Dog Swordfighting - Lesson 2: Holding the sword correctly

So in the previous blog I talked about how I was getting my dog to pick up a sword. I also pointed out how I wasn't focused on getting him to hold it right as long as he held it.

So that worked! And it worked exactly as I said. He grabbed it without any real care as to how because he has no concept as to how. He's a dog. And despite what videogames and anime may tell you, dogs don't know how to use swords.

Various dogs that actually know how to use swords.

Missile, demonstrating the ancient canine technique of having no clue what he's doing.

This means he will make up any method to hold the sword as long as it gets the reward, because of course. Reinforcement!

This puts me in a position where I need to make a choice. I can either:

1. Keep moving forward, and refine the position he's grabbing the sword at later. This would have us practice up to the point of lifting and holding the sword, and then afterwards adjusting the position.

2. Adjust the position now and don't think about it later.

This really depends on what you would want, but personally just knowing Missile, I'm going to pick option #2. If nothing else because I really don't want him to tear up the sword and force me to buy another one. He's really, really good at tearing things up.

"To shreds, you say?"

I would have an easier time replacing the super high tech pool noodle(™) , but I would have a much harder time replacing the actual sword. There's some other good reasons though:

• The weight of the sword is different grabbing it at the handle versus at the blade. I don't want to confuse him.

• He's going to be constantly rehearsing the grip throughout the whole series. It would be more of a pain to fix it after than to just fix it now.

So cool! Let's do it.

How...?

Yep! You guessed it! More shaping!

Shaping it up

A cool thing about shaping is that by doing it, you're constantly building the behaviors you do want, and extinguishing the behaviors you don't want. The dog is going to do whatever does not lead to reinforcement less and less. (In theory, anyways)

So this means that all I have to do is adjust my criteria so that where he grabs is something I care about. He's already built a habit of grabbing it in the first place, and he has enough motivation and enthusiasm to just keep trying, so my job is easy. I just pick out the behaviors I want and ignore the ones I don't! Sweet!

Here's what it looks like in action:

https://video.wixstatic.com/video/31b2fc_776dd9e7e2dc4cfa85a3fcd79e4f3cd2/1080p/mp4/file.mp4

Some takeaways here:

• It can be seen at the end, but it appears Missile is targeting the closest thing indiscriminately. In some cases that's just the blade.

• He's losing his mind

• Despite him losing it, he's actually starting to develop some hold skills for free! Saves us a little effort down the line even though it isn't necessarily part of our criteria

So hey! This is progress, but I think we can do a bit better. While it will eventually work itself out, I would prefer to figure out his grabbing the handle correctly before the sword gets so many bite marks it's unrecognizable.

This was after two minutes of practice.

What this means is we want to increase his success rate. Doing so should in theory help him learn faster, simultaneously sparing our poor sword from needing a replacement and my poor wallet from needing to buy another one.

So how do we do that?

We cheat!

Okay kinda, but not really. The proper term is "setting them up for success". Basically make it as easy as possible to do the right thing. This allows us to reinforce that behavior more, and in theory**** the dog will do the thing I want more often.

So again, how do we do that?

There's a small hint in the video. Missile is targeting the closest thing to him. It makes sense. That's a thing dogs do. He does try to think it through sometimes but he often falls back on his habits.

But that's okay! We can manufacture it! Here's the plan:

1. Create as many scenarios when shaping this where the piece he needs to grab is the closest piece to him.

2. As the successes mount, adjust the sword so that more of the blade faces him.

3. Keep repeating this process until he can nail it regardless of orientation.

I'll be honest, this part is pretty boring to watch so I didn't record anything until we nailed it down.

So here's a video showing it after all of the tedious work was done:

https://video.wixstatic.com/video/31b2fc_c8f5e7bd211e4b499f741f8ce4110c0d/1080p/mp4/file.mp4

That looks WAYYYYY better! Takeaways:

• He is intentionally grabbing it. In the video I reverse the sword to show that he's targeting the handle, not just the closest or farthest piece.

• At one point he accidentally grabbed at the blade and then immediately corrected himself

• He's absolutely not losing it anymore. Still fast though.

But how long did it take?

In terms of overall time elapsed? Between the first video and this video it just took a few days.

More importantly, it was a few days of intermittent 1-2 minute sessions. If I had to give you an actual estimate the total amount of time spent was between 20-40 minutes of active training. It really does take less than you'd think to make solid progress!

Wrapping it up

That's all! This time around we focused exclusively on getting Missile to focus on how to grab the sword at the right spot. That took some effort! Next time we're going to keep building this same trick. Now that we've nailed down grabbing, we can focus on holding!

If you're interested in more of this type of thing, I'd love to invite you over to DogKatas, where we're building a community of fellow Dog Training nerds practicing all kinds of things! I even go out of my way to break the steps down for this trick into individual exercises. Check it out! Look forward to seeing you in the next, really grabby blog post. Happy Training!