Every horse walks. Few get measured.

Twenty-one landmarks cleared the floor. Ten did not, and you can see which.

This is one real walk, ten seconds, 241 frames from a handheld phone. Each bar is the share of those frames a landmark was actually found in. The line at 0.90 is the floor a landmark has to clear before any figure is allowed to use it.

Three stages, and we will tell you all three

1. 01

   Film the walk

   Side on, on a phone, in the time it takes to walk your horse out and back. No rig, no markers, no special surface, and nowhere you have to take the horse to.

2. 02

   It measures

   Landmarks are found in every frame. Where a joint is not clearly there it is recorded absent rather than guessed, and the paths those landmarks sweep become the movement figures.

3. 03

   It tells you what it could not measure

   Every figure is checked against what is biomechanically possible for the gait you filmed. What fails the check is withheld, and the read names which check failed and why. What comes back is a record you can keep, and one your vet can read beside their own examination.

The horse’s own units, not ours

Every figure comes off the landmark paths, and every one carries the confidence the tracking earned rather than a flat number with no provenance. Measured the same way each time, they are meant to be held against the same horse a season later.

Stride length

How far the horse travels in one complete cycle, reported against its own frame so a big horse and a small one compare honestly.

Cadence and rhythm

Strides per minute, and how much each stride varies from the one before it.

Duty factor

The share of the stride a limb spends on the ground. A walk has no moment of suspension, so anything under half is impossible and is caught.

Overstep

Whether the hind foot lands ahead of the print the front foot left, and by how much.

Head carriage

The rise and fall of the poll through the cycle.

Fetlock drop

How far the fetlock sinks under load.

What is missing matters as much: nothing on this list compares one side of the horse with the other. That work is not finished, so those figures are not published.

Anything can print a number. The hard part is knowing when not to.

• It will not tell you what the movement means

  It measures and it reports the measurement. What a figure means about a particular horse is a veterinary judgement, and it stays one.

• It cannot yet compare one side with the other

  The landmark model was trained on standing photographs of horses, so it cannot reliably tell the near leg from the far leg as the two cross. Every left-against-right figure is therefore withheld, and stays withheld until that work is finished. What you read today is how the horse moves, not whether one side differs from the other.

• It will not score footage it cannot read

  If the tracking loses a hoof, or the measured stride would put a hind print closer than the limbs are mounted apart, the read is withheld and names the check that failed.

• It will not quote you an accuracy figure

  No validation study has been run. Until one exists, with its protocol written before anyone sees the data, there is no percentage worth printing here. When there is a figure, the study behind it will be published alongside it.

• It will never stand in for your vet or your own eye

  It is decision support on both sides of the consultation: for an owner, a record to bring to the vet; for a vet, a second record beside their own examination. It is never a diagnosis, and it does not replace a veterinary examination.

For vets, and for everyone who keeps a horse

One horse in a paddock is as much the customer as a practice with a list of clients. Movement is half the picture; the radiographs are the other half, and StrideVision is built to sit beside EquiRay so a horse can be described by both rather than remembered by neither.

Neither product looks at a horse and tells you what is wrong with it. EquiRay does not read films for pathology; every finding in it comes from a veterinarian. Together they give the fuller picture, movement and structure, both recorded rather than remembered.