reality becomes data rather than impression. The sport would have to choose: enforce the rule, or change it to reflect what is actually happening.
The contradiction has survived because it has never been measured. Measure it and it cannot survive.
Show Jumping: A New Vocabulary
The immediate application in show jumping is not judging — a clear round is a clear round. But AI gait analysis gives show jumping something it has never had: a data layer that makes the invisible visible.
Stride variability on approach: the horse arriving at every fence on a consistent stride is genuinely balanced. High variability means the rider is managing the distance rather than trusting the rhythm. Measurable. Meaningful. Visible to commentators and audiences.
Time in the air: which horses genuinely jump and which are fast and careful but essentially flat. The data makes the quality of the jump visible — not just whether poles stayed up, but how the horse moved.
Welfare indicators in broadcast: if a horse's stride is shortening progressively through a class, if its recovery after landing is deteriorating — that data can be visible to officials and vets in real time. Before the horse breaks down. Before the catastrophic failure that ends careers and appals audiences.
This gives equestrian broadcasting the vocabulary it has never had. Not "what a magnificent jump" but numbers that mean something to casual viewers. The Hawkeye of horse sport — giving audiences the language to understand what extraordinary horsemanship actually looks like in measurable terms.
The Trust Problem This Solves
Dressage judging has a trust problem that is destroying the sport's credibility. Riders from smaller nations do not trust that they score fairly against established names. Welfare advocates do not trust that irregular gaits produced by tension are being penalised. Classical riders do not trust that genuine self-carriage is being rewarded over managed spectacle.
The trust problem is not fully solvable by technology. But the part of it that comes from unmeasured things being measured imperfectly — that part is completely solvable. Measure them properly and the argument disappears.
Hawkeye did not make tennis less beautiful. It made it more honest. The same technology applied to dressage does not reduce horsemanship to data. It protects horsemanship from being faked by data's absence.