eileenanddogs

Tag: learning from mistakes

Summer Punches It

Summer Punches It

Targeting a plate with élan
Summer targets the plate so hard that her muzzle slips upwards. And look, her mouth stays closed!

I have learned a lot in the last three months. Yes, that’s how long it has taken for me to get Summer’s target behavior where I really wanted it.

Back in September I published a post about the many ways I had messed up Summer’s target behavior. I had lived with it a long time, but it really became a problem when I tried to use a target for distance behaviors because Summer kept biting it and trying to bring it to me.

So I decided to fix it “in public” and published a post with my training plan to fix the problems.

Here is the result!

What Happened

I made a few changes to the training plan along the way, but not all that many.

My training tracker spreadsheet first had the following columns:

  • date
  • criterion
  • number of reps
  • number of correct reps
  • correct reps as percent
  • goal percentage
  • comments

One of the first things I learned was that Summer was not the only one making correct and incorrect behaviors. I was sometimes marking the incorrect response! So I added two columns, one for my number of correct reps, and the other to express that as a percent. That made me clean up my act in a hurry, and pretty soon I was no longer marking incorrect behavior except once in a blue moon.

I also added a column for a moving average, to smooth out some of the noise in the graph and show the trends better. And just because I’m a nerd and I like that sort of thing.

Recall that one of the worst (out of six) problems we had was Summer’s biting the target, because of our retrieve work before we got target on cue.

I had picked a new hand position so as to change the picture completely for Summer, but my choice, which made it appear that I was holding a treat in my fingers, elicited even more teeth and biting from Summer at the beginning. (So ironic, since Summer is not at all a mouthy dog.)

After the first few sessions I came close to changing my hand position again because of all the teeth. But I decided to take the challenge and keep it. I really liked the touches I was getting from her.

Really, this was the hardest part and took the most time. It took a little more than a month of practicing only with my hand to get rid of the teeth. But I’m really glad I stuck with the new hand position because started getting much, much firmer and nicer touches from Summer than I ever had before.

After we got the hand touch, I tried transitioning to a target stick and that was disastrous. Bite city. The stick was a cue for the retrieve work we used to do. So I thought of an object that I could hold that she couldn’t bite. How about the back of a plastic plate?

Putting the spoon against the plate made it less tempting to bite
Putting the spoon against the plate made it less tempting to bite

So we did many reps with a plastic plate with a piece of blue painter’s tape on it. A good Internet friend points out that blue tape is nicely  visible to dogs. After about a month of that, I brought my target stick back out (it also has tape on it), and held it flat on the back of the plate. By making an interim step (splitting), I was able to transition her to the stick without having teeth. This was a huge step, and a good one towards my practical goal of being able to send Summer to a freestanding target stick to touch.

Where We Are Now

Over the weekend I tested Summer on hand, plate, and target stick touches and we got 100% correct! Not only that, but her touches have still nice and firm and she is eager to do it. No more drive bys for sure.

Link to the video for email subscribers.

You can see from the graph that I lifted from my tracking document that there are several dips in performance; those correspond to places where I raised criteria. But even counting those dips, her overall average was 86%. Keep in mind that my goal for percentage correct before proceeding each time was 95%, not the 80% that trainers typically shoot for before moving forward. You can see in the graph that we stayed on each step longer than we would have had to if that were our goal. It worked for Summer and me because neither of us minds repetition.

This graph covers 1,012 correct repetitions. Yes, you read that right. About 1,000 reps. Let that be a lesson. Try to train it right the first time!

Final Notes on Criteria and Method

I ended up changing one criterion from my original training plan. I had specified that I wanted Summer’s mouth to be closed. But  I got visually confused when I saw her approach with an open mouth, then close it just before the touch. I decided that was her business whether she wanted to leave her mouth open, as long as she touched my hand or the object with her nose/muzzle and not her teeth. This worked out for us.

I wrote in my previous post that I wanted to avoid negative punishment if possible. I did end up doing it a few times. Sometimes we would get in this loop where she would do an unacceptable touch and when she tried again, one of the undesirable behaviors would pop immediately back in. So a few times when I got a bite or felt teeth, I not only didn’t give her the treat, I pulled my hand back and paused, with a little break in the action. This was always followed by a correct response from Summer. The penalty did seem to communicate very well that I wanted touches and not bites. I probably did it fewer than 10 times in our 1,000 reps.

At the time it seemed more kind than letting her try over and over again without getting reinforced (extinction). A more skilled trainer probably wouldn’t have had to do either (and certainly wouldn’t have taken 1,000 reps!)

Notes about Future Steps

What’s left, following the Training Levels, is a foot touch (her nose to my foot), then touching a Post-it or piece of tape on the wall, with the final goal of pushing a cabinet door closed.

I don’t anticipate a problem with the foot touch, but the wall thing will be a challenge because we have done lots of wall touches with her paw. But I know how to be patient, and so does Summer.

Some final tasks will be a duration touch, mixing up Zen and target cues, and finally distinguishing target and retrieve cues. And of course I’ll need to generalize every one of these things and take them on the road.

Thanks for reading! I would love to hear more retraining stories. I’m not the only one, am I?

By the way, now that it’s done, here is the whole series in one place:

Eileenanddogs on YouTube

8 Common Dog Training Errors: Cautionary Tales

8 Common Dog Training Errors: Cautionary Tales

As the great trainer Bob Bailey says, training is simple but not easy. The principles are very simple and straightforward, but actually applying them in practice can be very difficult.

I’ve mentioned many times that I am not a professional trainer. But I hang out with some phenomenal ones. Plus, I am a student of life and tend to do lots of observation of myself and others. (What, you had noticed?)  And I don’t mind sharing my own errors if it can help somebody along.

Here are eight of the top booboos in dog training. All of which I have done myself, and many of them on camera! (And guess what, I have 15-20 more! You can expect more posts on this if people enjoy this one.)

And by the way, these are errors made by trainers who are using the most humane methods they know and can learn. The list would be much different if it addressed mistakes made by trainers who make heavy use of aversives. I don’t want to write that list right now!

What Can Go Wrong?

(1) Allowing the dog to rehearse the behavior you’re trying to eliminate. One of my trainer friends says this is the #1 problem she runs into with her clients. A client will say, “I bought my dog a bed for his crate and he chewed it. I bought him another one and he chewed it. I bought him another one and he chewed that one, too. I’ve bought him five beds and he’s chewed them all.”  That dog is getting really skilled at chewing beds, and has found a way to occupy himself in his crate.

If a dog has a problem behavior, it’s because that behavior has been reinforced. (Chewing is fun for dogs!) Usually not deliberately by us, but reinforced nonetheless. So we don’t tend to “count” it in our minds. But it’s just as if we had given the dog a cookie every time he chewed the bed, or jumped on the couch or whatever. But if we want to teach them to do something different, we also need to prevent those fun rehearsals of the “wrong” behavior and stop that reinforcement.

Desirable door behavior
Desirable door behavior

For example, you want to train your dog not to crowd up to the back door and rush out when you open it. That behavior is reinforced every time she does it if you immediately let her outside. When you start to teach polite door manners, you doubtless start the training with an “easier” door. You practice with an interior door, and work on the behavior  you want. You reinforce with food.

Undesirable door behavior
Undesirable door behavior–see the linked video–>

But in the meantime, every time you take your dog into the back yard to potty, she continues to crowd up to the back door and rush out when you open it. Getting outside is obviously a potent reinforcer and nothing has happened to stop that. To get the new behavior in place at the back door, you will have to prevent your dog from practicing the old one. That can be a challenge without using aversives, but can be done. With three dogs, our back door behavior is a work in progress, but those bad habits do on occasion get reinforced. And reinforcement “on occasion” is enough to keep them alive and well.

The classic example of rehearsed behaviors and conflicting reinforcers is walking on a loose leash. If you let your dog pull you, using the same gear you use when trying to teach loose leash, that behavior gets reinforced every time that pulling gets the dog where he wants to go, just as purely as if you had given him a cookie for it. That’s why trainers always tell you to cease walking the dog or at least use completely different gear if you must walk the dog, if you are simultaneously trying to teach leash manners. You are shooting yourself in the foot otherwise.

(2) Lumping. Lumping means failing to break the behavior you are trying to teach into small enough steps. For instance, you are teaching your dog to get on a mat and lie down. Your dog is beginning to understand your cue. You stand right next to the mat and give the cue. Dog lies down. You stand two feet away from the mat and give the cue. The dog goes to the mat and lies down. You put the mat in the corner of the room, go back to the center of the room with your dog, and give the cue. Your dog says, “Huh?” Lots of misunderstandings between humans and dogs could be ameliorated if we just took enough baby steps when teaching a behavior. Here is a post about lumping, with a video starring my dog Zani, who lets her disapproval of my behavior be known.

Small black and tan colored hound looking at her trainer with her mouth open. There is a piece of tape on the wall behind her.
Zani says, “Quit that lumping!”
Zani can sit on a crate
Sitting on a crate was Zani’s idea

(3) Neglecting to generalize cued behaviors. Dogs are great discriminators. They notice every little thing in the environment. It’s all pertinent to them. They are not as good as we are at generalizing. And it is very very hard for us to get that through our human heads.

If a dog knows “sit” when in the kitchen facing east, she may not know it while standing on the piano bench in your front room facing north. And she almost definitely won’t know it if you lie down on the floor next to her or stand on your head before giving the cue. All along you thought she was responding to your verbal cue, but she actually was responding to the fact that she was standing in the kitchen facing east, and you have a clicker and treats, and you said a word (which turns out that it probably could have been any word). All of that was actually her cue. I have several movies showing dogs who respond incorrectly to a cue because of the human failure (um, mine) to help them generalize.

And here’s my moment to plug Sue Ailsby’s Training Levels. The Training Levels are the best training program I have ever seen to build generalization into every step.

(4) And that leads to…not knowing what the dog’s cue actually is. It may not be what we think it is.  They are generally paying attention to our body language and props way more than we think. Almost all of us (border collie owners excepted, grin), think that our dogs understand verbal cues better than they actually do. I have lots of experience with this, having been blessed with two dogs who are probably less proficient than average at verbal cues. I have yet to teach them, despite many repetitions, some simple word discriminations that some breeds and individual dogs seem to pick up very fast. To really know whether your dog understands a verbal cue or not, you need not only to practice generalization, but get yourself out of the picture for the final exam. I have a fairly embarrassing post and video where I was trying to test my dog’s knowledge of the difference between her “crate” and “mat” cue. I didn’t realize until I saw the film that I was still cuing her with my body.

(5) Trying to train when dog is too stressed. Just about anybody who has gone to a dog training club has either done this or seen this. The dog is trying to take in this noisy, chaotic environment full of other people and dogs. Even if the dog is not scared of strange people or dogs, the noise and chaos make everything difficult. This is where it pays to know and observe your dog, and do some homework on dog body language.

It's not hard to tell if Clara is too stressed to train
It’s not hard to tell if Clara is too stressed to train

Here are some posts, photos, and videos of mine showing a stressed dog, a fearful dog, my own shut down dog, and some other shut down dogs. While some dogs can respond while stressed or fearful, a good teacher will help you work with your dog to get her comfortable in the environment before ever starting to teach “behaviors.” Most would agree that learning to be comfortable in a difficult environment is a more important lesson than learning to sit on cue. And it will really pay off in the long run.

(6) Not being aware of the dog’s responses. This is a more general version of #5. Once one starts learning about dog body language, it can be a revelation to watch videos of one’s own training sessions. You may learn that you are making your dog nervous when you ruffle his fur. Or that he ducks strongly away when you reach down to pat his head. You may learn that you are leaning far to much into your dog’s face. You may learn that you are unwittingly doing something to jazz up your dog when you are trying to teach a relaxed, duration behavior. Or like me, you may learn that your dogs are way more sound sensitive than you thought. (I had the camera on behind my back when I recorded this video.)

(7) Using a verbal cue too soon. I mean, we want to do it at the very beginning, and that’s exactly what not to do! It seems to be almost innate for us to assume that dogs do or “should” understand our own native human language. If you grew up in the “say ‘sit’ and push the dog’s butt down” generation, it is hard to imagine anything else. But remember, the word “sit” is a cue. It is a green light that says, “if you sit now, reinforcement is available.” But attaching that cue to the behavior is a latter step, not the first step. Chanting “Sit, sit, sit,” does not instruct your dog what to do. And doing the butt push move means that the first thing that Sit means is “Mom is about to push my butt down.”

The better way of looking at it, as Sue Ailsby describes it, is that we say to the dog, “That thing you are doing–we’re going to call it ‘Sit,’ OK?”

The first step is to get the dog sitting repeatedly. You can lure, capture, or shape. But it’s a good idea to keep your lip zipped and don’t let that cue come out. Some people say you shouldn’t use the cue until you are willing to be $100 that the dog will sit.

Here is a little video I made of Clara as a puppy. This is the very first time I had used a verbal cue for “down.” I had already gotten predictable and repeated downs by arranging the environment that way. Do you see how startled she is the first time I use it? It’s only the second verbal cue she has been exposed to. As far as she’s concerned, I have interrupted the game we were playing.

(Watching the video today makes me realize I know her better now. According to #6–watching the dog’s responses–I might have ended earlier. Her tail is wagging with a rather low carriage and she is working very hard for a wee one. She was more stressed than I would like.)

(8) Rate of reinforcement too low. “Rate of reinforcement” means the number of reinforcers per unit of time, for instance, per minute. When teaching a new behavior, and particularly to a dog who is new to training, it is desirable for the RoR to be very high. Dogs who know you have treats and are playing a training game but can’t figure out the rules of the game tend to get frustrated, lose interest, or even wander off. It is our job as trainers to arrange the teaching so that the pupil can have a high success rate and therefore a high rate of reinforcement.

Here’s a video of some behavior drills I did with three of my dogs. The highest rate of of reinforcement was in Zani’s first session: 12 treats in 43 seconds. That’s a treat about every 3.5 seconds, or 17 treats per minute. All of the other sessions shown came to a treat about every 5 seconds, which would be 12 treats per minute. Many of the behaviors had a little duration on purpose, so that’s not bad. I checked the video of puppy Clara I linked to in #7 above, and I delivered a treat about every 5 seconds in that one as well.

By the way, I made this list of booboos before I realized that I had examples of so many of them! How about you? Have any suggestions for my next post on this? Bonus points if you’ll step forward and show an example!

Other Resources

Here are a few other articles on common errors. I got a few ideas from them, but most of the above were from my own head and some trainer friends’ heads.

This post is part of a series:

Coming up:

Eileenanddogs on YouTube

Using a Training Plan to Retrain Summer’s “Target”

Using a Training Plan to Retrain Summer’s “Target”

Summer's new hand touch
Summer’s new hand target

In our last episode, I shared how I messed up Summer’s hand (and object) targeting behavior six ways from Sunday. Now I will share the process of retraining it.

When you follow directions from a book, such as the Training Levels, a lot of the planning is done for you. So I get a little lazy about training plans. I can just check little boxes off in the books.  (A reminder: I  acquired all these errors when I was brand new at training and using a mishmash of methods. Wish I’d known about the Levels earlier.)   But it’s a good idea to always have a plan, and collect data on what you are doing. Since I now need to do a unique retraining plan suited exactly to my dog and her needs, I am going to make a plan and share it, then share how well it works.

Training Plans

Training plans can be as simple or as detailed as the situation demands. For this situation, where I am trying to get rid of several superstitious behaviors that often follow my cue for hand target, I am going to make a thorough plan, and share it here.

Here are two posts about training plans, both by great trainers.

I combined parts of both of these to design the categories for my plan. I also made a record-keeping spreadsheet in Google Docs, loosely based on Melissa Alexander’s. Hers is accessible through her post above.

My Training Plan

  • Goal: a clean touch of Summer’s nose to my hand, followed by her generalizing that to similar touches to different objects. I want verbal cue recognition (will do tests with objects, see below).  But she doesn’t have to wait for a verbal if I do the hand signal.
  • Description:  A clean clear touch of nose to hand or object. She can be in any position that will allow her to reach the hand or object. It doesn’t have to be a hard touch, just definite touch of nose. No drivebys, and no just whiskers. No teeth, no open mouth. Minimal paw lifts. I define minimal as: her paw can lift about an inch higher than normal if she is walking or trotting to the target. Getting her mouth closed and preventing paw whacks are essential. A little leftover paw action is OK with me.
  • Methods: Capture the touch, then shape a firmer touch if necessary. I want to make the picture as different as possible for Summer from the very beginning, including changing the hand signal and verbal cue. I will follow the progression in Level 1 Target in the Training Levels. I will start with me seated. Use Sue Ailsby’s hand position (see “new position” above). Start off with my left hand rather than right, which I have used more often for hand targets before. I’ll drop treats rather than handing them to her (encourages mouth/hand contact) or throwing them (builds excitement).
  • Cue: Verbal. In the case of hand touch, presentation of hand.  Cue discrimination: the ability to distinguish from Sit and Down on verbal alone. For this I will use a standalone object, since the presentation of the hand will always be more salient than the verbal. When to start with the cue: TBD.
  • Sessions: Up to three sessions per day of 10 treats.
  • Criteria for advancement: In the early stages of the hand touch, 95% or above. This is because my goal is to clean out the old superstitious behaviors. Also I have observed that Summer doesn’t mind lots of repetition. Later I will build in her looking me in the eyes before I will give the cue. This is because of her habit of staring at the food or my food hand.
  • Duration? Not for this project.
  • Distance? 15 feet to object, or about that much if chasing me.
  • Distractions? Maybe near the end. Put down a mat for her to go by as she goes to touch an object.
  • Position: Hand touch from all different directions. Object touch from different positions. I will limit to objects already in her sight, i.e., she doesn’t have to turn around to find it. However, I plan to “try it cold” by cuing a Touch when she is not expecting it and when there is an obvious object to touch.
  • Where: Start in my den. Do other rooms in house, back porch, back yard. Possibly go on to front porch.
  • Reliability: I want 95% free from superstitious behaviors. Response to cue itself 80-90%.
  • Comments and caveats: Since we have an ongoing issue with staring at food, I will chain in eye contact after she is getting some fluency.  She is more likely to do the undesired behaviors if she is excited and moving fast, so I will start with her standing still. Observation: she is quite likely to offer an undesired behavior after failing to meet criteria and doing a light touch on the first one, instead of offering a firmer touch. I will need to be creative and use positioning to avoid errors. Also I stated earlier that I don’t want to use negative punishment at all if possible. That means I don’t want to rely on pulling the target away from her if she is approaching it with her paw or an open mouth. I want to prevent those things from happening to begin with. I want to tell her through reinforcement what is working.
  • Future:  Duration. Mix up Zen and target. Learn to distinguish target cue from retrieve cue.

The difference between my old and new hand positions for target:

Notes about Future Steps

In the Training Levels, what follows the hand touch is:

  • Foot touch:  (Dog’s nose to human foot) Probably no problems here.
  • Wooden object: I’ll need to prevent teeth touches and grabbing by using a large, flat object, as described in the Levels (p 187) Need to watch for feet movement. How to discourage? Careful height of object. Experiment with stationary vs moving.
  • Plastic object: ditto.
  • Metal object: ditto.
  • Spot on wall: I’ll have to modify the instructions: I won’t use a post-it note or painter’s tape. (Watch the Targeting Mishaps movie to see why.)  I’ll draw or paint a target on a piece of poster board with non toxic paint. Start by holding the board. Shape touching the spot. When that is solid, get it onto the wall.

We have practiced all of the above behaviors before, but many incorrectly because of superstitious behaviors.

Session Planning

Session 1. I’ll sit in a chair. Treats on my right on a desk. Proffer left hand in position described by Sue. Correct iterations marked by Yes and drop (don’t throw) treat.

Link to video for email subscribers.

My Notes after the First Two Sessions

Wow, real life comes crashing in. So Summer did one touch/sniff, then the very next one she took all my fingers in her mouth. (A “bite” but very inhibited. Her teeth didn’t close.) I wasn’t ready for that at all. I was in the middle of saying “Yes” but aborted it. I was so surprised I just got up and turned off one of the cameras and took a break. In the meantime Summer heard me say most of “Yes” and was sniffing around looking for her treat, which I had made a split second decision about and didn’t give her.

Dang! An important goal for me is no negative punishment, but abruptly getting up and stopping a training session can be a big dose of that….

But the video taught me a lot. Both the times (yes, it happened again) Summer took my fingers in her mouth, I had presented my hand kind of flat. Must have looked like I was handing her a treat.

Besides the position of the hand, I need to make its presentation a little clearer (I don’t need to leave it halfway out there). Make it very clear: on/off. I’m still struggling a little with the hand position; that’s part of why I am so stiff. Also I’m trying to keep my body very quiet. A couple times I was too slow and she was already moving forward when I presented my hand.

I’m really really glad I counted reps and successes. I would have overestimated our success rate otherwise.

Also, I chose to go with 10 treats rather than 10 total iterations. 10 treats means 10 correct responses, but puts no limit on incorrect responses. Sometimes not advisable at the beginning. But even looking at the video I had a hard time deciding what “counted” as an iteration or not, so I’m glad I wasn’t trying to count while training.

Third and Fourth Sessions

We have already had our third and fourth sessions, although they’re not included in the movie. Our success rate got better and went up to 10 correct out of 13 both times, which comes to 77%. I tried to loosen up a little and move in Session 4 but I immediately got an open mouth from Summer. I’ll need to continue to be very conservative since movement on my part has typically triggered mouthiness on hers. There’s always a fine line between getting the behavior and not wedding it to a certain setup. I’ll do some other things to introduce some variety.

Here is my training tracker document. I’ll keep it up to date and publicly accessible.

Thanks for reading.

Now that it’s done, here is the whole series:

Also coming up:

Eileenanddogs on YouTube

6 Ways I Messed Up My Dog’s Targeting

6 Ways I Messed Up My Dog’s Targeting

Targeting done right! --credit Marge Rogers
Targeting done right! –credit Marge Rogers

Hand targeting is usually suggested as a great behavior for new clicker trainers since it is easy to get and easy to define criteria for.

I guess I didn’t read the brochure carefully enough because I messed up hand targeting for one of my dogs six ways from Sunday!

From time to time I share in the blog mistakes I have made in the past, Continue reading “6 Ways I Messed Up My Dog’s Targeting”

Wordless Wednesday: Bopping the Wall

Wordless Wednesday: Bopping the Wall

This is a video follow up to one of my more embarrassing posts: “Lumping It: A Public Service Announcement.” In that post I described, and showed in the video, how I repeatedly raised my criteria too quickly when trying to teach Zani to nose target a piece of tape that I first had in my hand but was working toward putting on the wall.

Zani bopping the wall
Zani bopping the wall. Sorry it’s blurry: that girl was moving fast!

Despite my errors, Zani is now doing a great job finding and touching tape and sticky notes on the wall, and enjoying doing so.

We have varied the objects: tape, sticky note, band-aid. We have varied the height a little but she doesn’t know yet that it is OK to get up on her hind legs to reach the target. That’s next!

Here is a link to the video for the WordPress email subscribers. 

Thanks for reading! Stay tuned for:

Eileenanddogs on YouTube

A New Resource, and our Rally Weekend

A New Resource, and our Rally Weekend

I have published a permanent page on my blog that collects all the posts and videos I have made that I have been told are useful for dog trainers to show their students.

It can be accessed here:

Video Examples for Teachers

but also is listed on the permanent menu above. I hope it is helpful. I will be adding more material as I develop it.

Our Weekend

For those of you who saw my last post on practicing Rally with Summer and attempting to reinforce appropriately, here are some pictures of how we spent our weekend. We trial very infrequently for a number of reasons, so it is a big deal for me when we do.

On Saturday we both worked hard but it wasn’t fun like it can be. I tried hard to make it easy and fun for her, but there were various stressors. We held it together in a difficult ring with an 88 and fourth place.

But on Sunday it was magic. We had a lovely run, stayed connected, and Summer stayed happy despite the difficult trial environment. I am pleased with the sequence of photos below that show her eagerly taking the jump, then beautifully collecting and checking in on her landing (fourth photo).

We scored a 98 and took first place. My unlikely competitive obedience dog and I.

Summer jump sequence 1
Summer jump sequence 2
Summer jump sequence 3
Summer jump sequence 4
Summer jump sequence 5
In the ring for awards
Accepting our blue ribbon
Superstitious Behaviors in Dog Training

Superstitious Behaviors in Dog Training

Definition of a Superstitious Behavior: Accidentally or unintentionally reinforced behavior where a behavior is reinforced but the reinforcement occurred by random chance instead of in accordance with a specific contingency.   —From the Association of Animal Behavior Professionals glossary

My little black cat Arabella never brought me bad luck

I wrote when I started this blog that I was going to share my mistakes in the hopes of helping others learn. Here are some nice big embarrassing ones regarding superstitious behaviors, but at least they date mostly from my earlier training days. Hopefully you, or any beginning trainer, can benefit from the lessons I learned the hard way.

The Terminology

B.F. Skinner first described superstitious behaviors in experiments with pigeons in 1948. He set a feeding mechanism to trip at variable intervals that had nothing to do with the actions of the pigeons. The pigeons nonetheless started repeating behaviors that had been “accidentally” marked and reinforced by the feeder.

The term “superstitious behavior” now refers to any behavior that is accidentally reinforced. A couple of the behaviors in this post stretch the definition. But even if they aren’t technically superstitious, they are nonetheless accidental or at least poorly trained on my part.

Summer’s Nod

When one of my agility buddies encouraged me in 2008 to start using a clicker, I didn’t know that I should practice timing. I didn’t know that there were mechanical and observational skills involved. A clicker seemed like a fun thing and I had heard that dogs got motivated and enjoyed it. So I also didn’t know that it would be wise to start with a behavior that involved only gross motor movements.

Uh oh.

The very first behavior I actually got with a clicker was a head nod, even though I was trying to click for eye contact. I realized this after working on this a few weeks with Summer. Summer would move her head to look at me and I would click. I clicked the eye contact but apparently also clicked the nod.  Strangely, the nod drifted to the period after the click and before treat delivery. The sequence went: eye contact, click, head nod, treat. The nod, immediately preceding the food, accordingly got a ton of reinforcement. Even though she did learn (despite me) that I was trying to teach eye contact, the head nod remained.

Four years later, I still get little nods from Summer. Interestingly, she doesn’t offer it in shaping sessions. When it comes back, it returns in its old place between the marker and the treat when I have just clicked her for something else. This is an example of a superstitious behavior.  And it turns out that I am really good at creating those!

It has faded some over the years, but I found a couple of examples. Want to see?

Zani’s Weave Poles

The following is a behavior that would have been very difficult to teach, had I intended to do so.

In the course of teaching Zani agility weaves using the two by two method, I would tend to mark with a “yes” the moment she made the turn between the last poles. This was the moment I was absolutely sure she was going to complete the behavior correctly. That’s a natural time to mark. But early on in our training, she did a few little jumps through the last pair of poles. I marked, and you can see what happened.

But what is most fascinating is that she does it only when I am on her right side. When I am on her left, she doesn’t do her “jump thing” between the last poles. I speculate that since she is very spatially sensitive, she is less likely to hurl herself out of the weaves when I am close to where she will emerge. Or perhaps I just didn’t mark the exit as much when we practiced on that side.

Clara’s Circles

Those first two behaviors are pretty cute. This behavior of Clara’s that I accidentally reinforced is rather unfortunate.

Clara has always been pushy. When she was about three months old I started a training project of reinforcing her for walking a few feet away when I was interacting with another dog. I started off with the other dogs in crates and was very systematic about it. I drew lines on the floor for my own benefit so as to keep consistent criteria about Clara’s distance from the other dogs. We did lots and lots of sessions where she would walk away a few steps and reorient at some distance.

It would have been better if I had taught a default down or Go to Mat, or at least thrown the treat away from our immediate area. I ended up unintentionally reinforcing a circling behavior. She would walk a few steps away, turn and reorient at the desired distance. I marked the turn way too often, when what I wanted for her was just to back off. But the “backing off” was not a well-defined behavior, even with my lines on the floor. So I ended up clicking an observable behavior, and that was when she turned back to me. Dang.

What is so unfortunate about this is that the circling either morphed into a stress behavior or it was one already. Because I have seen a lot more of it ever since those sessions. Clara tends to do it when I don’t mark a behavior that she expects to be marked. She will immediately whirl around, usually counterclockwise, then often retreat to a mat.

It is impossible to tease apart how much of this is due to all that early reinforcement, and how much of it is a natural stress behavior for her. I do wish I hadn’t trained so many 180 and 270 degree turns when she was young. When I set out to teach her spinning as a trick, it was dead easy, but I gave that a second thought and decided not to use that trick.

Cricket, Too

I even taught superstitious behaviors with Cricket. I tried to train a paw lift as a “wave” trick, but then it started occurring in her “sit” position as a superstitious behavior. Once it started, I kept accidentally reinforcing it. She almost never put her left foot down when she sat for me again. I didn’t know enough at the time to fix the problem I had created.

The way I first taught the behavior was not great either. A friend had suggested holding a treat in my hand and clicking Cricket for pawing at it, then fading the hand and treat. Such a bad idea in so many ways. Reinforcing an enthusiastic digging terrier for pawing at my hand? Ouch.

How To Avoid Training Superstitious Behaviors

I wish I could give some succinct, pithy advice that could keep other newish trainers from doing this. When choosing what behaviors to teach and how to teach them,  it takes experience to learn to predict the ramifications.

Here are the best suggestions I have.

  • Answer a few questions. Is there a persistent extra behavior that is happening when I train this behavior? What’s going to happen if that extra behavior sticks around? How can I get rid of it? If I’m training a trick—might this extra behavior or the trick itself turn up where I don’t want it? Might it interfere with behaviors that are actually more important to train?
  • Video yourself. If you don’t have a teacher, you can learn a lot by recording your training sessions, and if you are brave, showing those recordings to online friends if you don’t have a teacher or local training buddy. People can give much better counsel if they actually see what you and the animal have been doing. Most of us humans could use a lot of work on our observation and description skills. Cameras do a lot better job for a lot of us.
  • Get expert advice. I didn’t have a teacher to ask when I taught most of the behaviors I’ve described here. A professional would have seen most of my mishaps in an instant and showed me how to head them off.

A Success!

One thing that I got right: I started training that “backing up the stairs or wall” trick that was going around a while back. Zani just loved it and started getting good at it. It was great for hind end awareness.

But then one day when we were practicing our two on, two off agility contacts, she overran them and happily backed up into position. That would be a fault in many agility venues. I immediately stopped training the trick. A more experienced or patient trainer could certainly have both behaviors, but sometimes I realize my limitations. The risk wasn’t worth it to me.

Here is one more superstitious behavior Clara and I collaborated on. This is an example of something that is cute when a puppy does it, but can get pretty tiresome in a grown dog. Of course it’s still cute, but who wants their fingers licked Every. Single. Time. They go to open a crate door?

OK folks, please tell me I’m not the only one who trains silly behaviors by accident. Does anybody want to say what they have done? Or are you all perfect?

(Here’s a link to a one minute video that shows all five behaviors, to make my humiliation complete.)

Many thanks to Joyce Loebig for suggestions that improved this post!

Coming Soon

Copyright 2012 Eileen Anderson

Lumping It: A Public Service Announcement

Lumping It: A Public Service Announcement

So maybe you are new to clicker training and you keep hearing people talking about lumping and how bad it is. Be a splitter, not a lumper, they say. You have a vague idea about it but maybe aren’t exactly sure what they mean except that lumping is bad.

Or maybe you are a teacher and you would like a really clear example of lumping vs splitting to show your students.

Do I have a video for you! Continue reading “Lumping It: A Public Service Announcement”

Ant-Sized Treats

Ant-Sized Treats

No tiny treats for my dogs!

 

 

 

 

 

 

 

 

 

 

 

 

 

Note: there exists research regarding the effects of using multiple smaller treats vs larger bites (aka magnitude reinforcement) when training.  But the basic premise that there exists a size of treat that is “too small” for an individual dog also holds true.

How often have you read the following words? “For clicker training you need some tasty treats for your dog. We recommend treats about the size of your pinky fingernail.”

If you Google “clicker training treat pinky nail” you will get page after page of hits with variants of this advice. Except some of them caution that the treats should be no larger than the pinky nail. Some say half the pinky nail. I saw one that said a quarter of the pinky nail.

If you spend any time on the clicker training Yahoo or other discussion groups, you will read conversations about treat size in which people practice one-upmanship regarding how many tiny pieces they can cut out of a single hot dog.

Guess what? This advice negatively impacted my training for years.

Many people have legitimate concerns about their dogs’ weight. Also, I think there is a bit of sensitivity in the clicker training community to criticisms by trainers who use other methods. I think some people want to minimize the whole treat thing.

At your peril.

Here is my story about treats. Summer, my crossover dog, borders on hypervigilance and does not appear to have been selectively bred for an abundance of the “joy with working with humans” genes. She is very environmentally turned on and it is very, very difficult to get her attention outdoors. So naturally, I decided to do agility with her. An offleash sport often pursued outdoors. In fields often bordered by wildlife habitat or in rodeo arenas loaded with animal smells.

For our first year or two, I didn’t have a private teacher, so we just struggled along. I did learn to use high value treats, but I cut them up nice and small as directed. Still, I doled them out fairly generously.

Summer and I got a great agility instructor after about two years. She encouraged the high value treats. I used higher value treats than anyone else in my class, and doled out more of them. Yet after about another year, I was still struggling to get Summer’s attention. My instructor started talking to the class about giving treats generously enough. I listened but I was sure she wasn’t talking to me. As I said, mine were better and I was giving out more.

Then one day in a private lesson, my teacher remarked that the treats I was using were pretty small. I immediately said, “Oh, those are my pocket treats. I have bigger pieces in the food container that I throw at the end of the sequence.” She didn’t say anything else that week. But the next week she took another look and said, “These are just too small. These are ant-sized treats!” I didn’t ask whether she meant they were the size of ants or or appropriate to feed ants. It didn’t matter. Both were embarrassing. I was still a little resistant to her comments since I always passed out several of the small treats. But I did as she suggested and started cutting up much bigger treats.

Around the same time I told her that I knew of something that Summer loved but I hadn’t ever tried. We had been struggling to get her attention for a year or two, remember. It was baby food. She asked me why I hadn’t used it and I said, and this is true, that I was afraid of “treat inflation” and that I needed to leave something at the top to use later. She kindly suggested I drop that concern.

I wrote about what happened next on a list in March, 2010.

But over the weekend, I tried something new. I took two dogs to our agility lesson. My highest value treats were pieces of commercial meatball, thrown in a food tube at the end of the run, and baby food in the jar for contacts. Even my
“pocket” treats were chunks (not little pieces) of chicken skin, hamburger omelette I made for them (just hamburger cooked with eggs), and hot dog. I know, horrifyingly fatty and gross.

And you know what: my dogs performed with the intensity and enthusiasm of my dreams. Like never before. Boy did I feel stupid. And I probably didn’t end up giving them that much more than usual in terms of calories, since one piece went a long way.

I had been like the proverbial frog in the hot tub, who ends up boiled since he doesn’t notice the rising temperature because it is so gradual. I have been settling for lackluster performance without even knowing it. Last week I would have called my dogs enthusiastic. Now I know better.

Dark meat chicken chunks for agility training

In addition to the very high value food treats, we also started reinforcing Summer in agility sequences by letting her play in water sprayed from the garden hose. It turns out she will do almost anything, with speed and excitement, for a chance to play in the hose. And the speed and excitement have “stuck” in her agility performance.

More than two years later, I still tend to use pretty high value stuff for training, but you know what? Summer gets turned on for training whatever I use. Indoors can be pieces of Natural Balance roll, kibble, goldfish crackers, or even bread. Outdoors, and for longer or more difficult behavior sequences anywhere, it is generally meat, fish, or purees thereof in a squeeze tube. And now even outdoors, my  high prey drive, curious dog keeps an eye on me all the time to see if we might do something interesting together.

Weaving for white bread

This is still a little difficult for me to admit to. It feels like I “bought” my dog’s attention. But either you’re a positive reinforcement trainer or you’re not. And if you are, part of the process is finding out what is reinforcing to your dog and using it. If you aren’t getting great results, you try something different. If I had had a typical border collie or retriever, I might have gotten equal enthusiasm from the start with something lower value. But I had Summer, and I (OK, my teacher did actually) figured out what turned her on.

While preparing this post, I needed some photographs. I cut up some hot dogs into “ninths,” then tiny pieces as in the pinky photo above. The dogs were excited by the smell of hot dogs, which they don’t often get. After the photo session I had a pile of tiny hot dog pieces, so I tried them on Clara.  But when I gave her the treats, even two or three at a time, she acted as though she wasn’t sure she had gotten anything. And this is a dog who will happily work for kibble much of the time. The pieces were just too small.

In some circumstances it seems to be very effective to dole out several treats over a time period instead of one big one. But I think even then, there is a minimum effective treat size. I’ve got two dogs (Summer and Zani) who clearly enjoy a nice big piece of good stuff for a difficult job well done.

I’m sure there are plenty of dogs out there who would be delighted with the hot dog treats I cut up today.  I’m not prescribing a treat size. I’m suggesting that we all listen to our dogs about what they want. Most importantly, don’t assume that the common recommendations apply to all dogs.

For two years now whenever I’m on a list and someone starts talking about tiny treats, I have a knee jerk reaction, and write a semi rant in response. Now I can just refer them to this post.

Has anybody else experimented not only with different foods but the size of the treats? Do your dogs like rapid fired smaller pieces or a big chunk?

Addendum, 8/24/12

Two astute readers have mentioned in the comments some things that I should have included. Pawsforpraise pointed out that you need to make sure not to make treats too large because of the danger of choking, especially in rapidfire situations. Good point. Marjorie M. also mentions pancreatitis and the dangers of too much fat in the diet. Also a very real concern. You can read the discussion in the Comments below.

Both of these points reminded me that I didn’t say anything about the need for balancing out the rest of the dog’s diet when they are getting some rich training treats. Summer and Zani, for instance, only get treats like the dark meat chicken above in one, maybe two (active, outdoor) training sessions per week. And I adjust their meals accordingly every time we train. I figured that to be self evident, but I shouldn’t have. My own problem with the tiny treats was caused by taking something too literally, so I sure don’t want to omit some practical concerns here and send anyone flying in the opposite direction!

Thanks for reading.

Related Posts

Copyright Eileen Anderson 2012

Eileenanddogs on YouTube

The Missed Cue: Attack of the Zen Field

The Missed Cue: Attack of the Zen Field

Eileen is seated on a short stool and Clara is lying on the floor. They are looking into each other's eyes. There are some training props on the floor.Up to this point, all of my Missed Cue videos have been set up. After I discover or suspect a hole in my dogs’ training I set them up in a situation in which I’m guessing they will fail, and record it as a teaching exercise. (I discuss why I don’t think this is a mean thing to do in the original post about missed cues: Dogs Notice Everything.)

But this one was not set up. It was during a normal training session. I thought I had the bases covered. And I had the camera running.

The behavior we were working on was Level 2 Go to Mat, Step 3 from the Training Levels books: Dog goes 5 feet to the mat and lies down. Clara has been getting on mats and being reinforced for that since the day she arrived. She can go to a mat on a verbal or hand signal from at least 20 feet away. She can stay on it for extended periods (20-30 minutes). She has a verbal cue, a hand signal, and two environmental cues to go to mat. She can do it when I run in circles around her, when the other dogs are excited, and in many other challenging situations.  So I really thought we had this covered. But when we are working on the Levels, we never skip steps. We train every step as if we’ve never done it before. You’d be amazed what we find out by doing that.

I was amazed today. We got to the Comeafter.  The Comeafter in this Step is to add a distraction. In the book, Sue talks about taking care in picking our distraction. And I thought I was being careful. I picked putting some food on the floor as our distraction. This is old hat for Clara. She has training sessions with plates of food on the floor, can do recalls past food, etc. She has very close to a default Zen during training. And this was only a 5 foot trip to the mat.

What could possibly go wrong?

(There is a synopsis of the following video at the bottom of this post.)

I managed to do exactly what Sue warns about in the book if you make a poor choice of distraction. I made Clara so crazy she wouldn’t go to the mat.

This problem is different from those shown in all the other Missed Cue videos. They involve generalization issues with behaviors for which the dog knows the cue in some environments/situations but not in others. This one is more like the conflict of two cues, one verbal, the other environmental. Clara certainly appears to understand what I am asking her to do and just can’t figure out how to reconcile it with other strong default instructions.

The more I think about it, the more understandable Clara’s behavior is as she shies away from the food and won’t/can’t go to the mat. We teach Zen by reinforcing the dog for moving away from the treat. That is a definable behavior, as opposed to “not eating the treat.”  And when we train it, most of us like to see the dog getting very distant from the treats, and we reinforce accordingly.

So how can I re-train this? Clara needs to know that she can pass close by the treats as long as she doesn’t eat them.

Also, why, in the second go round, does she not take the straight path I have made for her to go to the mat? She wouldn’t have to come within 2 feet of the treats. Anyone care to speculate about that? That part I don’t understand. I do note that in both cases she seemed to feel “safer” from the treats when I was standing near her.

I know we are not the only ones this has happened to. Sue has at least one photograph in the Levels book showing one a dog shrinking away from a treat on the floor. And Sharon Wachsler, a great service dog trainer, came up with a name for the thing that she modestly mentions lots of us have noticed: the Zen field. The Zen field is the invisible area around the treat that only the dog knows the boundaries of. Sharon is the only trainer I know though who deliberately manipulates the field during training: taking treats in and out of the field and extending the field by adding treats within it and changing its shape.

I am hereby asking for suggestions on how to retrain Clara to get closer to the treats, and not freak when she is asked to walk close by them.  In other words, we need to shrink the Zen field but retain its potency. Seriously, we need some suggestions. I have only one idea and it is very mundane. I bet some of you can come up with some clever ideas. I’ll choose whichever suggested method seems to fit Clara’s and my skill level the best and video the progress and results.

Discussions coming soon:

Synopsis of the embedded video 

The Missed Cue: Attack of the Zen Field

Scene 1: We see Clara having a training session with Eileen. Clara is practicing dropping a piece of knotted rope into a bowl, and there is a plate of treats close by on the floor.

Scent 2: We see Eileen calling Clara, who runs full speed past a plate of treats to Eileen.

Scene 3: We see Clara running to her mat with Eileen, but plopping down and staying without a verbal cue as Eileen continues running by and going out the back door.

Scene 4: We see Clara going to her mat and lying down on verbal cue from two different directions.

Scene 5: We see Eileen put some treats on the floor next to a mat, then verbally cue Clara to go to the mat. Clara looks at the treats and scoots a bit sideways away from the mat. She looks away, then looks back at the treats several times. Eileen changes her own position closer to the treats and cues mat again, and Clara slowly goes around and get on the mat, sniffing it as she does so.

Scene 6: A silly repeat of Clara shying away from the treats with animated flames coming from the treats and the music from the shower scene in Psycho.

Scene 7: Eileen again places treats on the floor near the mat, but this time on the other side, leaving Clara a clear path to the mat. When Eileen cues mat, Clara again slips off to the side and puffs with her mouth and circles around. Eileen encourages her to come to the other side (actually closer to the treats). Clara eagerly comes that way, then stops very short when she gets close to the treats. Finally Eileen puts her foot over the treats and Clara goes by and gets on the mat. Eileen is chatting reassuringly to Clara throughout this.

Theme: Overlay by Kaira Extra Text
Cape Town, South Africa