Sunday, January 29

Value Transfer Concepts

One of the things we discuss in my Basic Manners class is the idea of "value transfer." Forrest Micke, an amazing trainer whose online classes first introduced me to this concept, also refers to this as the "economics of dog training." There are LOTS of sources of reinforcement in our dogs' lives, and each has a different value to the dog - various treats, different toys, their food, squirrels, playtime with other dogs, etc., etc. Anything that grabs your dog's attention and is desirable to him is a potential source of reinforcement, and also a potential competing motivator (something that competes with you and your rewards for your dog's attention).

Playing around with Value Transfer concepts is helpful for teaching our dogs clearly when they are allowed to access reinforcement and when they're not. This makes it possible for you to tell the dog "leave it" and have him ignore the squirrel (or any other competing motivator) and accept treats or affection or another form of reinforcement from you, even if he'd rather have the squirrel, because he clearly understands that the squirrel is unavailable to him and that you are his only source of reinforcement at the moment.

In Manners class, we only have time to practice the first basic step: with equal value treats in each hand, we teach the dogs that an open hand is an invitation (reinforcement is available) and a closed hand is an automatic leave it (reinforcement is unavailable). We then switch back and forth between a closed and an open hand, encouraging the dog to ignore the reinforcement that is unavailable (the closed hand) and instead look for and commit to the hand where reinforcement is available (the open hand).

I wanted to include these video clips of Gideon and I playing around with this concept to show not only the first steps to anyone who is unfamilar with this training, but also to show my Manners classes where they can go with the training to increase the difficultly level and help the dog fully understand the concept.

In the first clip you'll see me training with food only. I have four different types of food/treats. Two are low value: one is Gideon's normal food (the rest of his breakfast that he hadn't finished yet), and one is Charlie Bear treats (small round crunchy treats). I also have two treats of higher value: a soft, meaty training treat that Gideon likes (although it's a bit chewy - you'll see him take his time eating it), and also some string cheese (relatively high value and nice and quick to eat).

Here is a description of what we're doing as you watch the video:
     We start out with a basic review of what the open hand and closed hand mean. You'll see Gideon really digging into the open hand (we also use this open-hand concept to develop luring, which we use to shape behaviors later), and you can see that he (somewhat) remembers that the closed hand means the treats are unavailable. Always review the simple stuff and then build from there only when your dog is ready.
     Seeing that Gideon is remembering this game, I then start with an open hand and switch it to closed, or start with a closed hand and switch it to open. This ensures that he fully understands the "rules" regarding the open/closed hand.
     Next I start presenting two hands at a time - one open, one closed. Gideon has to look for the open hand. My next step is then to close the open hand and open the closed hand, so he has to switch - he must recognize that the reinforcement that was previously available to him is now unavailable, and switch over to the other hand where reinforcement is available. He's doing pretty well, so I even start offering the closed hand closer to his nose so that he has to actively ignore it and move past it to reach the open hand.
     Up to this point, I have been using approximately equal value treats in each hand - not asking him to switch back and forth between high and low value. But since he's doing well, I increase the difficulty by putting a high value treat in the closed hand and a low value treat in the open hand. This is the first time that I've asked Gideon to ignore a higher value reinforcement that he'd prefer to have, and instead commit to the lower value reinforcement. You'll see once or twice when I mark "yes" to let him eat the treats from the open hand that he looks to the closed hand (high value) in hopes that I'll reward with those treats. You'll also see him stubbornly poking my closed hand with his nose and even pawing at it once. I wait him out, and when he makes the right decision and commits to the open hand with lower value treats, I reward him by giving him multiple treats and letting him chase them a bit. And at the end, I bring out the last few high value treats so that when we end the session he isn't feeling disappointed or cheated.

In the next video clip, we use very similar techniques but with toys instead of treats. Instead of physical cues (closed hand/open hand) to tell him which reinforcement is available to him, I use a combination of specific verbal cues and obvious body language to tell him which toy he is allowed to access. Just as we did with the treats, we start with two of the exact same toy so the value is equal.

As a side note, before you can do the rest of this type of work, there are two things your dog MUST be fluent in! This training will not work if you don't have these skills in place beforehand.
First, your dog must fully understand two verbal markers:
   "YES" means he can access a reward that I am holding
   "GET IT" means he can access a reward planted somewhere away from me
And second, your dog must fully understand the rules of toy play:
   1. Play with toys enthusiastically (the dog has to enjoy it for it to be reinforcing)
   2. Targets the toy well (bites the toy, not your fingers)
   3. Brings the toy back automatically (rather than keeping it for himself)
   4. Drops the toy on command (either to your hands or on the ground)

Here's what's going on in the video:
     I ask Gideon for a simple behavior, which he performs, and he is rewarded by an obvious presentation of the toy (clear presentations are the first step to good targeting skills!) I then ask him to "out" the toy and access the other one. Each time I cue "out" while playing tug, I make sure to stop moving the toy. A "dead" (unmoving) toy is not as fun to play tug with, so this encourages a fast out. When he outs the dead toy and switches, the new toy "comes to life" for him to play with.
     This work is pretty self-explanatory - we continue to switch back and forth between the two toys, sometimes outing and re-biting the same toy, and sometimes switching; sometimes leaving a dropped toy to come get the one I have, and sometimes ignoring the one I'm holding to access a dropped toy.
     During this work I can up the challenge a bit (and test how good our "out" command is) by continuing to tug and play with the toy after the "out" command rather than making it go dead. This puts more value into the toy and tends to make him less likely to out it, but he handles it pretty well and is rewarded by extra fun play when he makes the switch. He also adds in a bit of his own fun by adding the tunnel into the game, haha.

The next two videos REALLY increase the challenge for Gideon, and this work is BRAND NEW for him. He has never been asked to switch back and forth like this between treats and toys (third video) or between toys of such different values (fourth video).

This third video shows Gideon switching between treats and toys. He is a toy maniac, so the treats are definitely MUCH lower value. I make this value gap as small as possible by starting with high value treats and a lower value toy, but even a low-value toy is much higher value than even a high value treat for Gideon. Here's what happens:
     I start out with treats so that the first switch is an easy low-to-high switch. I ask for a few simple behaviors, and as soon as I "yes" him for the toy, you can see his entire demeanor change - he "lights up" for the toy. You can also see a bit of disappointment when I then make the high-to-low switch back to treats. I help Gideon here by switching pretty rapidly back to the toy, and increasing the value of the treats by letting him chase them and such.
     Then I really challenge him by asking him to drop the toy and telling him "yes" (come to me to get reinforcement). However, he'd prefer to have the toy than to come to me for treats, so he goes to pick it back up. In a perfect world scenario, I should have been holding his leash to prevent him from grabbing the toy again. I "out" him again and encourage him back to me by my movement, then grab the leash to help control his access to the toy and move him away from the toy. We do some "fun stuff" with the treats, chasing them, catching them, feed multiple treats as we run backwards, etc.
     We continue to work though his toy preference with more switching. You'll see him a few times glance back at the toy when we're working with treats. I also am careful not to switch him to treats every time I "out" him - I often reward the out with more toy play so that he doesn't become slow to out, thinking he's only going to get treats out the deal.
     A bit later you see me really increase the challenge - there are toy treats planted on the floor, and with Gideon looking between me and the treats, I show him a toy presentation (just as I would if I were going to tell him "yes" and let him have it), but I instead tell him "get it" and release him to the treats. I help him choose the right one by waiting to release him until he looks at the treats, and also making sure that the treats are closer to him than I am with the toy. As when I was switching between different treat values, I end the session with toy play so that he isn't feeling cheated.

The fourth video is even harder for Gideon (and definitely harder for me, too!) We are working with three different toys of different values: the long, thin tug (no handles) that we have been using previously is the lowest value of the three. The thicker, two-handled red tug is our mid-value toy. And the large jute "bite pillow" is, by far, the most valuable toy reward that Gideon has. Gideon and I are both pretty new to the bite pillow, so you'll see me a few times struggling to manage it properly, and you'll see with all three toys (especially the larger two) that I sometimes have to take a moment to make sure that I am presenting the toys in an appropriate way to keep all my fingers safe! (Gideon has pretty good targeting skills, but even the best dog can make mistakes if the handler isn't clear with the presentations of the toys.)

     Like the previous video, you'll notice Gideon and I making some mistakes (especially when he's asked to switch off of the bite pillow onto a lower value tug). Ideally, we want to try to block our dogs from accessing reinforcement that we haven't made available to them (just like the closed hand blocks the dog from accessing the "off-limits" treats, a leash could/should be used to block the "off-limits" toys.) However, it's a very real struggle trying to manage three toys (one of which is big and bulky) and a dog and a leash. The training won't ever be perfect, but we try to make it as smooth and clear as possible and are always seeking to improve. This wasn't too bad for our first go at this.
     At the beginning of the session, I bring Gideon out and he knows that I have the bite pillow ready. Part of Gideon's training includes offering me behaviors in order to "activate" me and get me to play. This helps build our training relationship by giving him a "voice" in the training process. He offered me an approximation of heel position (typically he offers me "easy" behaviors like spin or down), so you'll hear me praise him for this choice with genuine pleasure/surprise ("Ooohhh!") and then ask him to build on it by heeling a few steps before I reward.
     As we start switching between the different valued toys, you'll also notice that just like I tossed the treats and played around with them to help increase their value, I can also put extra effort and enthusiasm into play with the lower value tugs, and put less energy into play with the bite pillow. This helps even out the differences in value and continue to make it rewarding for Gideon to make the appropriate switches. I also gave some extra help as needed by blocking the incorrect choice or making the correct choice much more obvious to him.
     Another thing you might notice is that I use "yes" in two different ways in this video. I use "yes" to mean that he can grab the toy that I am holding (as I have been doing in previous videos), but I can also use "yes" when he picks up the correct toy ("yes" means that he is getting reinforcement from me, so if he already has the toy in his mouth, the reinforcement is a good game of tug).
     Unfortunately, the video stopped recording, so I didn't get the last bit of the session, but I once again ended the session with the highest value toy.

As a quick final note, it's not always possible to end a training session with the dog accessing the highest value reward. If you're at the park and the dog's reinforcement options are a tennis ball and a squirrel, you're likely going to make him aware that the tennis ball is available and the squirrel is not. At the end of the session you aren't likely to let your dog go chase the squirrel - you're going to have to play ball and call it quits with that. However, when you're in an environment that you can control, it's best to end on a high note with the dog accessing the highest value reinforcement available (of course, at the VERY end of the session, the handler always keeps the treats and toys).

Hopefully the combination of my explanations and seeing it in video form helps with your understanding of value transfer and how it can be used to shape your dog's understanding of opportunities for reinforcement. If you haven't done this kind of training in the past, consider giving it a try. If nothing else, it's a fun game and a good way to test how clearly you and your dog understand one another. :)

Happy training, everyone!
~ Danielle and Gideon

No comments:

Post a Comment