Obedience and agility require different behaviors, but both sports share some basic challenges that create grief for handlers.
Both sports require focus and impulse control without a leash and with significant distractions present. Agility competitors struggle to maintain connection when the dog is working at a distance under speed, often with a good deal of excitement going on in the rings around them. Obedience competitors struggle to maintain connection when the dog is working under the pressure of silence and formality for long stretches of time. And both sports require a balance between handler connection and exercise (or object) focus.
Yet, the culture of the two sports is quite different in a fundamental way. Agility handlers are trained to take responsibility for their dog’s failures (they are directed to change THEIR behavior), and obedience handlers generally hold the dog responsible (they are directed to change THE DOG’s behavior). Which is not to say that you won’t find significant exceptions to both of these philosophies, especially at the lower or middle levels of competition, but if you pick up a book, magazine, or online chat, you’ll understand what I mean.
Here’s a classic example:
Agility – “If the handler makes an error, reward the dog! He did what he believed was correct! If the dog makes an error, reward the dog! He did what you trained him to do!”
Obedience – “If the handler makes an error, try again! The dog has to earn those cookies! If the dog makes an error, withhold reinforcement or fix it!” How you fix it depends on your approach to training. But reinforcement for errors? I don’t know anyone in obedience who actively proposes that solution, yet it is very common in agility. A consolation prize, of sorts. The “screw up” cookie.
It’s not that agility people are morally superior, or nicer, or like their dogs better; it’s that the sport of agility requires a highly confident dog who has no fear of making an error, because worry impacts speed. When the winner is usually much less than a second faster than the next dog in line, it’s critically important that the dogs believe they are absolutely correct. Indeed, you will often hear agility instructors tell their students never to let their dogs know when they have made an error. Simply continue on.
Obedience? Not so much.
People say the two sports are different so cannot be compared, but these are issues of training specific behaviors, not of competition. One dog doing an obedience directed jump for Utility and another dog learning an agility sequence are practicing the same skill; take the jump you are directed to. Yet even in exercises with the same base behavior, errors are approached differently. If you do some reading on email lists or magazines you’ll quickly get what I’m talking about.
There are exceptions in both sports and I represent an exception in obedience. For a long time now, I have rewarded dogs for my handler errors in training because I see what happens to the self confidence of dogs that are held responsible for handler error. They deflate, especially the less driven ones.
But I have not rewarded dogs for making errors when I believed I handled or trained correctly. And now I am re-thinking that.
Two weeks ago I wrote a blog on rewarding a dog who was wrong when working scent articles. You can see it here. http://denisefenzi.com/2015/03/09/scent-articles/
And I saw more response to that training blog than I have seen in months! People tried it and dogs that have had long standing issues with scent articles turned around almost instantly. Which doesn’t mean they didn’t continue to make errors, but now they were still in the game! Willing to try! If a dog is in the game then you can train through problems, but when a dog opts out, you’ve got a much bigger problem. Dogs that give up simply cannot be trained using positive methods, and what fun is it to make a dog do things that they do not want to do? And how well does that approach to training hold up in the ring with a dog that is totally disengaged and lower drive to begin with?
But this reality brings up more questions than answers for me. For a long time I have used a Cheerful Interrupter (CI) to address errors. The point of the CI is to interrupt the dog’s behavior while maintaining their enthusiasm and willingness to perform the task within behavior chains. CI’s have worked extremely well for me, and as a result I have almost completely removed No Reward Markers (NRM’s) from my training because I place such a high value on a motivated, enthusiastic and fully engaged partner.
But what if I take that one step further? What if I not only interrupt, but add a cookie to it? How will that affect the dog’s willingness to withstand pressure? And what happens if I do not interrupt the chain at all? What happens if I allow the dog to finish, reward and then set up the exercise in a manner that makes success more likely on the next attempt?
In agility it works. Consistently. If the dog goes over the wrong jump then the handler throws out a toy, works to understand the error in either handling or training comprehension, and then the team either tries again with a better plan or returns to more basic behaviors to find the weak link. So how about when my dog takes the wrong jump in Directed jumping? Could that work for me as well? Something makes me think…yes.
So where does this approach come into play? Not in the shaping phase – dogs learn the basic component behaviors by marking correct behavior with classic reinforcement (cookie or toy) and incorrect behaviors by waiting. Same in agility.
How about the next phase, where dogs are performing their known behaviors under distraction (proofing). Can you reward if the dog fails to perform correctly? Does it matter if the dog fails multiple times in a row? I don’t know but I have some guesses.
How about the next phase, where the dog is experiencing stress during the work, possibly as a result of the environment, or simply by virtue of being very sensitive? I think that is a place where dogs might be rewarded for error. But all dogs or only some dogs? I don’t know.
How about in the behavior chain phase – if you interrupt the chain where the dog makes an error, give a reinforcer and then start over, what happens? How about if you allow the dog to finish, reward, and then change the exercise to create a better chance of success? I don’t know.
What are the relevant factors to consider? Should the Consolation Prize Reinforcer (CPR) be of a significantly lower value? For example, a cheerio vs a piece of chicken? I don’t know.
Should the CPR be offered with a significantly different emotional response from the handler? A piece of chicken and a party for a correct repetition vs. a piece of chicken and a chance to do it again for an error? Should the CPR be one piece of chicken whereas the reinforcer for correct work is several pieces of chicken?
And how much do each of these factors vary by the given dog and their stage of training?
I don’t know those things either. Of course, I have a lot of guesses about all of this, and I have two fine subject dogs to try stuff out on. So we’ll see what happens.
If my part of training is to communicate expectations, and if his part is limited only to learning what I teach, what are the possibilities?
I’d like to find out.
And on an unrelated note: Fenzi Dog Sports Academy has a new term starting on April 1st! Registration starts March 22nd; some classes fill very fast so be ready to register at the correct time for your course.
This term I’m teaching Heeling Games and Bridging the Gap between Training and Competition. You can see the full schedule and read more about your choices here: http://www.fenzidogsportsacademy.com/index.php/schedule-and-syllabus