Improvisation in Dog Training: Lucas’s Service Dog Tuck

One of the most undervalued must-haves in a dog trainer is the ability to improvise effectively.

Almost anyone can follow a recipe to get a basic behavior.  If it’s a good recipe, even a novice dog trainer will be successful a reasonable amount of the time.

But what really separates the wheat from the chaff is what a trainer does when the standard recipe isn’t getting the job done.

Do they escalate to more risky methods?

Do they keep pushing something that obviously isn’t working?

Or do they find a way to improvise on the recipe so it makes sense to this particular dog?

Full disclosure: My training mechanics could be cleaner.  I’m slower to dispense treats than I’d like.  My patience with proofing is frustratingly low, especially with my own dogs.  But the place where I shine as a trainer is coming up with creative work-arounds when something doesn’t go the way it’s supposed to.  I have learned to love the opportunity to troubleshoot a recipe gone wrong because it always, always teaches me something valuable.

Teaching a service dog tuck: Lucas style

Two years ago, I borrowed my friend’s collie Lucas for a weekend.  Lucas is now a working service dog, but at the time, he was still in training.  His tasks were well under way and his public manners were already excellent, but there were a few specific behaviors that his handler was struggling with.

One of the goals was to teach him a service dog tuck.  This is when a service dog squeezes behind a person’s legs when they’re sitting at a chair to take up less space.

At 25.5″ tall and ballpark 75 pounds of rough collie, Lucas is not a small dog.  In crowded spaces like restaurants and public transportation, handlers often have their service dog tuck behind or between their legs to keep them out of the way of other patrons as a courtesy.  That’s even more important with a large dog like Lucas.

I had taught a service dog tuck behavior several times, most recently a week before my session with Lucas, and was pretty confident that it was going to be a quick fix.  For the dogs I’d taught before, it was a very straightforward two step process: (1) lure the behavior, and (2) fade the lure.  As easy as sit.

Except it wasn’t easy for Lucas.

Lucas wasn’t at all confident that his body could fit through the small tunnel between my legs and the chair.  And he was even less sure he was comfortable with the feeling of his body being lightly squeezed between my leg and the chair.  He felt pretty squicky with the whole situation — not frightened, just a practical “Nope, I’m going to get stuck if I try that, better not,” type of response.

Well, crap.

At the time, that was the only recipe I knew to teach that behavior with luring.  I could have shaped it from scratch, but Lucas was relatively new to learning with pure shaping.  Shaping the service dog tuck would have been frustrating and stressful for him.  I needed a way to teach it to him using luring or capturing, at least for the first step.

Step one: Find a starting place that is “stupid easy” for the dog.

I decided that his biggest problem was the amount of spatial pressure. So I went back a step. Could Lucas go between my legs and a chair if I was standing up, so that I was crowding him less?  Grudgingly, yes, but he was still obviously skeptical about the contact along his sides, and it was going to be very difficult for me to gradually change my position from standing to sitting over the course of the repeated trials (think slow motion squats).

So I went back several more steps to find a stupid-easy entry point. Could Lucas go between two pieces of furniture with plenty of room to maneuver if I was completely out of the picture?  Yes, with enthusiasm.  Perfect!  That’s our starting point.

Step Two: Show them that “stupid easy” pays.

So for several minutes, Lucas got treats for going through a wide channel between a chair and an ottoman, going clockwise around the ottoman.  It looks absolutely nothing like a finished service dog tuck behavior, but at this point, we’re just looking for an entry point.  Clockwise circles between furniture: Accomplished.  Lure: Faded.  Confidence level: High.

Now the dog has a baseline incentive to stay in the game and work with you as you approach the levels where the dog is actually struggling.

Step Three: Change tiny variables in the direction of the full behavior, one variable at a time.

Think of this like a game of degrees-of-separation.  Your goal is to change one small detail at a time until you have connected the starting place with the end goal.

Once Lucas was offering this with confidence, I moved to sitting in the chair, but let him pass in front of my legs between the two pieces of furniture on the same path he’d been on before and still probably twice as wide as his body so there was very little spatial pressure.

Slowly, I moved the ottoman closer so that he had to lightly press against either my legs or the ottoman, and later both.  Clockwise circles with light body pressure: Accomplished.

But for the final service dog tuck, I wanted him behind my legs, not in front.  Once he was getting more comfortable with circling in narrower spaces very close to me, I put my legs up as a bridge, laying between an ottoman and a chair with my knees high enough that a full-grown 25.5″ tall collie was passing under my legs without touching them.  (And yes, this position was exactly as uncomfortable as that description makes it sound.)  Gradually, inch by inch, I lowered my legs until I was sitting normally with my feet on the floor and he was belly-crawling to continue to successfully go under my legs in increasingly small spaces.  Clockwise circles with light body pressure while belly-crawling under my legs: Accomplished.

Step Four: Fade out any props

I faded the ottoman further away so that it was just me, the chair and Lucas.  And suddenly he was happily crawling through the gap between the backs of my legs and the front of the chair, totally confident, completely okay with his sides/back being touched and quite pleased with himself to boot.

The only thing left to do was to teach him to pause for increasing amounts of time when he was under my legs, and finally to hold a stay in tuck position. And then alakazam! a large dog who can’t do spatial pressure suddenly has a totally functional service dog tuck for public access.

In dog training, slow is fast and fast is slow.

My usual training plan for that behavior has two steps.

The version I wrote up for this session with Lucas had seventeen.

But, the whole process still only took about 30 minutes from start to finish to teach the finished behavior.  That’s almost exactly how long it took to completely fade out the lure and add cues for my own dog when he learned the same behavior — the time was just allocated very differently.

People usually expect the micro-management version to take substantially longer than the broad-strokes way, and it didn’t.  And it almost never does, if you’re careful with your criteria-setting.  By the end of half an hour, both dogs were consistently and comfortably performing the same finished behavior, and I don’t think an onlooker would have been able to tell which dog learned with two big steps and which dog learned with seventeen tiny ones.

Learning how to create a new training plan or troubleshoot a recipe is one of the most useful skills a dog trainer can have.  Because I was able to adjust criteria to meet the needs of the dog I was working with, Lucas quickly learned the service dog tuck that he was struggling to master — without having to escalate to more risky training methods or wasting time trying to force a square peg into a round hole.