What is back-chaining in dogs?

It's a technique for teaching a sequence of tricks: you learn the last step first, then work back towards the start. Because the end is the best known, your dog is always moving towards something familiar, which steadies the whole chain.

Why does my dog drop out halfway through a sequence of tricks?

Almost always because a link isn't solid enough on its own: it collapses once it's put into a chain. Break it down, rework that link in isolation, then chain it back in. Never any insistence or force to 'reposition' your dog.

How do I get started with dog dancing (freestyle)?

A routine is a long chain built up in small blocks with back-chaining, then set to music. Work on a non-slip floor and limit jumps and twists ; nothing acrobatic before growth is finished.

Is it better to learn the first or the last step first?

The last one. That's the whole point of back-chaining: your dog moves from the less familiar to the more familiar. Starting at the beginning, it moves away from the familiar and the end of the sequence falls apart, exactly where back-chaining makes things solid.

Chaining tricks: teach the last step first

Teaching your dog tricksPart of · Teaching your dog tricks

Chaining tricks: teach the last step first

Q: How do I teach my dog to link several tricks together?

First check that each trick holds up on its own, under its own cue. Over-learn the last link, then add the second-to-last one while rewarding only at the end, and work back one link at a time. Never add a link while the previous one is still shaky.

Q: Should I reward every trick in the sequence?

At first, yes, you can mark a few links. Then you gradually fade towards a single, generous reward at the end of the chain. That's what pays for the whole sequence, and the run-through eventually becomes rewarding in itself.

Q: How long does it take to teach a dog a routine?

Count on weeks , not one evening. Short, spaced-out sessions help learning (Demant et al., 2011). Two or three successful run-throughs per session beat one long session that ends up unravelling.

11 min read

Back chaining is a simple way to link several tricks together: you teach the last link of the chain first, then add the earlier ones one by one, working backwards towards the start. That way your dog is always heading towards the part he knows best, until he can run the whole sequence off a single word.

How back chaining works

A chain is several already known behaviours performed one after the other off a single cue (go to the toy, pick it up, carry it to the basket, drop it in = "tidy up"). With back chaining, you teach the last link thoroughly first: it becomes the anchor point for the whole chain. Every time you add a link in front, your dog moves from the less familiar towards the more familiar, so towards what reassures him. Start from the beginning instead, and he moves away from what he knows as he goes along, so the sequence ends up falling apart at the far end.

Check every link before you chain

This guide starts from tricks your dog can already do, each one on its own: it doesn't teach them from scratch, it teaches you to link them up. Before you chain anything, check that each trick stands on its own, on these four points.

0 / 4

Building the chain

You always start from the last link and work backwards, one link at a time; the reward moves towards the end of the chain.

Strengthen the last link

Your dog should already be able to do this trick on its own, on its own cue (if that isn't the case yet, teach it first with the foundations of reward-based training). Here you are not teaching it: you are making it smooth. Ask for it on its own, several times over, at different moments, and reward every success generously. Sign of success: he offers it without hesitating, several times in a row, before you move on.

Add the second-to-last link

Say the cue for the second-to-last link out loud, as if it were a trick on its own (for example "fetch" to go and get the object). Your dog does it, then carries straight on into the last link, which he already knows, without you saying anything in between (he drops the object in the basket). Let him run both links right through, then say your marker word ("yes!") and reward at the end of this two-link sequence, never in between. The last link becomes the reward for the one before it.

Work backwards one link at a time

The principle stays the same with each new link: say out loud the cue for the one link you are adding at the front, let your dog carry on by himself through everything already in place, then say your marker word and reward once the whole sequence is finished, never before. Golden rule: only add a link when the part already built runs without hesitation, every single time.

Fade the rewards along the way

At first, keep marking and rewarding two or three links along the way, as well as the end of the chain. Then remove the markers in the middle one by one, starting with the one in the middle of the sequence: let your dog run that link with no marker and no reward, straight on to the next one. If he hesitates when you take that marker away, put it back and try again later, in the same session or the next one. Carry on until there is only one reward left, a generous one, right at the very end of the chain: running the whole sequence then becomes a pleasure in itself for your dog, almost as much as that final reward.

One single cue, in several places

Say your single cue ("tidy up", "routine") just before the first link of the chain. To fade out a cue in the middle, start with the one in the middle of the sequence: say it more quietly at first, almost under your breath, then wait a little longer before saying it, one second more, then two, letting your dog carry on by himself during that pause. Once he carries on without paying attention to it, say the cue only every second or third time, until you drop it altogether. Then run the whole sequence in two or three new places, going back to a shorter version in each new place if you need to.

A chain is built over weeks, not in one evening. A 2011 study on laboratory dogs shows that a spaced-out rhythm (one or two sessions a week, one session a day) leads to faster progress than daily training repeated several times over, without changing what the dog retains a month later. Aim for two or three successful run-throughs per session, then stop: that beats a session that drags on.

The link that drops out

A dog who stops, barks, whines or improvises in the middle of a sequence is not being stubborn: it is nearly always a shaky link collapsing once it is put into a chain. The fix is always the same: take that link on its own, work on it again, then reattach it to the rest of the sequence. After three failures in a row, shorten the chain and let him succeed before you start over. Never push, and never physically force your dog to "put him back in position": in 2021, the American Veterinary Society of Animal Behavior (AVSAB) restated that physical force has no place in training a dog.

Three chains to practise on

The first two examples assume the basic link, dropping an object on cue or carrying it in his mouth, is already in place for your dog, independently of any chain. If that isn't the case yet, teach it first with the dedicated guides, Tidy up your toys and Go find & fetch, before using it here to build a chain.

Tidy up your toys

The last link is dropping the toy into the basket: that's the one you teach first, thoroughly.
Picking the toy up is added next, working backwards from that first link you taught.
You reward only after the toy is dropped in the basket, never at the moment he grabs it.

Take it to someone

The final link is placing the object in the person's hand: that's the one you work on first, thoroughly.
Carrying the object, then picking it up, are added next as you work backwards.
The reward comes from the person who receives the object, never while your dog is on his way back to them.

Go to your place (at a distance)

The mat is the final link: your dog already associates it with something pleasant.
You move the starting point back by about a metre at each stage, and you only move on to the next stage after three successes in a row at the current one.
You reward on the mat, never by calling your dog back to you.

AVSAB — Position statement on humane dog training (2021)
Demant et al. — The effect of frequency and duration of training sessions on acquisition and long-term memory in dogs (2011)

Going further

guideThe foundations: teaching a trick with rewardsThe base to lock in before you chain anything: teaching a single trick cleanly, with no equipment, using reward-based training.Read the guide guide“Touch”: teaching your dog to boop your handPutting his nose on your hand (or a stick) often serves as a first link: a versatile move for building any sequence.Read the guide guideTidy up your toysThe very first chain to try at home: putting his toys away, starting from the end and working back to the beginning.Read the guide guideGo see & fetchA concrete example of a chain that works: he spots the object, goes to get it and brings it back, all in one smooth sequence.Read the guide guidePositions and sending your dog at a distanceThe logical next step once chaining is mastered: the same links, but asked for from a distance.Read the guide

Frequently asked questions

What is back chaining?

It's a method where you teach the last step first, then work backwards towards the start, link by link. Because the end is the best known part, your dog is always heading towards what he knows, which keeps the whole chain stable.

How do I teach my dog to link several tricks together?

First check that each trick stands on its own, on its own cue. Strengthen the last link, then add the second-to-last one while rewarding only at the end, and keep working backwards one link at a time. Never add a link while the previous one still wobbles.

Why does my dog drop out in the middle of a sequence of tricks?

Nearly always because one link isn't solid enough on its own: it collapses once it is put into a chain. Break it down, work on that link separately, then reattach it to the rest of the sequence. Never push and never force your dog to "put him back in position".

Should I reward every trick in the sequence?

At first, yes, you can mark some of the links along the way. Then you fade that out towards a single generous reward, right at the very end. It's that final reward that closes the whole sequence, and running the chain becomes a pleasure for your dog, almost as much as the reward itself.

How long does it take to teach a dog a routine?

Think in weeks, not one evening. A 2011 study on laboratory dogs shows that a spaced-out rhythm, one or two sessions a week and one session a day, leads to faster progress than daily training repeated several times over. Aim for two or three successful run-throughs per session, then stop.

How do I get started in canine freestyle with my dog?

It's a choreography danced to music, which is really just one long chain built in small blocks, starting from the end. Work on a non-slip floor and keep jumps and twists to a minimum: to be avoided for a growing puppy, an older dog, an arthritic dog or an overweight one.

Is it better to teach the first or the last step first?

The last one. That's the whole point of back chaining: your dog moves from the less familiar towards the more familiar. Start at the beginning and he moves away from what he knows, so the end of the sequence eventually falls apart, whereas starting from the end keeps it solid.

Loading your progress…