Why does my dog's heel fall apart when we walk?

King is the case the distinction was made for. His heel was his weakest command all program and the best-documented one, and the split is what makes his record legible: the failure is located precisely rather than generally. On a nature trail his handler narrates it in seven words - King, heel, so he lost it on the turn - which is a heel-in-motion failure, and specifically a turn failure, not a bad heel. Earlier in the program the entry is run as its own named drill for a stated reason that has nothing to do with duration: a few heel entries, to clarify the position for him, with free sniffing between them. And at the two owner lessons the split is taught to the family in full, including the line that makes it usable at home - technically the heel entry could be done from not moving. The repair for the turn is taught rather than demonstrated: mark it, pat your leg so he can see where the spot is, keep going. His file also contains the cleanest example of correcting the right thing: he comes out of ball play still holding the ball, and the handler separates the two problems out loud - you can hold the ball, I don't care; that is the exact opposite of heel.

Rocko's material, preserved in the Kodak source folder, distinguishes entering heel from maintaining a moving position. The trainer describes a progression from entry to movement, added difficulty and turns. He also compares entries while the handler continues walking with entries beside a stationary handler, and names Rocko while explaining practice from ahead of the handler. A separate Rocko passage describes lagging behind and the trainer's interpretation of that response. These examples explain the curriculum and handling choices in these lessons; they do not establish that this is the only complete curriculum in the wider archive.

Kodak's separate session describes retaining recent heel work. The trainer reports an initially uncertain entry followed by a little leash clarification and says no collar input was needed for that repetition. This is a Kodak example of the work in practice, distinct from the Rocko footage explaining the progression. The selected passage does not by itself establish the whole training timeline or outcomes in other settings.

Zeus's file states the teaching model as an explicit numbered sequence before the dog is asked for anything, which happens nowhere else in this reader's material: one, teach where the heel is - the position relative to your body; two, teach him that the position moves and it is his job to fix it; three, once he owns both, deliberately try to shake him out of it and give consequences, positive and negative, for keeping it. The entry is treated and named as a separate repetition from the heel itself, and it is given a second function beyond convenience - it is how the dog learns what the position is at all. It is also the half Zeus struggles with: the handler calls the rep out loud as the heel entry in one short session, and at the owner lesson the progress report names both halves and then describes the entry in terms of the picture the owner will actually see - he is out there walking ahead of you, which is where he is probably going to walk a lot of the time. That lesson also gives the only spatial definition of the position in the corpus: the head and shoulders inside a box, not back here and not up here.

Bullet supplies the clean evidence that the two halves dissociate, because the record follows him. The split is named on a facility session - staying in the position is the heeling motion, coming into the position is the heel entry, and those two things go together - and both are taught with the handler's own body position and body language plus a little negative reinforcement from the leash, with the e-collar deliberately not used at all yet. The criterion at this stage is not a position but an awareness: all he is looking for is the dog paying attention to where his body is in relation to his own. Weeks later, on a long proofing walk in a busy park, the assessment does exactly what the naming makes possible - the motion is good and the entry is the part that still needs work: he does know where the position is.

Dolce's file does not name the two halves, and it supplies the thing the other files lack: a criterion for the motion half that an owner can check without a trainer. She was going past him a little when he stopped hard on the command, and what he names as the requirement is tracking - she needs to track it and make sure that she stops when I stop. That converts 'entirely focused on you' from an abstraction into something visible, and he says the two in one breath, which is why they belong on a page together. Her heel word is beifuss rather than fuss, deliberately, because fuss is the bite command - a separation kept obvious for clients.

Roxy's file is included because it shows the entry being drilled by a different route and states the hazard the whole curriculum is designed around. The route: let her go out ahead, then call her back into it - which is training the entry by deliberately manufacturing the picture the entry is needed for, the same logic as training hardest the entry the dog will actually use. The framing is the obligation one - she needs to learn that it is her job to keep the position - which is the same sentence as Kodak's phase two in different words. The hazard is that heel is a grey criterion: you have to reinforce something as clear as possible to the dog, and negative reinforcement applied to a heel that is still unclear causes real problems. That is the argument for phase one existing at all - the entries are what make the position unambiguous before anything is enforced.

Louie is here for one moment, and it is a good one: on his first heel session the handler tries the stop test before the dog could possibly have learnt it - I did a stop, but Louie didn't really know what that was - and calls the result encouraging. It shows the tracking criterion being used as a baseline probe rather than as a standard, and it dates the criterion to the very beginning of a dog's heel work rather than to the end of it.

Sage is here for the same reason as Louie and with the same test: a stop dropped into an early heel session to see what happens, on a nervous dog, with the result noted rather than corrected. Her handler's plan for what comes next reads as the same curriculum in lay terms and in the right order - a little longer, then abrupt stops, then fast and slow.

Why does his heel look fine and then fall apart as soon as we are actually walking?

Check which part changes. King’s trainer separates coming into position from maintaining that position while the handler walks (0Ryn-sQhWK0, 24:15). Bullet’s explanation uses the same distinction (epglfzzRDaQ, 14:07). A dog may therefore need more work on one part than the other. Compare the entry, the first steps and later movement before choosing what to practise; this distinction is a useful observation framework rather than a diagnosis of every heel problem.

How do I practise the entry, exactly?

King’s lesson demonstrates that an entry can begin with the handler standing still (0Ryn-sQhWK0, 24:33). Another session describes a few entries with sniffing between them (3F2ALQ6pFME, 04:17). Rocco’s lesson compares joining a moving handler with coming into position after the handler stops, and emphasizes the approach he commonly needs (7_ZqvaivpPU, 27:54–28:44). Use these recordings to understand the different pictures; choose the cue, assistance and amount of practice from your own lesson.

How do I know whether the heel in motion is actually good?

Stop walking and see what he does. If he stops with you, the motion half is working; if he sails a stride or two past you, it is not - and that is the criterion a trainer gives on camera in exactly those words: she needs to track it and make sure that she stops when I stop. It is a tracking test rather than a tidiness test, which is why it works: a dog holding a position by habit will overshoot, and a dog attending to you will not. Two other tests the trainers use deliberately, and you can too: change pace - speed up and slow down, and see whether he matches you - and turn, which is the single most common failure point. One handler slowed down on purpose specifically as a test, and the dog went out ahead. Try it before you assume you have a finished heel.

He does not follow me round corners. Should I correct him harder?

No. The trainers treat a missed turn as the dog not having the information, and they repair it with information: mark the error, then pat your leg so he can see where the spot has moved to, and keep walking. The same instruction is given twice on camera in the same words, once demonstrated and once taught to the owner, which suggests it is house practice rather than one handler's habit. This is also the reason one of the two stated curricula puts turns and figure eights in a phase of their own at the end - they are their own body of work, after pace changes and distraction, rather than something a dog with a good straight-line heel should already be able to do.

Where exactly is he supposed to be?

Honestly: the corpus does not agree, and it is worth knowing that rather than being given false precision. One trainer describes a box for the head and shoulders - not back here, not up here. Another describes it as near his feet and pretty close, without touching, and adds a practical scaling rule: bigger dogs are held tighter, and smaller dogs are given more room, because a small dog trained to follow your foot exactly is a small dog you will tread on. A third says he is not looking at a place at all but at whether the dog is paying attention to his relative body position. The one thing they agree on is that vagueness is the problem, not the solution: heel is a grey criterion, and correcting a dog against a standard he cannot read is named as causing real problems. So pick your version, make it the same every time, and teach it with entries before you enforce it.

Do I need an e-collar to teach this?

Not for the part this concept is about. One trainer says it plainly while teaching both halves: he is using his own body position and body language plus a little negative reinforcement from the leash, and not the e-collar at all yet. Another, grading a good entry weeks into a program, says he clarified with the leash and did not use the e-collar for it - no need. Where the collar does appear in heel work it is deliberately low: the handler working a lagging evasion out of a dog says outright that he does not have the level up high relative to what the dog feels, because the point is that the lag stops paying rather than that the dog is impressed. The thing that does the teaching, on this evidence, is clarity about the position - which is what the entries are for.

Only ever practising the heel by walking. The entry is half the command, it is the half that defines the position, and it can be practised standing still in a kitchen.

Escalating when the dog misses a turn. The trainers treat a missed turn as missing information rather than as refusal: mark it, pat your leg, keep going.

Correcting the heel while the dog still cannot read where the position is. The criterion is grey by nature, and enforcing a grey criterion is named as causing real problems.

Turning one problem into two. If the position is wrong, correct the position - not whatever else the dog happens to be doing at the same time.

Only ever showing the dog one entry picture. Handler walking and handler stopped are two different pictures, and a dog who has only seen one will fail on the other.

Holding the heel and hoping, when the dog has quietly discovered that hanging back ends it. Raising the level is the wrong answer; making the strategy fail is the right one.

Doing the whole job for him. If you are steering him into position and keeping him there, you have not taught the command - the point of phase two is that keeping it becomes his job.

He gets into position fine and drifts out of it once you are moving - or the reverse, he holds it well and never gets into it cleanly. Either way you have a named half to work on rather than a bad heel.

He walks nicely beside you in a straight line and loses it every time you turn a corner. That is the specific failure one dog's whole record turns on.

You stop walking and he keeps going a stride or two past you. That is the tracking test failing, and it is the criterion the trainers use.

He falls apart specifically when you speed up or slow down. Pace change is used as a deliberate test by the trainers, so it is a known pressure point rather than an oddity.

He starts drifting a long way behind you when you ask for a heel, and the heel tends to end shortly afterwards. Consider that he may have worked out the connection.

He comes in beside you when you are already stopped but cannot find the spot while you are still walking, or vice versa. Those are the two entry pictures, and most dogs have only been shown one.