Constant reinforcement is only necessary during the learning phase of a behaviour. Constant rewards are not necessary once the behaviour is learned. Constant reinforcement will result in an animal that puts in the minimum amount of effort to get his piece of kibble. For example, think of an employee that regularly gets his pay check and only his pay check. He gets the same money week after week. There are no bonus incentives, no words of encouragement from management, no donuts at coffee time, no lunches with friends, no parties, no business trips or raises.
What would happen to this employee’s performance? He would do the minimum amount of work required and would likely start looking for another job!!
Variable reinforcement is imperative to have a dog that wants to work with you. This is the principle behind gambling. It is the possibility of the big score that keeps the gambler coming back for more.
Once a behaviour is learned it is crucial to vary the rewards the dog gets for complying with a request. We not only vary the type of reward but we can vary the amount, timing and delivery of the reward.
|Type of reward||Amount of reward||Timing/delivery of reward (examples)|
|Food (liver, cheese, kibble, carrot pieces, hot dogs||1 piece, 0 pieces, 3 pieces
mix them up
|Every time dog does behaviour (only in learning phase)||Every 3rdor 5th or 2nd time he does behaviour he gets a food treat|
|Play with toy||Length of time short or long||Dog gives a quick response to command
|Dog does 5 behaviours in a row > throw frisbee|
|Chase owner||Time, intensity, hide and seek||Dog coming on recall, owner turns and run away||Dog holds a wait for 10 seconds=finds owner|
|Valuable commodity. Do not give away for free||Dog looks at you instead of squirrel>reward||Walking on a loose leash > reward|
|Couch time||Occasional, only specific furniture||Dog given couch time after complying to request.||Dog gets off couch, given command then allowed back on couch|
The training process works like this:
For beginners (using sit as example)
- Lure dog into position> as soon as dog’s rump hits ground>bridge with “yes”> reward with food.
Once dog responds reliably to lure: (dog responds 9 out of 10 times to lure)
- Verbal command “sit” >lure into place>bridge when rump hits ground> reward
Once dog responds reliably to verbal cue and lure
- Get rid of lure. Hand movement becomes a signal. We need to get rid of the lure as soon as possible because luring creates an expectation of reward which decreases the dog’s motivation. Remember the gambler. We want the dog wondering what the reward will be so he is tuned in to his handler.
Verbal or hand cue > dog sits>bridge>REWARD
It is now crucial to look at the chart above and vary what comes after the bridge.
How to know when to progress in training? Outlined above is our training method at Dealing with Dogs.
Here are some universal principles to training:
- Go forward slowly in training so your dog gets reinforcement opportunities
- Train one thing at a time
- Make sure the dog is responding reliably before making things more difficult ie by adding greater distractions
- Have a clear picture of what you are actually teaching the dog.
- Be creative when trying to get the behavior you want
- Interrupting a training session in progress is confusing for the dog.
- If behaviour is getting worse, go back a couple of steps so the dog has an opportunity to be rewarded.
- End each session with something the dog is able to do well.