Tag Archives: ian dunbar

05/17/12

Seven Steps to Off Leash Reliability

This post is part of the series in response to Dunbar’s 2012 Australian seminars. See index.

A reliable recall is often difficult to train.  Dunbar has many suggestions on training a dog to be reliable off-leash – however, he doesn’t teach a recall as such, more cues at a distance.  The logic is that it is just as useful for your dog to sit immediately, on cue, in any context, as it is for you to train a recall – and perhaps even better if you do not want your dog to move towards you (for example, if there was ongoing traffic or other hazards associated with approaching you).  Dunbar’s logic is partly based on not removing a dog from their rewards, as well.  For most dogs, being off leash is associated with a lot of fun, and calling a dog away from that fun is inherently punishing, despite any rewards you think you may offer.

Without further ado, here is Dunbar’s seven steps to off leash reliability.

 

Small back and tan crossbred dog running towards camera.

Photo © Ruthless Photos.

 

Continue reading →

05/10/12

Repetitive Reinstruction (Dunbar)

This post is part of the series in response to Dunbar’s 2012 Australian seminars. See index.

Dunbar advocated what he called ‘repetitive reinstruction’ as a form of verbal punishment for dogs that do not obey a trained cue.

He called this technique negative punishment (taking away something good), but this only happens if you are always rewarding the behaviour in the first place. Repetitive reinstruction is taking away the ‘nice sweet speaking person’ and you get ‘annoying and over the top’ person. It works quickly, and response reliability increases after successive trials.

Staffordshire bull terrier crossbreed smiling at the camera.

This is Dexter, a young Staffy X available for adoption through Adelaide All Breed Dog Rescue. Click his picture to visit their Facebook page.

 

The best way to illustrate this technique is by describing training a distance sit.  Here is the process:

1) While the dog is playing, grab the dog by the collar, give them a treat, then release them to continue play.

2) While the dog is playing, grab the dog by the collar, ask for a sit, give treat, and then release them to continue play.

3) Without touching the dog, cue sit, then grab the collar, give a treat, and send them to play.

4) From a casual position, cue “sit” as many times as necessary (perhaps decreasing distance and increasing urgency in voice) until the dog sits.  Once the dog sits, release the dog, ask for a second sit, reward when the dog sits on the first cue, treat the dog, and send it to go play.

 

In this way, the dog is repetitively reinstructed to perform the behaviour until it does so.  Though this may be a number of cues in the first instance, over time, the number of cues will reduce to 1-2 cues.  He assures us that this does work – the number of responses increases per the number of cues, over time.

Dunbar acknowledges that collar grabs should always be associated with good things, in order to reduce the likelihood of dog bites from this interaction (touching a dog collar is often a precursor to a dog biting).

Obedience competition has made a second cue a ‘crime’.  In reality, we are aiming for one cue eventually, but it’s okay to use multiple cues in training. Dunbar has collected figures to graph the improvement and there is an improvement over time (i.e. the number of cues diminishes).

Not only is this repetitive reinstruction, it is also specific redirection.  It is instructing the dog what to do in order to stop the negative punishment.  This can be a useful way to direct the dog during ‘crazy times’ (when the dog is over aroused, like when the doorbell goes).

Dunbar argues that the cue doesn’t become ‘irrelevent’, because if the dog doesn’t perform the first time, the cue was irrelevant anyway.

This is one of the more controversial suggestions from Dunbar, with many dog trainers advocating for one cue only.

 

Suggested further reading:

Patricia McConnell on repeated cues.

“Rover, sit. Sit. SIT. SitSitSit!!!” – a review of Dunbar’s suggestions by Boulder Dog.

05/7/12

Dunbar on Classical Conditioning

This post is part of the series in response to Dunbar’s 2012 Australian seminars. See index.

Dunbar believes Classical Conditioning is a big deal in dog training.  He doesn’t believe in separating Operant Conditioning from Classical Conditioning. To Dunbar, the dog learns the same thing: In case of x, good things happen, and in case of y, bad things happen.

White poodle.

For a basic overview of classical conditioning:  Classical conditioning was named by Pavlov, who learnt that if he rung a bell before feeding dogs, overtime, the dogs would start to saliva

te at the sound of the bell alone.  Basically, classical conditioning is associative learning.  Dogs will associate things with certain stimuli.  In Dunbar’s view, however, he thinks that the brilliance of classical conditioning has been lost over time.  In his opinion, just because Pavlov worked with a reflex (i.e. the dog couldn’t choose to salivate, it happened itself), doesn’t mean that classical conditioning is only used for reflexes.  (This is normally the distinctive difference between classical conditioning and operant conditioning, but Dunbar ignores it and finds it irrelevant. Confusing!)

 

Classical conditioning should happen all through a dog’s life.  Treats should never be phased out for classical conditioning.  Dunbar said, “Don’t take you dog’s temperament for granted” and “If your dog is friendly, it can be friendlier”.

Dunbar particularly advocated Classical Conditioning for improving handling of dogs.  Particularly, classically conditioning collar grabs and other contact the dog may find unpleasant.

He talked about dog trainer Bill Campbell’s ‘jolly routine’, which is an over-the-top play response an owner ‘performs’ when a stimulus is seen.  The idea is the dog things, “Whenever my owner sees x, they get so happy!” – which in terms classically conditions the dog to like x.

While he believes there is better ways of getting rid of problem behaviour, he did talk about using ‘lightning strike’ verbal feedback for poor behaviour.  He described this as punishment in a praise sandwich.  For example, the dialogue would go: “good dog, very well done, lovely dog, excellent, good, good, yes, ASSHOIFJIDSHKLJ!!!, yes, good dog, good, very good”.

He also liked the use of classical conditioning in shelters, particularly in the Open Paw program, where dogs are rewarded for just being in the shelter and seeing people.

Dunbar describes classical conditioning as a ‘winning strategy’.

Further reading: my lecture notes from Paul McGreevy on Classical Conditioning.

External link:  Roger Abrantes’ post on “Unveiling the myth of reinforcers and punishers” 

Updated with additional notes 13/10/12.

05/2/12

Lure Reward Training

This post is part of the series in response to Dunbar’s 2012 Australian seminars. See index.

Dunbar promotes lure-reward training.  This training uses a lure in order to accomplish a behaviour, and a reward to communicate the desirability of the behaviour lured.

 

 

Young red dobermann puppy sucking on a finger.

Even young puppies can be lured by food on fingers, or even the sucking reflex itself. Photo © Ruthless Photos.

The Steps of Lure Reward Training

Any behaviour we can teach on cue by a 1234 process in lure/reward training. The 1234 steps are:

1: Request (i.e. use the cue – e.g. say “sit”)

2: Lure (i.e. get the behaviour – e.g. use food to get the dog’s nose up and bum down)

3: Response (the dog performs the behaviour)

4: Reward the dog

 

1,3,4 is always clear, but the lure to get a behaviour (2) may need some creativity.

 

There may also be several lures for one behaviour.  If this is the case, they should be ranked from most desirable to least.  For example, if you were training a dog to eliminate on cue, you could ‘lure’ the behaviour by confinement (i.e. putting the dog in a crate for a period of time) or by an olfactory lure (pee smell).  It’s probably easier to crate your dog than to carry pee in a bottle, but it’s up to you to determine the best lure for you and your dog.

 

Once you’ve got the behaviour, you can start to progress through the stages of luring.  Lures should be phased within a few repetitions, and rewards should become ‘life rewards’ instead of extrinsic rewards in a brief period of time.

 

The stages of reward training is phasing out, first, the lure and, second, the reward, and then enforcing compliance (without pain).

 

 

Stage One – Phase Out the Lure Continue reading →