Tuesday, September 14, 2010

Tiny little box that makes a happy sound = $2.50

$2.50 9yrs ago...This past weekend was really interesting and the theme each day came back to the fine balance of raising criteria and fading clicker and use of food once a behaviour is established.  It's a process that needs to be repeated in each new environment until both you and your furry pal, know the drill with your eyes closed.  Even D & Cc issues or BAT you've got to be selective and increase criteria sub-threshold otherwise you'll feel like a pez dispenser and their behaviour either plateau's or gets worst.

I LOVE this blog post written by Dr. Ian Dunbar. reinforcement-schedules
Please read Dr. Ian Dunbar's reply to Lee Charles.. I think it is gold in summing up how to streamline the process.  "Hi there Charles. Ask Bob. However, I would imagine he means that "every time a dog meets criterion you would reward him." However, the criteria are progressively refined (tightened) as the dog's behavior improves. For example, if training a dog to pay attention,  first you would reward if the dog glances at you, then a look for one second, then two seconds, then three, five, eight, ten, fifteen, twenty and so on, until when the dog is cued he is expected to look at you for 30 seconds. So, when the dog looks at you, he is NOT rewarded every time. However, if he meets criteria, he IS rewarded every time. If the dog looks at you for less than 30 seconds he is not rewarded but if he looks at you for more than 30 seconds he is rewarded, i.e., we only reward above-average responses. Carefully and progressively refining criteria in clicker training is exactly the same principle as test-train-test and rewarding the dog on a differential reinforcement."

I learned with Daisy my brainiac 9yrs ago that I needed to make a conscious effort to increase criteria each interaction otherwise she gets frantic and starts offering all sorts of stuff up that has been rewarded in the past while working her self into a frenzy.  The other side of the coin is a dog (George...) will just shut down and offer nothing.

Being systematic and deliberate pays off. It's essential to ensure I don't create a routine that the dog becomes conditioned too. I find that once you do...if something "out of the routine" happens...then it causes an over reaction (or shut down) which is counter productive to what you are trying to achieve in the first place.

Back to the reason I started this blog post...clicker stuff!  Sunday I went to our Squamish Pause for Cause w/ Keegan and Daisy to support our local SPCA.

Here's a highlight from our soggy day...a trick competition. We were the last to show a couple tricks.

I did a quick talk on clicker training,  talk about how I use a clicker...to ensure people hear me say...it's used as a tool to capture a behaviour you'd like to see more often and put it as quickly as possible on handsignal or verbal cue (within 3 reps with Daisy), vary the rewards asap (play, environment, toys, sweet talk, cuddles), then fade the clicker and transfer the value for the behaviour to that cue word, add distractions and take the show on the road.

In my experience if you do not raise criteria asap and breakdown the behaviour so that it's a piece of cake for the dog to quickly grasp what the end behaviour is, you end up using a clicker as a crutch.

Just like the bad rap of using food has with some people... a clicker can become so exciting that it becomes a distraction and arouses the dog up to the point that they're no longer really concentrating on achieving the next step in the sequence of event you are trying to teach.

No comments: