One thing this weekend really drove home to me is that I've been incredibly complacent in our training. Oh, sure, I train a lot, but I suck at raising criteria. This is kind of silly, because I understand ways to do this: click only better efforts, ask for more duration before clicking, etc. And yet...
So, I've devised a new training plan for the next two weeks. I chose two weeks as my timeline for two reasons. First, we have a trial in two weeks, so I'd like to see some improvements before then! And second, because it seems like a nice amount of time to reassess our progress.
In addition to devising a training plan, I've also developed a method to track our success, so that at the end of our two weeks, I can see exactly how much progress we've made. (This also appeals to my geeky graph-making side.)
I'm going to work on four main behaviors. I will work on them in two minute increments (I have a timer and everything, and let me tell you, two minutes is a LONG time), once or twice a day. The behaviors are:
1. Sitting in heel position ("get-ins"). The criteria for a click is that she is sitting STRAIGHT. In the future, I'll work on straight AND the correct position, and then add in closeness. For now, though, she just has to be straight. I chose this as the criteria because she often wraps around me or sits crooked. The way I'm going to train this is to do 90 degree pivots (then 180, then 270, then 360) in both directions (working on one direction at a time, and eventually including finishes), clicking only for straight sits.
2. Heeling with eye contact. I'm going to use the 300 peck heeling method. I do not require strict eye contact for heeling in general, however, I am going to require eye contact for the first 5-10 feet because I often lose her at the first step.
3. Straight fronts. The criteria is that she must be sitting straight. Okay, there's a second one: she must be sitting straight, and somewhere between my two feet. Closeness and centeredness doesn't matter; as long as she's sitting straight, she'll get the click. I'll reset the exercise by pivoting 90 degrees (and then 180, etc.) to the right (and later the left).
4. Duration heeling. My house is laid out such that I can make a giant circle through five different rooms, for a total of about 40-50 feet. The criteria here is that she stays roughly in heel position (position and closeness don't matter much as long as she's roughly at my left side) and paying attention. I won't require eye contact, but I will require signs of attention- ear set, head position, etc. It's hard to describe, but I know it when I see it.
I'd also really like to work on fast pace heeling as well, but I have no idea how to teach it. Ideas? I've tried throwing toys and treats ahead of us, and she's terrified of a target stick... I'm not sure what else to do, other than just click and treat when she changes speed with me. And... maybe that's enough? Thoughts?
After we've done our training session(s) for the day, I'll wait at least 30 minutes, and then do a test so that I get some data. For the fronts and get-ins, I'll do ten trials and count the number of clicks she gets. I'll increase the difficulty when she gets a 90% success rate.
For the two heeling exercises, I'll do one trial, and measure the distance she covers before failing (and then, of course, do an easier version so she ends on a good note). I'm a bit uncertain about doing the test this way- I hate to set her up to fail- but it seems like the best way to see what her maximum duration is each day. Any thoughts on this?
I won't be updating this every day, but here's today's baseline:
1. Get-ins (left, 90 degrees): 40%
2. Attention heeling: 2 steps
3. Straight fronts (re-set with a 90 degree pivot to the right): 60%
4. Duration heeling: 1.25 laps
I'm really pleased with how well she picked up the attention heeling for even one step. I've never required that before, and she's never really offered it either, so it's cool that I got two steps! I'm also surprised by the number of straight fronts she got. I did feel that she was frustrated with the duration heeling, which means it's good I'm working on this, although I'll need to be careful not to raise that criteria too fast.
Anyway, I'd love some feedback! Does this sound reasonable? Is there anything you'd do differently, or not at all? Let me know!