About our 1st routine: Shrek
Dog dancing, or freestyle, is a sport so packed that it is claimed by many (myself included) to be the hardest dog endeavor there is. The goal is to have the tricks cued by verbals so that you can choreograph and do alternate (and sometimes counter) movements yourself. I’ve videoed this routine a number of times, and each one has some point(s) where something goes other than planned. It’s a fascinating study of behavior and learning.
One of the first little chains I wanted to use was backing away from me and then sitting up at a distance. I chose it because it was solid, and a hard skill (that I wanted to showcase). It was so solid, in fact, that I can back away from her the opposite direction, for a long distance. And sitting up is a trick she’s done since puppyhood. We also practiced the chain regularly and it was pretty fluent. It’s inherently difficult because of the distance, the counter motion, and then the anticipation of the sit up cue, followed by the anticipation of running forward once released from the sit up (which is really a “fancy stay”).
Throughout the evolution of the routine, though, this little sequence of behaviors went through a number of stages. It got so bad it had to be pulled out of the routine and trained on its own, slicing the chain down to its component parts, totally away from where we practice our routine. After trying unsuccessfully to “repair it”, my strategy was to just give it a break. Let latent learning do its thing and remove the pressure. In the routine, we skipped it entirely for a while.
What I learned is: this is common. Now, there are plenty of people pursuing the sport with untrained behaviors. Those will, of course, fall apart. But, at the risk of sounding too full of myself, I don’t think that was our problem. I know how to train, I know when things are incomplete/untrained, I have a good perception of how my position and movement effect a dog’s understanding/perspective (agility remains a huge training focus, and those qualities are inherent to the sport), and how those things can contribute to failures and successes. But she knew these behaviors. Could perform them under pressure, in multiple environments, in different positions relative to me, on verbal cues alone…although anticipating the sit up might waver. Was that enough of a crack to cause it all to topple?
When trying to repair a situation like this, you’ve got to be careful. Push too hard and your dog starts hating it. Some dogs come to hate the song you’re using, might develop an unwanted emotional/biological reaction to hearing the cue, or shut down in some other way. If need be, you might be better off to just permanently remove something and alter the routine. You’ve got to be willing to let things go. That is not a bad idea. But, for me, this was a fascinating degradation of behavior, and experimenting with it was worth more than the routine. I care more about understanding how animals work and learn, and this was ripe for a study.
I had a hard time giving up on the behavior and putting it “on a shelf”, even for a little while. But she was so mixed up, and I simultaneously had a competition coming up where I needed her to perform a single paw lift that I wanted to tidy up to be purely verbal, so I finally shelved it. However, I’d kept pulling it out every now and then to see where it was at. We focused on different things for a while, and then started to ease back into just the backing up part.
Backing up morphed into a behavior that had a limit of about 4 steps. Incidentally, this is generally how an average person trains the skill: dog backs 4 steps, stops, persons marks and rewards. That dog believes the skill is back up AND STOP after __ number of steps, not back up until you hear a mark, when you throw the reward to the dog who’s ideally still in motion. We didn’t begin training the routine with this limit, but it was sure cropping up during our “repair training”. A huge challenge with the predictable number of steps is that if you wait for more steps, and your dog doesn’t ever offer them, they get frustrated and usually stop (maybe run forward/offer other behaviors you’ve been working on), and you have only so much frustration to play with before you have to fabricate or settle on a success. In this situation, some would say ask for an alternate behavior that the dog knows well and reward that. That’s something to try, and probably the only option when things are this far gone. I believe that the problem with that strategy is that you don’t get the understanding about the failing behavior to develop, so it’s not my first response, I employ it rarely. But if your dog is dying for some sort of success, by all means.
Eventually we built her duration of backing back up. But it wasn’t easy. Timing is really critical and can screw things up if you don’t get close enough, often enough. And, ideally, the verbal cue should initiate a non-stop reversing, on a single command. But I leaned on repeating the cue to help get over the hump; uttering it right before she’d usually stop, and then looking for times where she continued on her own and making a huge deal about it (jack-potting). Also, I believe praising in motion helps. It might initially confuse your dog, thinking she got marked and can stop, but hopefully delivers the information about what the right choices are—to keep backing up.
Separately, we needed to address sitting up. Keep it sharp and keep her listening for the cue, not just defaulting and offering it because we were practicing it recently and she’s been rewarded for it a lot. And dare I mention that the sitting up behavior I taught involved the criteria of sticking her front paws in the air? I totally forgave that criteria for a while. That was one battle I didn’t pick, though when you sit down and analyze, you see that there are a lot of options about which battles to pick, and who knows what every best choice is.
Beyond that, sitting up needed to be snapped back into the chain of behaviors. And I’m a stickler for not letting a dog guess what the next piece of a chain is. It wouldn’t be useful for her to insert the sitting up whenever she felt like it. You can let a dog get patterned, but with my agility background, I rather have responses to cues than a patterned dog. So, drills included backing up and asking for alternate behaviors occasionally, not just sitting up. Whew, that’s a lot of work already…and that’s just a few tricks being trained outside of a whole routine! Can you see and appreciate how hard freestyle can be? Even more so if little to none of the routine is patterned, and each behavior is a response to your unpredictable cues!
When the problematic chain started to look strong enough, tested away from the routine, it was time to start working it back in. By the way, in case you don’t know, trainers are rarely supposed to practice a complete routine. The protocol is to train segments thoughtfully. My plan was to reward a really nice backing away, placing the reward BEHIND her. Because it’s already difficult to back up while I move the opposite direction, the last thing she needed was the idea that next she runs forward for reinforcement.
I’d occasionally practice the routine with the reverse into sit up with duration (there’s an expected “Stay”). Then, I’d release her but instead of continuing with my routine, I’d release and send her to a toy behind her hanging on a fence or toss the toy (or hunk(s) of meatball) to her right in position.
In the finished product, the routine involves her being released from that “fancy stay” to come forward, circle around me, then comes a send forward to circle around a prop. Dogs naturally find it much more rewarding to run forward and race around something (especially this dog), and being in closer proximity to the handler makes a dog feel more secure. That was why I picked this super difficult maneuver—I wanted to show off. But, that anticipation of running forward to be close to me, and race around me and then an obstacle was too much to not tear apart that backing away to nowhere skill. It would have been easier if she backed up to a target/mark. But I never ended up experimenting with that for the repair. I have this thing I probably picked up from agility training where I want the dog to be able to follow commands midst changing contexts. I hope I didn’t make it harder than it needed to be. Sorry puppy! I guess I can always go back and add some sort of prop that will be a “landing pad”…it’s not like we’re going to forget our routine. I think that practicing it less will probably keep it from getting sour and might end up in us capturing our best execution yet.
We recently participated in a demo where we practiced the routine with breaks for reinforcement. Demos are great because you can break things up and reward, so take advantage! Anyway, the first time I asked for the backing up, it wasn’t great. After a little “ice breaker” where we played for a sec, I held the toy and went back to that segment, and she shot across the floor backwards with isolated footsteps like a pro. It’s funny how bad it was before and how phenomenal it was next try.
I’ve decided to give it my best but not be too attached. That’s why I’m finally uploading an imperfect video of my routine. You can see her struggle with the sequence we’re talking about. Sometimes that part goes great but we flub something else. In the ending of this shoot, where the arm weaves happen, there are debarked dogs off camera that are clamoring at the front of their x-pen and worrying her. I was upset to lose a few otherwise good video attempts to those annoyances. But, that’s the kind of thing that can happen.
Moving forward, I want to attempt a different style of routine and focus more on dancing. In the very last bit of this routine, I’m so busy cuing circle around, leg weaves, circle around the opposite direction, that my upper body is just dead. Still, not bad for a beginning routine.
I’m really excited about my next routine, but you’ll have to be patient. A trained routine takes a lot of time…especially when things fall apart along the way ;)!