Engage-disengage: look at the trigger, then look back for pay

Engage-disengage is a threshold skill: the dog notices a mild trigger, then looks back to you and gets paid for that return of attention. On-leash threshold owns finding a distance where the dog can still look, breathe, and eat. Pattern games at threshold own predictable 1-2-3 style games at that distance. This page owns the look-at / look-back loop — not the counting game.

I train on month-plus residential stays from Russia. I am not a veterinarian. No affiliate harnesses, no fake reviews.

This is a reactivity protocol. Dog-to-dog cases also cross-link leash reactivity to dogs; people and bikes cross-link leash reactivity to people and bikes. Emergency U-turn is the leave pattern when the loop fails.

How does engage-disengage work?

The dog is allowed to see. You pay the return.

At a distance where the mouth stays soft, the dog glances at a person, dog, bike, or van. You mark either that glance or — once the glance is cheap — the look back to your face or hand. Then you feed. The trigger stays legal information. Your job is not to block every look; it is to make looking back pay better than staring until the body locks.

Distance first. The look-back is useless inside the scream zone.

What does an engage-disengage ladder look like?

One criterion at a time.

1. Quiet outdoor spot. Mild trigger far enough that the dog can still eat. Confirm soft mouth.
2. Dog glances at the trigger. Mark the glance. Pay. Reset. Short set.
3. When glances are cheap, switch the mark to the look-back to you. Same distance.
4. Add slightly harder pictures one raise at a time: closer meters, busier backdrop, shorter leash slack — never all three.
5. Body tightens or food refusal — leave. U-turn or exit. Management owns the next route choice.

Do not drag the dog toward the trigger while saying “watch me.” That is flooding dressed as attention work.

How is this different from pattern games?

Different structure. Same distance fence.

Pattern games give a handler-led rhythm the dog already knows: count, feed, leave. Engage-disengage pays voluntary notice-and-return. You can stack them across weeks, but do not blur them in one muddy session where you count, demand eye contact, and walk closer at once.

US sidewalk versus UK narrow pavement?

Same criterion. Different escape geometry.

US (typical): wider sidewalks and parking strips let you open lateral distance. Use that space before you shrink it.

UK (typical): narrow pavements, estate paths, and canal edges trap you fast. Build earlier exits and reset blocks so a failed look-back is not a forced face-off in a funnel.

When is engage-disengage the wrong tool?

If the dog cannot eat, you are too close — return to on-leash threshold. Fence fighting and barrier runs belong on fence fighting. Injury, fear biting, or a dog who never softens outdoors needs management and a qualified professional. See when to call the vet if pain or sudden change is in the picture. This page is not a diagnosis.

What are the usual failure modes?

• You started inside the scream zone. No look-back protocol survives that. Add meters.
• You mixed glance-mark and look-back-mark in one muddy set. Pick one criterion.
• You cloned pattern games here. Counting is a different page.
• You walked closer while the mouth went hard. That is flooding.
• You skipped the exit plan. Emergency U-turn and route management are part of the skill.
• You treated stare as progress. Locked staring is not engage-disengage.