Clean Run - June 2017 - 52

Keeping Track of Your
Dog's Learning
We're not going to discuss the "how" of
training here, but rather keeping track
of the training. There are three important assumptions I am making about
your training process:
1. You are breaking down the
behavior you want to teach into
small pieces;
2. You are rewarding your dog for
correct behaviors; and
3. If you have two failures in a row
you are either going back to an
easier step where you had success
or further breaking down the behavior to insure success.
Working on these assumptions, you do
not need to constantly be writing down
how many repetitions your dog is doing
and what his success rate is. Most of us
have a sense of when it is time to move
on to the next step. It is at this point
that you check for an 80% success rate
and keep track of the results.
The easiest way to do this is in a set
of ten repetitions for two consecutive
training sessions. These sessions do not
have to be on the same day. As a matter
of fact, they don't even have to be sets
of ten. When I trained my Beagle and it
was time to test behaviors to see if she
was ready to move on, I trained in sets
of five. Obviously, to get to twenty repetitions at an 80% rate, I had to do four
consecutive sets of five, instead of two
consecutive sets of ten. If you get 80%
on day one, and 63% on day two, and
83% on day three, you do not have two
consecutive sets of 80% or more! But
83% on day three and 90% on day four
is indeed two consecutive sets of ten.
(Remember: the reason for two consecutive sets is simply to be sure you didn't
get lucky when your dog was having a
really good day.)
The other thing that is important when
testing for 80% is understanding how
percentages work. For those of you
who are good with math and percentages, you can skip this. For the rest of
us here is a simple question: If I have

ten treats in my hand, and in a session
I give my dog those ten treats for correct behavior, what is my percentage of
correct behaviors? If you're like me and
sometimes a bit math challenged, your
immediate answer is 100%. And sadly,
like me, you would be wrong. Why?
Because yes, you gave out your ten
treats, but in doing that how many mistakes did the dog make that you didn't
count? Your percentage isn't ten out
of ten, it's ten out of thirteen if the dog
made three mistakes, which is 77%.
So when you are testing your dog to
see if he is at 80%, you need an easy
way to count mistakes, assuming that
you have ten treats in your hand to
mark correct repetitions. Some ideas
are getting poker chips, dominos, or
marbles to use to mark mistakes. Just
stick the marbles in your pocket and
every time the dog makes a mistake
you drop a marble on the ground in a
designated spot. When you're finished
with your set of ten correct repetitions,
you count your marbles and figure
your percentage from there. I actually
have a little cheat sheet made up for
myself: 10/11 = 91%, 10/12 = 83% and
so on. I really don't care about anything worse than 10/12 because it isn't
going to hit my 80% correct level.
If you're rewarding with toys/tugging
instead of food, you may need two different things to make your counting
easier. Maybe put ten dominos in one
pocket for the correct repetitions and
have marbles in the other for the mistakes. For each correct repetition drop
a domino; for each incorrect repetition, drop a marble. (Aside: I really like
dominos because they are easy to see
on dirt or grass!) When the ten dominos have all been dropped, count your
lost marbles. If all your marbles are still
in your pocket-hurray!
That's really all there is to it. But something so simple gives you so much
information about your dog's training,
why wouldn't you do it? Remember,
you do not have to do this all the time,
only those times where you are training something that requires sequential
steps or those times where you think

2o/2o, but didn't
wait for release

3 off 1 on

NO! NO!
3 on
1 off

NO!

All 4 on

Yes! NO!
2o/2o, but
waited for
release

Stopped
above
contact

Train Your Dog in
Black and White
When your dog is asking a
question-such as "will this
behavior pay?"-your training
will go faster if your answer to
the dog is clear. Your answer
comes in the form of whether or
not you reward the behavior.
Being clear about what you
expect is fairly simple: define
your behavior! For example, I
want a 2-on/2-off performance
on the dogwalk and for my
dog to wait in position until I
release him with a verbal cue.
Anything else-stopping with all
four feet on, stopping but selfreleasing, stopping with all four
feet on the ground, stopping
above the contact and waiting
for more information, etc.-is
non-rewardable. By rewarding
only the correct behavior and
ignoring all other behaviors you
avoid introducing shades of gray
that create even more questions
in the dog's mind and often
result in slow, uncertain, and
unreliable behaviors.
you dogs has a skill-such as backsides on verbal from 15 feet-and you
want to know how well the dog understands the behavior. Then it's easy to
set up a test for your dog so you know
where to go next.
Happy training! D

Elaine and Bob Coupé live in Oakland, Tennessee with three Border Collies and a Beagle. Kate ADCH, Dew, and Tug spend as much time as possible trying to train the
humans. Brassy Beagle ADCH prefers to eat and nap. You can find Elaine at memphisagility.com.
52

Clean Run | June 17

http://www.memphisagility.com

Table of Contents for the Digital Edition of Clean Run - June 2017