Thursday, December 15, 2011

On Dialouge

Wouldn’t it be better to say, “In my experience, positive training does not work.”? That phrase opens up a place for dialogue and the possibility that you may be confronted with evidence, which might, over time, allow you to change your perspective and try something new.

Denise Fenzi, What is Possible

This is what this blog is about.

Tuesday, December 6, 2011

On Scaffolding

Please excuse the rambling, incoherent quality of the old posts. I am in the process of cleaning them up so they make sense to people who didn't write them.

Thursday, December 1, 2011

Why Dominance Theory is Useless for Training Dogs

The senior wolf does not send the junior wolf three hundred yards away to bring him a dead duck. If the junior wolf finds a dead duck three hundred yards away from the senior wolf, he eats the fucking duck.

Assuming performing the actions typically prescribed in dominance theory (always eat before your dog, always pass through a doorway before your dog, always be situated higher than your dog) will cause your dog to respect you as the natural leader of the pack, it is completely irrelevant to dog training. The point of training dogs is to get them to do distinctly unnatural things. Wolves do not march in formation.

What I Want

I want a training system that works for 90% of dogs, and has the ability to deal with the outlying 10%.
I want a training system that a novice can apply to their dog and get a safe dog.
I want a training system that an expert can apply to their dog and get an excellent dog.
I want a training system that has a low risk of fallout through well-intentioned misuse.
I want a training system that is structured enough to not leave dogs or people guessing about what comes next.
I want a training system that is flexible enough to be used for any sport, work, or behavior.
I want a training system that is internally consistent.

The Dogs Win

Bad R+ training is ineffective and creates bad habits.
Bad R- training is abuse.

On Owning Knowledge and the Purpose of Mentorship

The role of academia and texts in dog training is to allow new trainers to stand on the shoulders of giants. Dog training must always be based upon the actual training of real live dogs, but saying "you will understand with more experience" is not an acceptable response from a mentor to an apprentice. The foundation of mentorship is to keep someone else from wasting time making the same mistakes you did so that they might surpass you in ability.

Counterpoint: The teacher appears when the student is ready. I am a young, idealistic and modern person with a bias to having access to unlimited knowledge at my whim, without qualification. You cannot own knowledge until you earn it, and being told facts is not the same as internalizing knowledge.

It Just Bugs Me

"If I offered my dog a treat for herding the cattle, she would look at me like I was crazy."

This is because positive reinforcement training is not about treats. It's about REINFORCEMENT. If giving your dog treats is not increasing behavior, treats are not a reinforcer.


If you want your dog to do things, you need to use reinforcement. Positive or negative, behavior comes from reinforcement. Think about how your training paradigm uses reinforcement, and you will instantly understand how to put it into practice better. Success builds on success. Even if you are using corrections, they are meaningless unless the dog can succeed at the task following the correction. The important part about the choke chain is not when it is tight, it is when it is LOOSE.

"Some dogs just need corrections."

What bothers me about this is not so much that dogs are being corrected, it's the implication that if you bolt away from the dog so that he flips over and is dragged five feet, the dog will now instantly respect you and all your training will go swimmingly from that point on for ever and ever until you ride off over the rainbow together on matching unicorns. I am not against dogs getting physical corrections, provided the correction is administered within a system that the dog understands. I am against training that doesn't do anything, that only addresses the behavior in the moment and not the behavior in the future.

Corrections are not ABOUT punishment. They ARE punishment, in the psychologist's use of the term ("don't do that"), but what makes a correction more than just punishment it that puts the dog back in correct action; the sit correction is a sharp jerk upwards because it will cause the dog to sit and thus be correct (this chain is an example of P+ -> R-, since the collar is tight when the dog is not sitting (P+) and loosens when the dog is sitting (R-)). A correction is punishment with information about how to receive reinforcement. It's the difference between a teacher marking the wrong answer in red, and marking the right answer in red.

Do What you Know in your Heart to be Right

There are people I don't like talking about dog training with.

I don't like it when you train dogs sloppily, and since they're Labs and Goldens they tolerate it and work anyway.
I don't like it when "you'll show her," and "he'll never do THAT again."
I don't like it when you set a dog up to fail, and nail her for falling for the trap.
I don't like it when you two do that subtle exclusion silence when I walk up, because I have an internally consistent training philosophy.

The world would be boring if everyone was like me.


Dog Training Law

If you click, you must treat.

This is dog training law because this is the contract you make with the dog. "If you do something I want, I will do something you want." (Shirley Chong) The click/treat is the manifestation of that contract in it simplest, most concrete, most black and white terms. It is the basis of all communication with your dog.


To the dog, not the blog, because I never make any mistakes.
Personally, I feel if you don’t make fair, well-timed corrections part of your training you’re setting yourself up to go into the ring with a dog who will probably do as he pleases once he realizes no tangible rewards or additional handler help are coming. Depending on the alignment of the planets, this may result in a qualifying performance or it may not. (Exercised Finished - Are Corrections Really Necessary?)
What IS a correction?
...• A correction only needs to be strong enough to get your point across; if it doesn’t make an impression, you’re just nagging your dog and that’s not going to fix anything.
• A correction addresses the problem at the point where the error occurred (for example: at the point of pickup on a retrieve or during a slow response to signals)
• It is better to make 1 effective correction than 6 naggy ones.... (Exercised Finished - Corrections Part I)
Punishment techniques should not be taught to novices. It is an advanced technique. (Bob Bailey, The Fundamentals of Animal Training, paraphrased from memory)
In the example of the ubiquitous of the leash pop, the timing, magnitude, and attitude of application are of paramount importance.

It must be strong enough to make an impression (too low a level will require more frequent application and risks habituation - the "punishment callus") but not so strong as to overwhelm and shut down the dog. It must be strong enough for the dog to wish to avoid it in the future, but no so strong as to overwhelm the dog's ability to think through how to do so.

Timing of a leash pop is even more important than timing of a click. Mistimed clicks lead to frustration but generally if you're doling out good enough goodies you can keep the dog with you, mistimed pops lead to a frustrated dog that is more likely to say "Screw you! I quit!" than work through the frustration of handler error to figure out what IS wanted.

Attitude of application is something that I have heard varying reports on. Some say corrections should be impersonal, the dog should think they come from the environment. Other say the dog should know corrections are issued by the handler. Most agree that you shouldn't feel anger towards the dog, that the dog is a "bad dog" and that "you'll show him, he'll never do THAT again!" But there is a well understood connection between actions and emotions. Looking for things to correct puts you in a different mindset than looking for things to reward. It sets you up for a more confrontational attitude with the dog in training.

Applying fair and effective corrections that end unwanted behavior and decrease its frequency in the future is a mechanical skill. Novices have CRAPPY mechanical skills. Philosophically, I do not have a problem with skilled trainers applying fair corrections to their dogs to answer questions about an exercise they have performed correctly hundreds of times before. "Do I have to when there is a fox in a box? Yes, you really actually have to." In my training I prefer to avoid that if I can, but when I am Queen of the Universe I would allow other people to do so. I do have a problem with pet owners training their first dog popping their dog because he didn't auto sit, because they're going to do a piss poor job of it and confuse the heck out of their dog.

When they can effectively handle clicker, treats, leash, dog, and a prop, maybe they are read to start learning about leash pops. But at that point, they probably don't need them.

Safety of the dog is paramount. In immediately stopping, dangerous, bad behavior - not just unwanted or naughty, but outright BAD - I am not above using punishment to stop behavior in the moment. If my dog is trying to eat an entire dark chocolate Easter bunny, you bet your ass I am going to yell and fling him away by the scruff. But I do not consider it especially effective in preventing him from trying it again next Easter.

Corrections are reactive. Positive reinforcement is proactive. Be as proactive as you can, but I do not think it is a dog training sin to have reactions in your bag of tricks.

The Problem of Dog Training: Theory and Practice

This is a response to Sam's post on the dog in group classes that is not ready for that situation.

Ideally, when I am running my own training center, I would tell the person to sign up for some privates and credit  what they paid for the class towards private instruction.
Sometimes I really wonder: does the ability to understand classical conditioning methods require a genuine and whole interest in dog training and behavior, or even learning and behavior as a whole?  I don't say that disparagingly.  The fact is that we live in a society where it's largely accepted as OK to deal with dog misbehavior with a pop, a snap, and perhaps a verbal hiss.  Can people who just want a nice house pet wrap their mind around the idea of not JUST rewarding or punishing behaviors, but shaping emotions and associations?
This is why I want to separate pet and performance classes.

What makes the difference between a Dog Trainer and a Dog Owner is interest in the process vs. interest in the product. You can drive a car without knowing how the engine works. I like driving my car. It fits my needs. I really don't have the time or inclination to study how engines work, what a spark plug is or where it goes. As long as it goes when I step on the gas and stops when I step on the brake, I'm happy. It's similar with dog owners. As long as it doesn't shit in the house, doesn't bite people, and walks on a leash, they're happy. And they can really take or leave the leash walking thing.

As a Dog Trainer, I am highly motivated by the theory behind the method. Training dogs is an intellectual exercise for me, and it's one of the reasons I stick to positive methods. Anyone can train a dog with a choke chain! People have been doing it for decades! I need to make it harder! Yes, there's also the practical benefits and I wholly believe in the validity of the method, but the idea of limiting my tool box to accomplish goals is very, very appealing to me. Constraint forces creativity.

I am teaching my very first class, Clicks & Tricks. I have designed the class all by myself and am teaching alone. I have three students. It is kind of a disaster.

The conflict is one of theory vs. practice.

On one hand, I want to promote my training philosophy. I want the hallmark of my classes to be that you never need to take one again: you should have a solid enough foundation of theory to teach your dog anything. I think my understanding of dog training theory is one of my strengths as a dog trainer, that is something that I bring to the table that no one else I have seen in the area is really doing.

By on the other hand, I am weak in practice. I am the first to point out my lack of experience: I have two dogs, and no titles. They are relatively civilized dogs, but Gatsby got issues and Marsh has no recall. My strength in theory also bites me; I like this quote from Sam:
How can I stitch up that big gap between what I know and what the handler knows in the most effective way possible?
I know more about what I'm talking about (which is why I'm talking about it) than who I'm talking to. I can barely organize my thoughts in a way that makes sense to other human beings, let alone ones that don't have a background in whatever the hell I'm talking about. I can't separate what is actually important knowledge to complete a task because ALL of the information is vital. So I end up infodumping on the student (which, if you've read any other post on this blog, should not surprise you) and watch their eyes glaze over.
Fiesty Fido or Shy Dog classes sound great in theory, and that's because they are.  But they're not offered nearly enough.  Subsequently, those teams who need a little bit of extra help are thrown in with the teacher's pets and valedictorians.. and the result isn't pretty.
Training people should reflect how you train dogs. One of the things we stress in clicker training is "raise one criteria at a time." So you don't go from a ten second sit stay toe to toe with the dog to a three minute sit stay thirty feet from the dog while someone is bouncing a tennis ball behind him. You don't hand a person a clicker, a leash, treats, and a dog and say "you'll figure it out." That is sloppy training.

In many pet dog classes, there is just too much covered. The dogs (and people!) are supposed to learn rough forms of all the AKC Novice Obedience exercises, how to manage their dog at home, basic dog safety, socialization, AND how to read dog body language. In one hour a week for eight weeks. If you're lucky, you get a puppy class and a basic obedience class out of any one dog and if you're REALLY lucky you'll see that person in another eight years when they get their next puppy. You just can't get all of that in, period, let alone to any degree of nuance.

In my opinion it is a mistake to lead pet owners to believe that one class will cover all their needs. Yes, people are always told that training is for the life of the dog, there are more advanced classes, etc, but at least in my club the number of people who follow through on that are very small. The general consensus of trainers seems to be "let's hit on all the topics so if we never see them again at least we said SOMEthing," but I think that is giving owners just enough knowledge to be dangerous. Especially when I consider the information you're giving them to be dangerous, like the idea that you need to be the boss of your dog, he will work for you just because you are the boss, or that noncompliance is disobedience.

I would much rather see smaller, more tightly focused classes that address the core needs of the pet owner.

Puppy Kindergarten: How To Not Kill Your Puppy in the First Six Months, Accidentally or On Purpose
Household Manners: Go to Mat, Recall, Down Stay*
Zen and the Art of Dog Training: Leave it!
Loose Leash Walking
Canine Good Citizen

Can a DOG learn all of these skills more or less simultaneously? Yes. Can a PERSON learn to TEACH all of these skills to their first dog? NO. Most people that have been in my puppy class don't understand that if you are going to be teaching dogs with food treats, HE GETS TO EAT THE TREAT. Looking at treats is not reinforcing for dogs, EATING them is.

*for the average pet dog, I consider the down stay a better option than sit stay.

"Whatever Works"

is unacceptable language in dog training. Famously, shooting the dog works. It is the only 100% reliable solution to dog behavior problems. Instead, whatever you are doing must work. All the perfect application of scientific principles doesn't mean a thing if the dog's behavior is not improving.

If it is not, reevaluate. If the dog's behavior is getting worse, stop what you are doing and try something else.

"He Only Listens When I Have Treats!"

And why this is a nonsense argument for not using treats in training.

The dog that only listens when he sees the treats is the same dog that only listens when the leash is on. When the leash comes off, he knows you can't pop his collar. When the treats go away, he knows you can't pay up. In both cases, the solution is exactly the same: set up the situation to teach the dog that consequences are still in effect even when the leash is off and the treats are invisible.


I have recently been following the discussion on [click-bite], a R+ group for training dogs for protection sports. I am also reading a book called "Drive" by Daniel H. Pink.

So this is the encapsulation of my training philosophy: perfect performance comes from confidence. Confidence comes from competence. Competence comes from clarity. So my goal will be to train my dogs with maximum clarity and minimum stress. I want my dogs to be so competent at their behaviors that they are not worrying about why they can't do them, but instead do them for the intrinsic motivation of doing them.

The Book with the Turtles

I have a fondness for collecting old dog training books full of advice that I have zero intention of taking.

Beyond Basic Dog Training
It's the one with the turtles!

Game Dog
I do wish this guy had more descriptive titles for his books.

Training Your Dog - Volhard
Hey, Volhard! I bet this one isn't totally useless.

Dog Handling & Judging
Oh, these look kind of useful too!

I bought a dSLR this weekend (got a REALLY good deal on craigslist) so hopefully I'll have some more pictures to share with you.

White Unbalanced
It's good to know that I can manage to mangle white balance on a $400 camera.

When I am a Dog Trainer II

When I am a dog trainer, classes that don't differentiate between pet dogs and performance dogs will be called "Non-Regular Classes."

When I am a Dog Trainer

Although I have always loved dogs, and I love people who love dogs, I am but an adoptee of the dog people. My first home, my people, are Engineers, the nerds, the geeks, the gamers, the wedgerats, the techs, the fraternity of Lamda Nu Lamda. And, once upon a time, I spent a year and a half residing at Worcester Polytechnical Institute, the University of Science and Technology. And Life.The motto of this fine institute of many sciences is "Lehr und Kunst." It is upon these words - Theory and Practice - that my school will be founded.

I will teach pet classes, because honestly you have to, but there will be a Pet track and a Performance track and there will be flow charts and it will be awesome. Some classes, like Clicks & Tricks and other things that can stand alone will be trackless, but entry level pet classes won't see much actual clicker, if any.

Novice A
A Performance class for people who have never titled a clicker-trained dog in Obedience. 8 weeks, 6 teams.
Prequisites: Performance Foundations I and Obedience Foundations

Week 1: Bring Dog, crate, clicker, treats. Review Novice exercises in slideshow and assistant + demo dog forms.
10 minute warm up: Rally-style doodling. Do you and your dog remember the things you've already learned?
15 minute lecture: Dogs in crates. What areas do you need to polish? Most of this class is about heeling and proofing, since sit-down-stand are covered in the preqs. Leashes are a safety net! If your dog needs a leash to stick with you, your rate of reinforcement is too low. Demo footwork and other handling signals with dog, then without for right-about turn. Leave slide up showing footwork while handlers pair up and walk 3 paces - about turn - 3 paces with "dog" watching only feet. Switch. Repeat, watching face/shoulders/upper body.
15 minute practicum: Get dogs out, set up for heeling. Handle about turn as practiced. Treat often, and work at your own pace. It is more important that your dog is successful and is keying in on your movements than keeping up with the class.
10 minute lecture: Dogs in crates. Demo right turn as before. Homework: 300 peck heeling with verbal "Name, sit." Practice right and about turns - not in the context of straight line heeling.
10 minute free-time: Leave, or stick around and get questions answered, a little extra practice, etc

Week 2: Warm up before class. Think about your routine before going in the ring to compete.
10 minute homework check: Who can 300 peck the farthest? When you drop out, stop where you lost and practice sit-stays. Winner gets some kind of credit that can be applied to stuff.
15 minute lecture: Dogs in crates. Left turns.
15 minute practicum:
10 minute lecture: Dogs in crates. Demo linking straight line heeling and turns. Be sure to lower distance/duration since you are making it harder. Homework: Can you work up to where your 300 peck was at the start of class with a right, left, and about turn? Don't add them all at once!
10 minute free-time

Week 3: Slow

Week 4: Fast

Week 5: Figure-8

Week 6: Long Sit

Week 7: Long Down

Week 8: Heeling Off Leash/Run-Through

I got to our Obedience class early tonight, and after letting Gatsby run out a bit of his crazy left him in the car to watch the Beginner Obedience class. I had thought about putting Gatsby in this class instead of Pre-Novice, and I'm glad I didn't. For one, it's MUCH larger and fairly chaotic.

I really enjoyed the chance to observe though, because there were a lot of interesting things I saw.

First of all, you can tell who breeds what around here - there are a fair number of Dachshunds (1 in agility, 3 in Beginner, 2 in Pre-Novice) and the Danes! I think the instructor (same for my class and this one) breeds them, from what I picked up in conversation. She uses them to demo a lot, there was a beautiful Harlequin tonight.

Secondly, because the class is so big, there is a lot of confusion for people and dogs. I don't think more assistants would have helped all that much, the room was too full. People didn't really know what they were doing and were trying to keep up with the rest of the class, when everyone's dogs really needed different things. It was just TOO MUCH. Maybe I'm just sensitized to it, but most dogs needed to come way dowwwwn, not more up happy playful yay! But then there were a few people like the lady with a young Dane who was shutting down under the excitement (I suspect). The dog was terribly confused, she'd barely get herself into a sit before starting off again. Each time she sat slower and slower, I suspect she needed to be told she was RIGHT for sitting, but now we're going to move on.

As a side effect of size, I'm beginning to realize that the standard structure of a training class is all wrong. Nobody knows why (or exactly what) they're supposed to do what they do, so they don't know how to change it if it's not going to help their dog. Understanding the why behind something is a huge soapbox of mine, which is why I'm such a theory buff (and why I can know all this and still have only a half-trained dog, I never did have much use for practice). Tonight they sent dogs through tunnels as a confidence booster. For the Dane who started out apprehensive but by the third time though was bounding? Yes, and it was delightful to see. For the timid Miniature Poodle who slipped through at first and by the end was trotting calmly (if unenthusiastically)? Yes! For the collie mix who had her leash pulled through the tunnel and still threw back her head and dug in her feet? Not so much.

There's too much lumping - because there's so much to cover in a class period (and let's face it, it's easier to get by with lumping in correction-based training) - and not enough splitting. And trainers have been saying for as long as I've been around that short, high energy, successful training bursts are more effective than dragging through for an hour. I know my energy goes all over the place during class, and there's still a lot of downtime.

So that's my rough outline of how I'd do a Pre-Novice class. Obviously there's a lot left to fill in, because this is my first time through an obedience class with an eye towards competition (and I don't know what I'd cover in the foundation), but the structure is the most important thing. Talk - do - talk - do. Split, practice, combine.