Maple hates to travel.
She hates riding in a vehicle and she doesn’t like being away from
home. Through Novice and Open, she
mostly moped through the exercises. We
got our Qs with scores in the mid-180s to low-190s. This is a dog that loves training, and, in
her own yard, looks like OTCH material.
In Utility, things got worse. Not only was she unenthused, she became less
and less confident. She was even doing
things that seemed deliberately wrong, like veering drastically off course to
get glove #1 instead of #3. It dawned on
me that maybe her problem was not entirely a travel issue, as I had been
assuming. Maybe, without food
reinforcement after an exercise, she was unsure of whether she was doing it
correctly and was actually experimenting to see if doing something differently
got a treat.
Now, if you had asked me, I would have said that, yes, we
did training sequences without treats.
Thinking carefully about it, however, I realized that I rarely did more
than two or three exercises in a row without a treat and, more importantly,
Maple never knew in advance that she was not going to get treats after each
exercise. I did a lot of jackpotting, in
which I would randomly give a big reward.
I tried an experiment at home. I did run-throughs in which I would go
through a warm-up outside my ring area, with a couple of treats, followed by a
phrase that would always mean that we were entering the ring and there would be
no treats until a big reward at the end.
The first time or two, she did things well, although she would become
increasingly uncertain by the end of the run.
She soon figured out the pattern and started making the same odd
mistakes and showing the same uncertainty she had in the ring. For example, she’d get the correct glove,
then stop halfway back, apparently thinking she must be wrong because no treat
was going to immediately follow.
I know many trainers would conclude she would only work for
food. Of course that was true. Few dogs work only for praise, but I think
there was more to it. She was working
for a treat at the end. What she did not
understand was the concept of a delayed reward and that the reward would be
proportional to what she had done. If a
person is not gambling for a living, he expects that his pay check will be
bigger if he works more hours. I think
dogs are fully capable of understanding a proportional, delayed reward. My dogs, who do not care that much for mice
(they are not terriers!), will sometimes expend some effort to chase voles if
it easy and there is nothing distracting them.
They will work obsessly for HOURS trying to scare a rabbit from under a
deck. The rabbit chase (and sometimes the catch) is a huge reward for them and
they are willing to expend a proportionately larger amount of effort to get it.
I had never really taught Maple that there could be big
delayed rewards in training. I think teaching delayed rewards is not the same
as teaching according to the gambling theory that is so prevalent in dog
training. That’s the idea that, if you
randomize rewards, and sometimes make it a very big reward, the dog will never
know when the jackpot is coming and will therefore work hard when no reward is
in sight because they think there might be a random big reward.. I think a dog would have to be dumber than a
box of rocks to not quickly learn that, in a real trial, there are no
unexpected jackpots. The reward is always at the end.
Strangely, I had never had the ring issues with my other
dogs that I had with Maple. Alder is a
happy dope, who loves being with me in the ring. He wants the food reward, but he considers
praise and play to be great, too. He
doesn’t have it in him to do Utility, but it’s not because of the reward
issue. Camas, my bygone mixed-breed
girl, whom I started training for Novice when she was nearly 10 yrs old, VERY
quickly learned that a complete run-through meant a big reward at the end. It probably helped that she tended to gag on
small rewards fed at my hand level and she preferred the one big payoff to a
sequence of small rewards.
I think another factor with Maple was the time I spent on
fronts and finishes. Because she is my
first real OTCH prospect, I spend far more time on F&F than I ever have
with any other dog I’ve trained. Because
“straight” is not a concept easy for most dogs to grasp and F&F is
intrinsically motivating, it takes a ton of repetition and a lot of treats to
keep attitude high for all those reps. I
think I really drilled Maple to expect treats regularly during training. I also used lots of random jackpots and, I
think, reinforced the idea that the value of the treat was random and not
proportionate to the amount of work done.
I am well aware that there are a lot of animal behaviorists and
theorists that are strong proponents of the random reward and they get good
results with a variety of animals. I may
be totally wrong and may eventually totally recant these ideas.
In the meantime, with Maple, I plunged into the deep end of
the pool on teaching delayed and proportionate rewards. I would make it very clear when we were
starting a run-through with a delayed, big reward at the end. My goal was to
get her working briskly without the expectation of an immediate reward and to
make her understand that a lack of reward did not mean she had made a
mistake. Initially, I ignored all the
fine details, like straight fronts and finishes, but I absolutely insisted
that, on each exercise, she HAD to work briskly. If she moped on any step of the heel, I
backed up, put the leash on, and insisted she look up and move briskly. Then, I’d take the leash off and start
again. Any moping, the leash went back
on. Repeat until she finally realized
that she needed to move briskly and keep her eyes on me. On gloves, if she got the right one, but
stopped, uncertain, on the way back, I’d put the leash on, have her hold the
glove and back up saying “Front, good girl”.
Then we’d do it again. And again, and again, until she’d get the right
glove and come back without hesitation.
We’d do the entire utility routine with no treats and without moving to
the next exercise until she had completed an exercise with alacrity and no
uncertainty (but not caring about F&F).
The first couple of times, it took us as long as 45 minutes to get
through a whole routine. Then
understanding finally clicked in.
I stopped giving random big jackpots. For individual exercises when I’m teaching,
she gets only small parceled rewards of dog treats. When I tell her that this is a run through,
she knows nothing is coming until the end, but it will be a big payoff, like a
bowl of cut-up hot dogs or pieces of meat.
I say “It’s our turn. Show me what you can do.” I go through the
ring-entry sequence of handing a leash to an imaginary steward and talking to
an imaginary judge.
She was a changed dog in the ring by the end of summer. Sure, she made lots of NQ mistakes (those
articles are killing us), but she looked good doing it. For the most part. She still has the travel phobia, and it still
causes problems. She tends to get dog
show diarrhea and an upset stomach from the stress of traveling. (I bring water from home and give her the
same diet. It doesn’t help.) So, sometimes, if she’s not feeling well, she
gets lackluster, but she still tries.
Last fall, after a miserable trip back from Wenatchee, during which she
panted for 5 hours straight, I went to the vet and got a prescription for
Valium for the long travel days. The
initial results were positive. She was a
lot calmer on the long drives for the last couple of shows. My hope is that will begin to associate
traveling with (Valium-induced) pleasure and it will improve her travel
attitude, long term.
I wouldn’t recommend the total immersion method of teaching
delayed rewards. I should have started
early, during Novice training as a puppy, stringing together 2, then 3, then 4
exercises. Live and learn.
On to 2015 resolutions.
No comments:
Post a Comment