Behaviour Chains


Aka reducing the amount of treats you use!


Let’s get started and learn how to use behaviour chains to turn your pup into a gambler! (not knowing how many behaviours he might have to complete before earning a reward).

 

A behaviour chain is simply a series of behaviours performed one after the other for a reward.  A simple example would be a Sit, Wait (while you walk away) and a Recall (‘Come’) performed in a row for just one reward at the end.

 

One of the first things many people train their dog is ‘Come’, then as soon as the dog comes, they ask him to Sit, and as soon as he sits, they ask him to ‘Give a paw’ and then finally, once the dog gives his paw, he gets his treat reward.

 

Without realising it, what they have done here is a behaviour chain. They have merged three different and separate behaviours into one action:

  • Come, Sit, Give a paw

 

Although this simple approach may work in the beginning, soon things may change.  

 

In fact, if you were to shape any other type of behaviour chain based on this approach, you would have limited success and it would probably take you a very long time to do so. 

 

The reason is simple - what are you actually rewarding?

 

Focused heelwork is a good example for this. It requires three different elements in 

order to be fluent:

  • Training the dog’s speed to match the handler 
  • Training a certain position in regard to the handler’s body
  • Eye contact (or looking at a given focus point on the handler’s body)

 

If you were to try and train this all at once, how would your dog know what you are actually rewarding for? 

 

In order for your dog to properly understand the sequence of actions (behaviour chain) the best option is to train each behaviour separately and then merge them together.  

 

It's very important to wait until you have a variety of behaviours trained under distraction before you attempt to start sequencing behaviours together.  

 

Steps:

  1. Go to your familiar training area with no distractions (or food rewards) visible.
  2. Ask your dog for a single behaviour
  3. When your dog responds with the correct behaviour, tell him he’s great and then instead of giving him a treat as you usually would quickly ask for another behaviour.
  4. Once that second behaviour is completed then reward your dog.
  5. Now in your training sessions you can gradually work up to asking for anywhere between one and ten behaviours before offering a reward.
  6. As always, increase the challenge slowly and systematically. 
  7. If your dog fails just start over but don’t always reset to an easier level – repeat the same challenge .  This is important.  If you were asking for five behaviours and failure took place after three, you still want to ask for five on the next go otherwise your dog won’t learn to keep trying.  
  8. Another tip that works really well when asking for lots of behaviours before a reward is to sometimes make the reward at the end much bigger than normal.  This gets your dog thinking, still no reward…still no reward…wow I bet something really great is coming soon because I’ve been working for ages!
  9. Also keep in mind that if you’re performing the same sequence of behaviours, remember to reward the middle steps every now and then, as well. If you consistently reward just the last behaviour you may get a lack lustre performance of the first behaviours as these ones never get reinforced.  By rewarding the middle steps just occasionally, you get a correct and reliable behaviour chain, every time.

 

So, does all this really work in real life?  

 

You bet it does.  Have you ever watched the dogs at Crufts performing heelwork routines and running agility courses in a huge crowded public arena, under pressure and with no treats in sight?  

 

The truth is these dogs always get treats, forever.  No effort is ever made to wean them off treats because their trainers place tremendous value on having a dog who will perform with great enthusiasm and speed no matter what.  It’s just that the dogs will never earn treats in the competition ring.

 

These trainers turn their dogs into extremely optimistic gamblers.  In training, the dog might get rewarded for performing one behaviour or he might have to work for several minutes before being rewarded.  The reward could be anything the dog values, treats, toys, praise.

 

This is one of the big keys to getting a dog that will work for you under distraction.  Continue to reward what you like in a variable way so that your dog doesn’t know how long he’ll have to work to get his motivator.

 

Here’s an exercise to practice at home: 

  1. Call your dog to come in from the garden every day and you pick which of those times you’re going to reward him/her.  Maybe one of them, maybe three of them, maybe all of them.  
  2. Vary the reward you offer for each, you might reward with a dry biscuit or maybe you have a big piece of leftover chicken breast – jackpot!
  3. You dog has no way of knowing what you are offering for any given request and he will keep co-operating and trying because this time he/she might win big! 

 

Here is a link to a video showing some simple sequences of behaviours you can practice:

https://www.youtube.com/watch?v=Y8Ur5rD7V-8

 

Finally on this subject, behaviour chains can also be used as a means of discouraging undesirable behaviour.  Bad behaviour doesn’t happen in a vacuum – there’s always something that leads up to it. You could call it a cue or a trigger, or in keeping with the metaphor of a chain, a link. What you need to do is make every link in the behaviour chain dependent on the link that preceded it.

 

Some dogs, for instance, can display some very annoying behaviour when it’s time to go for a walk – whining or barking when you get your jacket, spinning when you try to attach the lead, and rushing ahead of you to get out the door.  So…, don’t put your jacket on until the dog stops vocalising, don’t attach the lead until he stops spinning, and once the lead is attached, don’t move toward the door until he’s sitting calmly beside you.  In other words, wait for the behaviour you want before moving on to the next link in the chain.