Failure

I was really excited to go to a Daisy Peel seminar on running contacts the other day.  I got a lot out of it, from uses of the Manners Minder (I ordered one that night) to how to see the proper gait of a dog running across the dog walk (still need some practice there).  One of my biggest take aways though was the importance of pushing your dog to fail.  Daisy (formerly a high school chemistry teacher) compared it to a class of students taking a test.  If every student gets 100% then the test was too easy, but if every student fails then it was too hard.  A teacher should want most (maybe all) of their students to pass, though not all with A's.  If all of the students got 100% then none of them had to work all that hard, and certainly didn't learn anything from taking the test.  The teacher doesn't learn where their students are lacking either which would allow them to improve and progress their teaching.

It's the same with dog training.  If your dog gets a skill right every time then they're not learning anything new, and you don't know what gaps there are in your dog's training.  For example, you practice weaves with your dog at home and they get it right every time.  Then you go to class to show off your dog's skill, but suddenly they're getting it wrong every time.  There's a gap in your training, in this case they don't know how to weave in a new location, or with different poles, or on mats instead of on grass or carpet.  Had the dog been pushed at home at least one of these differences could have been addressed and a better training session could be had at class.

Daisy suggested the 80/20 rule.  Your dog should succeed 80% of the time and fail the other 20%.  It's as easy as getting 4 successes in a row (4 out of 5 is 80%) and then changing something that might make the dog fail.  Get 4 nice weaves in a row then add a front cross, or a little distance, or move the poles to a different part of the yard.  You'll eventually have to get really creative to get your dog to fail!  Then you'll know that your dog really knows their job.

When a dog is just learning something, 80% success may even be too much.  If they're just starting to learn weaves, say by the 2x2 method, it's likely much more helpful to move from 1 set to 2 to 3 when the dog is only 60% successful than to dwell too long on one part of the process.

I am definitely going to keep this in mind when training with my dogs.  I already tried it at open weaves with the girls last night.  With Gidget, I decided I was only going to reward "fast" weaves.  We started with just the weaves by themselves both directions, she was nice and speedy so we added a jump at one end and a tunnel at the other, she was still successful and speedy.  On the next turn we added a tunnel between the jump and the weaves with either a post turn or front cross after the tunnel, still speedy and successful.  Time for a front cross after the weaves, first time successful, second time she popped out at the beginning.  Possibly because of the tunnel beforehand, or I was speeding up early to get ready for the cross, both things I know we need to work on.  Now if I keep track of what makes her fail, I'll know what I need to work on next time and will be better able to come up with ways to make her fail so we can learn even more!  This sounds like a good tactic to me!  Happy agility!