Reliability Deficieny Type A: Inadequate Training

This week, I'm focusing on reasons dogs don't listen to cues for "known behaviors." Today, we'll discuss inadequate training, as this is by far the...

Casey Lomonaco  |  Jul 12th 2010


This week, I’m focusing on reasons dogs don’t listen to cues for “known behaviors.”

Today, we’ll discuss inadequate training, as this is by far the most common reason dogs don’t respond to cues for behaviors.

I remember attending my first agility class with Mokie. I was very excited, and proud to show off how smart my girl is. I was flabbergasted and, frankly, embarrassed when she failed to respond to my cue for a behavior I thought she “knew” quite well…sit.

My agility instructor (who is now a friend and colleague) chuckled at my frustration. She asked me, “Has she done it 5,000 times?” Ummm…no?! “Then she doesn’t know it!”

Last year, I attended ClickerExpo in Providence. I had the chance to chat with a colleague who specializes in clicker training dogs for police and military service. Bomb detection dogs can’t be “semi-reliable,” lives are, quite literally, “in their noses.” I asked him, “When do you think a dog actually knows a behavior?” “After he’s done it 10,000x,” was the response.

He also said that to train to a level where most pet owners were satisfied (65%), it took approximately 5,000 repetitions of a behavior in at least 20 different environments with at least 20 different distractions.

*insert picture of dog owner with eyes bugging out of her head, feeling hopeless, here*

I’ve got to admit, even for a dog trainer these lofty goals sound daunting.

WHY DOES THIS HAPPEN?

Why do dogs need this many repetitions? The answer most trainers will give you is that “dogs don’t generalize well.” What this means, in practice, is that a recall in your quiet living room is vastly different to your dog than recalling away from play at the dog park or mid-deer chase. A living room recall is a kindergarten level behavior, a dog park away-from-play or mid-deer-chase recall is a PhD level behavior. While your dog very well may “know” to respond to your recall cue in your living room, you will have to teach him that the cue is a green light for reinforcement in a variety of situations before you can count on the behavior in new environments.

When I heard these numbers tossed around, I was understandably overwhelmed and a little discouraged. Yeesh, how long would this training take?

YIKES! HOW LONG DOES ALL THIS TRAINING TAKE?

In actuality, surprisingly little time. I did a test with Mokie on an easy behavior, “sit.”

After a bit of practice, I was able to get ten repetitions of the “sit” behavior in 30 seconds. I like to advise my students to practice behaviors in multiple, short training sessions throughout the day. I figured if I practiced 10 sits, 3 times a day, I would have 30 sits a day and this would take me approximately 90 seconds.

If I did this seven days a week, after one week I would have 210 sits by the end of the week.

After a month, I would have approximately 900 sits registered in our “sit account.”

In order to get those 5,000 repetitions needed to achieve 65% reliability, it would take me approximately 167 days or about 5.5 months. Hey, that’s not too bad!

To get to the ultimate reliability goal, we would double the amount of sessions necessary, so we would probably be looking at 11 months to 1 year worth of dedicated training, perhaps slightly longer depending on our ability to train daily and consistently.

A part of me really does believe if you don’t have 90 seconds a day to spend training your dog, perhaps you should consider a pet rock as an alternative next time you’re on the market for a companion. If you have six behaviors which are important to you (let’s say sit, down, recall, leave it, name response, and targeting), that amounts to approximately 9 minutes of training per day. Looking at it this way, it’s not like you have to quit your day job to have a reliable, well-trained family companion.

Remember, this is 5-10,000 opportunities to cement your relationship with your dog while strengthening the reliability of his behavior. Rome wasn’t built in a day, nor is good behavior. I would venture to say that even after those 10,000 repetitions, I would occasionally still reinforce my dogs with a special treat for a particularly spectacular effort. If the slot machine never pays off again, the game is no longer fun. This is not a conflict of interest for me, as I am accustomed to feeding my dogs every day anyway.

If we assume that the average dog will live to be ten years old, investing a year in training important behaviors to reliability is really only investing 10% of your dog’s life span to lay a foundation which will improve the other 90% of your time together. The average parent of a human child has to “train” her child for 18 years, far more than 10% of a child’s lifespan (unless you have an exceptionally healthy child that lives to the ripe old age of 180). Every minute we spend complaining about a dog that does not respond to a cue is a minute we could have spent training him to do the right thing.

Think of this next time you think your dog “knows” a behavior. Ask yourself, “has he done it 5,000 times? 10,ooo?” If not, your definition of “knowledge” might be vastly different from your dog’s. Practice makes perfect!

Stay tuned for tomorrow’s installation on “reliability deficiency” – yucky reinforcers.