 |
-
Be sure to raise the criteria slowly enough so that
the dog is sure to succeed. Raising the criteria should
be in relation to the level the dog already offers you
independently from time to time. This means that if for
example the dog heels sloppily but every now and then
correctly at the correct spot on your left side close
to your leg, the next thing to teach is to only reinforce
the best behaviors. The dog will understand that even
when the rules become a little harder, it is still possible
and quite easy for him to earn the click and treat, with
just a little more effort.
-
When you are teaching something to the dog, reinforce
only one criterion of the behavior at a time. For example
when teaching the correct show posture, if you first
do not click when the head is not right and next time
do not click because the feet are not right, the dog
cannot understand what he did wrong. One criterion at
a time!
-
Before you raise or increase the requirement level,
maintain the already-learned level by random reinforcement.
When the dog has learnt the behavior (the behavior is
under a cue) you should start reinforcing it only occasionally.
This is an important principle of the shaping process.
Then you can choose only the best performances of the
behavior to be reinforced, which again will drive the
performance towards perfection.
-
When you add a new criterion to the behavior lower the
requirement level of the already-learned criteria. If
for example the dog does not perform equally well when
you start training in an environment with more distractions,
return to the lower level so that the dog can be successful
again. Or when teaching show stacking - when the dog already
knows the correct head position and you want to add the
feet, do not feel bad if the dog at first forgets the
head when he is busy concentrating on the correct feet
positioning in order to get the click and treat. This
is only temporary - what he has learned he will remember,
but under the stress of increasing requirements an old
learned thing can be forgotten for a while. When he learns
the new criterion, the other parts of the behavior will
return to the learned level also.
-
Make yourself a plan on what to reinforce next. This
is in case the dog progresses faster than you had expected.
Sometimes the dog makes a sudden breakthrough in the
learning process and rapidly progresses to the right
direction. In these cases it is very important to be
able to raise the criteria immediately. If the criteria
remain the same the dog might get bored and there is
a possibility that he will not co-operate that well
in future.
-
Do not change the trainer during learning of a behavior.
The dog may have many trainers but only one for each behavior.
Different people have different habits, requirements and
expectations for progress, which may cause the behavior
to deteriorate or even disappear until the dog and the
trainer have reached the same wavelength.
-
If you cannot establish progress with one form of shaping,
try another one. For any behavior there are as many
ways to shape it, as there are trainers to invent them.
-
Do not end a training session without a reward. When
training, the trainee and the trainer "make a deal"
and if the trainer breaks the deal by ignoring the dog,
for example by starting to talk with a by-passer or
just daydreaming, the dog takes this as a punishment:
he does not get a reward even when he did the behavior
and cannot understand why.
-
If a learned behavior declines, revise the shaping
process. This is done right from the beginning, step-by-step
but in a faster mode with a couple of reinforcements
by each level to refresh the memory.
-
Finish the training session when the animal has made
a breakthrough. The dog will remember the last thing learned.
It is a general mistake to not have patience to finish
the session when the dog does the behavior really well
(breakthrough) but to insist him doing it again and again.
This will make the animal bored and the behavior deteriorates.
But if you are patient enough and stop training and try
this behavior the next day again, the performance is not
just equally good but often even better. Training should
always be finished on a high note.
As with all training, train only
short periods at a time to avoid boring the dog. Over the
same period of time you can teach the dog several behaviors.
I have noticed that with a Bloodhound it is much easier
to focus on one thing at time to get the behavior fluent
and under cue before starting to reinforce another behavior.
Use also so called jackpot-method
which means that you have some extra special treats for
rare occasions. These can be given for especially good performances
or when the dog makes a breakthrough, but also sometimes
randomly for a nice performance of a lower criterion level.
The principle of randomizing makes practicing more surprising
and enjoyable for the dog. An easy way to motivate the dog
is t give a jackpot for a nice performance right in the
beginning of the training session. This makes the dog try
for another jackpot, i.e. to work harder (compare this situation
to a person on a slot machine - if he wins on his first
try, it is more likely he will try again and again to win
more money).
The most important principle: what
you reinforce is what you get. If behavior starts to go
to wrong direction, you should look back and find out what
exactly were you reinforcing. Here is one experience I had.
I was teaching my Bloodhound to turn her head to the left
(she did this very well already) and then I decided to change
direction and try right. The first couple of times when
she turned her head to the right she happened to move juts
a bit backwards while sitting so after a few repetitions
I noticed that she had moved so far from me that I could
barely reach her to give the treat! Without noticing it,
I had been reinforcing her moving an inch backwards while
sitting while I thought I was reinforcing her turning her
head to the right!
|
 |