Wednesday, January 27, 2010

Revised Training Plan

In my last post, I set up a training plan for the next fortnight. Laura and I had a nice conversation in the comments about my criteria on heeling, and as I've both worked with Maisy and thought about what she said, I've revised my plan slightly. I'm still training each exercise for 2 minutes, but I've changed some of the criteria and testing slightly.

Exercise: "Get-ins" (sitting in heel position)
Criteria: Sit straight. Position and closeness do not count.
Testing Method: Put Maisy in heel position. Pivot 90 degrees to the left. If Maisy sits straight, click and treat. If not, try again. See how many trials out of ten she does correctly.
Notes: Unchanged from original (but I've seen a 20% increase!).

Exercise: Eye contact during first step of heeling.
Criteria: Maisy must maintain eye contact when I take the first step of heeling.
Testing Method: Put Maisy in heel position and get eye contact (she offers it the vast majority of the time). Step off. If Maisy maintains eye contact for the first step, click and treat. If not, try again. See how many trials out of ten she does correctly.
Notes: This is a pretty big change. Originally, I said I wanted to do 300 peck heeling with eye contact, but I've changed my mind. I don't want eye contact the entire time, just during the first step, and then intermittently afterwards.*

Exercise: Fronts
Criteria: Maisy must sit straight in front of me. She doesn't have to be centered or close, but she does need to be between my feet (so, generally centered).
Testing Method: I call Maisy front, and then pivot 90 degrees to the right. If she moves and sits straight in front of me, I click and treat. If not, I try again. See how many trials out of ten she does correctly.
Notes: Same as originally set up.

Exercise: Heeling with duration.
Criteria: Maisy will maintain attention* while heeling.
Testing Method: I am teaching this one with the 300 peck heeling, except the pecks don't correlate with a single step but rather in roughly 8-10 foot increments. (The reason these increments vary is because I have a large circular route I can follow through my house. Well, it's more of a square, really, and each side of the square varies slightly.) So, I set off in heel, and if Maisy is attentive for one side of the square, I click and treat. Then we do 2 sides, click and treat, then 3, and so on. If she loses her attention at any point, I start over at one side. We do this for two minutes during the training phase. During the test, we work backwards. We try for 3 sides. If she makes it, the test is complete. If not, we try for 2, and so. I record how many sides she completes during the test.
Notes: I know this sounds complicated, but it's not, really. The reason I chose to do it this way is because I wanted to set her up for success. Previously I was going to see how far she could go until she lost attention, but that meant that in order to measure her progress, I had to wait for her to fail. This way, I can build on success, and then test to see if she maintains what she achieved during the training phase. If she doesn't, I lower the criteria and try again. This way, she is always rewarded for doing well, and thus is set up for success.

*Laura rightly pointed out that I had two different criteria for heeling: eye contact and duration, and you're never supposed to work on two criteria at once. Since I don't want eye contact for the entire time we're heeling, it was kind of pointless to do 300 peck heeling with eye contact. Still, I need Maisy to pay attention during heeling, so I've spent quite a bit of time thinking about what attention looks like if she's not making eye contact. I'll post later this week about why I don't want constant eye contact, when I do expect eye contact, and what "attention" looks like. (Sorry to be such a tease, but it really deserves its own post!)

No comments: