October 31, 2019

A marker can be any sound that has been paired with a primary reinforcer such as food so that it can precisely indicate to the dog that what it has done at that point in time will earn reinforcement.  A marker can be a word or a sound such as a whistle or a clicker. 

What anyone who has done training with food rewards knows is that it works.  Dogs who get treats for doing things tend to do those behaviours more often in the future.  However, just dishing out cookies for stuff that the dog does is imprecise.  Let’s take a sit for example.  The dog sits and you give it a treat.  It sat, right?  But what exactly did your dog associate with getting the treat?  She knows she got a treat and can guess it was for something she did in the last several seconds. But there can be a lot of choices in this time frame.  It could be that she was looking at you before she sat; it could be movement of her feet; what about standing up immediately her rump hit the floor?  Given enough repetitions she would probably work out which of these options is present most often.  And in the end, she would probably figure out that sitting was what you wanted her to do.  

But the question now is what did the sit look like?  Was it a nice square sit?  Or did she sit on her hip – a sloppy, puppy sit?  She might start out with a square sit but quickly collapsed into a sloppy sit which just happened to be the time when she got the treat!

On the other hand, if you use a well-timed marker, she will know that it’s about her rump hitting the floor because each time she does this the marker always happens just then and the marker predicts a reward is coming.  This is much clearer communication with your dog.  Consistency and Clarity.     

Behavioural Inhibition

February 28, 2017

One of the most enjoyable things about training with positive reinforcement is that as soon as the dogs sense that a training session is about to start, a kind of buzz happens.  They can be fast asleep or lolling about in the sun, or just sitting gazing out at the garden, but as soon as they hear me getting my gear together (and they hear it – no matter how quiet I am), all of them are standing at the door, tails wagging and eyes bright.  This is because they just love training – training is not a chore it is a game we play together.  It gives them an opportunity to have a conversation with their trainer, to problem solve, to earn reinforcement.

Using positive reinforcement with a bridging stimulus has another advantage.  It gives the dog freedom to experiment with behaviours.  This is a trait that is not present in any dog that shows signs of Behavioural Inhibition.  There is nothing sadder than a dog that is so anxious about doing the wrong thing that it ends up doing as little as possible.  It becomes so inhibited that teaching it anything outside its experience is really difficult and time-consuming.

Some features of a behaviourally inhibited or shut-down dog would be a resistance to trying anything new or foreign to its experience.  A 14 week old puppy that has been positively trained from the get-go will have absolutely no problem rapidly offering behaviours such as climbing on to a raised object, offering position changes voluntarily, readily taking food rewards, making eye contact and so on.  The inhibited dog will, on the other hand, even when lured on to a raised object, will, after many, many sessions be very hesitant to try it and will often start the session simply sitting or standing and doing as little as possible.  This same dog will take many, many sessions of luring the down before offering the position, and even then will lower itself cautiously without making eye contact as though not sure what the consequences will be.

It is possible to make things better for these sad little dogs.

  • Identify the situation that tends to activate behavioural inhibition and avoid that situation. Change the environment.  If the dog is unhappy in the area that you train, move somewhere else where the dog is more comfortable.
  • Identify anything that the dog will participate in willingly. And gradually use this participation to ease the dog into working more confidently towards goals that you have set.
  • Observe the dog closely and avoid anything that causes the dog to shut down. Once the dog becomes more confident these things can gradually be reintroduced.

Exercise LOTS of patience.  Even a hint of frustration or irritation will lead to a shut-down dog.  End the session immediately you start becoming frustrated.

Taking Your Dog’s Behaviours to the Next Level

November 2, 2013

If you have just started the adventure that is clicker training you are probably having such fun clicking and treating all the behaviours that are being “thrown” at you that you haven’t given much thought as to how to make the behaviours reliable and “fluent”.  In fact you have just entered the first stage of the learning process.  Any learning of a physical skill is more than just experimenting to see which behaviours can earn reinforcement.  It also includes:

  • Developing physical proficiency,
  • Adding cues and contingencies
  • Adding duration, distractions and distance
  • Developing the ability to perform under stressful conditions.

In order to cover all these bases thoroughly, you need to be pretty systematic in your training.  One of the ways of doing this is to break your training into three training stages.

Stage One:  This initial stage consists of formal sessions of more than one repetition in a non-distracting environment.  You will gradually, with shaping or successive approximation,  be building the behaviour you have visualised and when this is done adding a cue.  This is the stage where every single correct repetition is reinforced.  Your rate of reinforcement should be high enough that the dog is learning rapidly without disengaging.  No corrections at all.  Any “mistakes” are simply not reinforced.  If there are too many “mistakes” then you are raising criteria too quickly and need to go back a step or two.

To check whether the dog has learnt the cues and the basic behaviour is being performed to your satisfaction, look at the four corners of stimulus control.

  • The dog does the behaviour immediately upon perceiving the cue (bearing in mind that the cue does not always have to be a verbal cue).
  • The dog does not offer the behaviour without being cued
  • The dog does not offer the behaviour in response to some other cue
  • The dog does not offer any other behaviour in response to the cue

 Stage Two:  This stage is entered when your dog is performing the required behaviour to your satisfaction and will respond to your cue quickly and accurately.  You’re now ready to add elements like distance, duration and distractions.  You might even start generalizing the behaviour (see article on Generalization)  by “taking it on the road”.  You’re still aiming for a high rate of reinforcement.  No corrections if your dog guesses wrong.  Corrections at this stage of the learning process might slow down progress and this would be counter to what you’re trying to achieve.

Stage Three:  You will now start cueing the behaviour outside of formal sessions.  There is still room for formal sessions at this stage of the training, but now is the time to start introducing real life situations.  To do this you will only ask for the behaviour when your dog wants something.  You will start using environmental rewards for these behaviours such as asking for a sit before throwing a ball, or asking for eye contact before allowing him to play with other dogs; or asking him to wait at doorways and gates.  In this way you’re teaching your dog that if he does what you ask him to he will get what he wants.

Shaping Behaviour

June 1, 2011

Imagine you are in a country where people do not speak the same language as you.  Because you can’t understand the language, following verbal instructions are obviously very difficult.  Imagine being instructed to draw a straight line on a board with a piece of chalk when you can’t understand verbal instructions. Very difficult, I’m sure you’ll agree.   Your instructor’s job, without actually showing you is to break the job down into slices and build it slice by slice.  This is shaping.

In this example, the slices would be like this.

Slice 1:  Look at the board

Slice 2:  Move toward the board

Slice 3:  Look at the chalk

Slice 4:  Touch the chalk

Clice 5:  Pick up the chalk

Slice 6:  Touch the chalk to the board

Slice 7:  Draw a line on the board.

Shaping a dog to do something works the same way.  The dog is equally disadvantaged in that he can’t speak your language.  So you break down the behaviour you want into small slices and reward each slice, gradually building the behaviour by raising your criteria as you go.  For example, the bow would be sliced up the following way:

Slice 1:  Look down

Slice 2:  Bend elbows

Slice 3:  Put elbows on floor

The interesting thing that I noticed about this video, apart from the excellent timing of the click was the presentation of the food.  What we tend to forget when we use the clicker to mark behaviour is that the food also influences what the dog is learning.  Bob Bailey said about the primary reinforcer “As much thought should go into the delivery of the primary reinforcer as goes into the secondary, or bridge (clicker).  Generally, to make this work for you, Click For Action, Feed For Position.”   This is exactly what the handler in the video clip is doing and it is very effective in getting her dog to understand what she wants.

You can also use shaping to build on an existing behaviour and this is how distance, duration and distractions are incrementally added to an already existing sit, down or heel.

There are other ways of getting behaviours that I can think of.  Luring, Moulding, Capturing and Targeting.  However, none of these are as flexible as shaping.  It is limited only by the animal’s ability and the trainer’s skill.

Sometimes the best solution is to use a combination of luring, capturing, targeting and shaping.  Get the dog on the right track by luring or targeting, then free-shape.  As a dog trainer you need to be creative and use whatever tools are available to you.

Clicker Games to Improve your Skill

January 27, 2011

In a previous post on clicker training, I suggested that one of the ways you could improve your clicker and timing skills, would be to practice without the dog.  In this way you can make mistakes in your learning process, and it doesn’t have to affect your dog.

Some of these games need a partner or friend, but there are some that you can play solo.

Game 1:  Mirror, Mirror on the Wall.  – One of the most common mistakes  beginner clicker trainers make is to deliver the food reward at the same time as the click.  The dog is, understandably, so focussed on the food in your hand, that he simply does not hear the click and it subsequently loses its value as a secondary reinforcer (or marker).  This is most common when working at close quarters with your dog and is often the reason that your dog is taking longer than expected to learn a skill.  Even movement of your food hand during an exercise will distract your dog and slow down the learning process.  One of the big advantages clicker training has over lure and reward training, is that it makes the dog “think” about what it’s doing rather than simply “doing”.

This exercise will need a handful of dry beans or any objects that are similar and an empty jar or cup.  Place the handful of beans and the empty jar on a table in front of a mirror.  Stand facing your image and click and transfer one bean from the pile into the empty jar or cup.  What you will be watching for is ANY movement of your food hand while you are clicking with the other.  One thing that helps is to put your (empty) food hand behind your back.  Pick up on even the slightest movement.  You must practice keeping your body, and particularly your food hand still while you are clicking.

Game 2:  One-Armed Bandit – For this game you need a friend, partner or child to help you.  Stand facing one another, about 10 paces apart.  Instruct your partner to very slowly raise one of his or her arms from its starting position at her side, sideways until it is pointing at the ceiling – in other words her upper arm is next to her ear.  Your job as the clicker trainer is to click when her arm reaches the horizontal, at which time she must stop moving.  If she is moving very slowly your clicks will be on the button, but as she speeds up, you’ll find that your clicking is too slow, and her arm stops above the horizontal position.  If it were a dog’s behaviour you were marking you would have been too late to capture it precisely, and probably would be marking another behaviour altogether.  You can see from this game that is rather better to click too early than too late.  It’s better to catch the behaviour on the way to completion, than after completion.  Remember practise makes perfect and even when speeded up, you’ll soon be able to catch the position with a great deal of accuracy.

Game 3:  Anyone for Tennis? – For this game you need a tennis ball and a friend.  Your friend or partner will drop the ball to bounce on the floor, or throw the ball against a wall.  The object of the game is to click when the ball makes contact with the surface.  Remember that if you think “Click” as the ball contacts the surface, you will be too late.   The average person’s reaction time is o.3 seconds, which when you are trying to mark small behaviours is actually a lot of time!

To check your reaction time, you can play these games on the internet.



Happy clicking!

What is Clicker Training?

January 19, 2011

The clicker is a small hand-held tool which makes a clicking noise.  The clicker is simply an effective hands-free way of communicating with animals.  It is a very distinct, unamibiguous sound (CLICK!) that tells the dog that what he has done (or is busy doing) is right and there is a reward to follow.  The clicker bridges the gap between the behaviour and the reward. It is not a magic “quick fix”, although in the hands of an experienced clicker trainer it looks magical.

Using a clicker has quite a few advantages over just using food without the event marker.  The clicker is…

  • More precise at identifying behaviour or parts (splits) of behaviour.  For example, if you want to reward your dog for making eye contact with you, you can catch the precise moment that his or her eyes meet yours.  The dog then understands that it is that brief contact that is being rewarded.  If you train only with food and lures the timing is difficult, as the second you move your food hand to reward that momentary eye contact, the dog will look at the hand holding the food hand and looking away is what you are actually rewarding.
  • Allows for a time lag before delivery of the primary reinforcer.  Because you have associated the clicker with the presentation of food (the primary reinforcer), the dog understands that the click predicts that the reward is on its way.
  • Allows for treat delivery to be more flexible.  Depending what you are trying to achieve in your training the treat delivery can be while the dog is in position (if you are working on sustained positions) or if you are training for some activity such as going out to touch a target and returning to you, the contact is clicked, but the dog has to return to you to get the treat.
  • Allows for easy resetting of the behaviour.  When you are working for repetitions to build fluency, the treat can be tossed away from the dog so that the behaviour is reset,  The training session becomes more efficient.
  • Affects Demeanour.  Tossing the treat after the click can activate a passive dog whereas feeding in position can calm a busy dog.
  • Enables movement to be marked and reinforced.  For example, you can click the dog while it is moving towards you, and even though you are delivering the food when the dog gets to you, it knows that it is the movement towards you that is actually being rewarded.
  • Allows for distance building.  The clicker is useful when teaching the dog to sit or lie down at a distance or to touch a target at a distance.

Some of the difficulties a lot of people have with using the clicker is the timing.  If the click is even slightly late, you might be reinforcing some other behaviour than you thought you were.  You get what you click for.  However, this timing can be practised without your dog and we all know that practice makes perfect!  I will describe some of these games in a subsequent post.

Dogs, even dogs that have never heard the click of the clicker before, take to it incredibly quickly.  It makes teaching your dog a breeze and fun for both of you.  It teaches your dog to think – and think they do.  There’s nothing more exciting than to see your dog have that “Eureka” moment when it knows exactly what you are clicking for.

Clicker Training or Using a Marker

November 24, 2010
Clicker training is simply a way of communicating with your dog.  The special thing about clicker is that your dog understands what you are telling it.  You can tell it “Put your paw on the tile” a million times without any response, but clicking the dog for touching the tile with his paw and following the click with a treat, will get the job done in no time at all!  The clicker is not magic, it is simply a tool which we use to make communicating with a non-verbal animal easier.  It is not a cue – it doesn’t tell the dog what to do, it simply tells the dog that what it has done, or is in the process of doing, is right, and that that particular behaviour or part of a behaviour, has earned a reward.  And, because your dog likes getting rewards that particular behaviour will very likely be repeated.  

Of course, there are other ways to get dogs to repeat behaviours, but dogs just seem to LOVE doing stuff for the CLICK!  There is just such joy and enthusiasm when the clicker comes out and clicker games are played.  I honestly think that it’s more than just the food handouts.  I’m pretty convinced that it’s the fact that you and your dog are communicating on a level which is so much more than your dog following commands.
You can also use a word as a marker – “Yes” or “Good” would both work as markers so if you like the idea of using a marker as a training tool, it doesn’t have to be a clicker.  I will discuss on this blog why I believe the clicker is more effective as a marker, and also possible pitfalls of using a clicker in your training.        

%d bloggers like this: