Tag Archives: shaping

Different ways to teach things to dogs


DSCN1091.JPGIf, as I read in a summary of one study, 90% of the dogs in America know the command for sit and that’s about it. Then I’m not sure why we’re having such a hoopla about training methodology. It would seem, if that’s the case, then whatever is being used isn’t working very well as far as expanding communication. I do hope that at least 90% of dogs are house-trained? Unfortunately a lot of the dogs I met in the shelter system weren’t, but then they didn’t seem to know ‘sit’ either.

Below is a listing of the methodologies I’ve used. The first one is what I was initially trained to do and got quite good at … then when I found out how much better the others worked I was incredibly irritated that I had been introduced to and worked so hard to get good at a method I needed, for the most part, to discard. Maybe that’s why there is so much hoopla. Even though I was good at this first method, the dogs I trained tended to want to quit when we got to the upper levels – I assume they didn’t want to work through any more negatives. The other methods have not had that as a barrier. But I have noticed more hijinx, more fun and less ‘stay there and don’t do anything.’

Mostly I use shaping now, although I’m wondering how much of the shaping works because my dogs understand my minor cues in looking at or adjusting my body to influence their choices … so maybe some sort of mimic/mirroring is occurring. I have also rewarded handling choices, like collar grabs, enough to mostly make them not negative, so some of the first kind of training has morphed into a more positive vein.

1. Training Strategy: Physical placement  – Say “SIT” Pull up on leash (probably with training collar on), optional push down on rear. Release pressure when they comply. Praise for compliance. Type: Negative Reinforcement (increasing behavior by removing something bad (neck pressure) or sometimes positive (meaning ‘added’) punishment because timing is poor and pressure doesn’t get released. Side-effects: Some people like the immediate ‘making them do it’, which increases its use as a technique. Some dogs find this form of training quite unpleasant and stressful, and because the dog doesn’t enjoy the training, after while the handler doesn’t enjoy it either. The use of pressure must be faded as a cue – dogs need to learn to do the action without the pressure, but the risk is the dog continues to need a pressure prompt and the trainer ends up getting harsher and harsher in their efforts to make the dog do what they say. This is a very traditional method, so there are many older references to it. This is where most of the electronic training devices fit in – so there is lots of marketing for this strategy. Dogs tend to become phlegmatic with this method or ratchet up to wildness if told no, often anxious about training or new training, wise to the usual tools used and unwilling to try new things.

2. Training Strategy:  Luring: Show dog reward (or otherwise encourage their attention) and use it to get the body position desired, wait for them to, for example  – sit, because looking up the reward is easier while sitting. Reward with treat, or toy for sitting. Do luring only initially (say less than 10 times), then start waiting for the desired response and reward after it occurs. When you can predict the sit 90% of the time, then add the verbal cue, “Sit”. Type: Positive Reinforcement (increasing behavior by providing reward for it) Side-effects: Dog becomes much more attentive and interested when training is offered. Improves relationship, and increases drive to learn. Creates consistent responses. If luring is used after the behavior is established (for example: food presented first, instead of after the dog offers a response), then the dog may learn to wait it out for a better or bigger offer. Thereby reversing the training process and shaping the owner’s response. This becomes the main complaint about this kind of training – lack of fading the lure…which is unfortunately the owner’s misunderstanding about how to use a reward. Another risk is over-use of treats and subsequent obesity.

3. Training strategy: Shaping.  Set up environment for likely response (for example be close to something, recently handle something, have a prior training prompt or value on something) and when dog looks at or goes toward or steps on or sits on or touches the new thing – reward. Progressively rewarding behaviors that are getting closer to what you want is called “shaping.” This is a build-a-behavior from the beginning (or from a foundation of other behaviors already built) process where the dog is offering actions and only getting rewarded for those that match the steps to what is wanted by the trainer. The dog tries stuff and the trainer responds by marking the right or closer to right behaviors. Type: positive reward and negative punishment (which means the dog gets nothing he wants for the wrong actions). Side-effects: Dog becomes very knowledgeable about what the criteria are for completing a task. This system is quite motivating when done well. The biggest issue with this is beginning trainers not knowing the progression of steps and so not rewarding early enough to keep the dog interested in trying to figure out what is wanted. Also dogs can become very interested in offering novel behaviors, which depending on what you want, may be undesirable.

4. Training strategy: Capturing. Observe and capture it. Most of the things we want on verbal or signal control are things the dog does in general life. If we watch we can mark and reward the behaviors we like as the dog does them. Unfortunately this strategy is most often used in reverse of the above, marking the behaviors that are not desired and punishing them. Side-effects: If using rewards and capturing this is great for identifying calm behaviors. It can be clear, but difficult to repeat (especially quick actions that are cute or funny) because observation is the only set up. If used alone as system – to mark and punish unwanted behaviors – this method tends to produce  anxious, hyper-active, unsure, disinterested dogs.

5. Training strategy: Mimic/copy what I do. Person does something and dog copies it for a reward. If dog is watching and realizes how it works then this can be a very fast way to train. This is often seen in ‘give me your paw,’ person puts hand out and dog may do the same, or lie down – person lays down and so does dog and jumping … there are actually quite a few things that dogs will often copy in action. Side-effects: Some things you don’t want copied by the dog.  …I have not used this method as a stand-alone, so I’m not a good judge of what could be accomplished.

What should you do? Well it depends on you and your dog. Back when the first method I listed was typically the only method used in training police dogs, seeing eye dogs, war dogs … many of these dogs washed out. The change to using more positive reward systems significantly increased the success rate of the programs and increased the working-span of the dogs. But I know that people like to ‘make their dogs’ … whatever and in many ways that is easiest, not the most effective or efficient, but easiest for people to understand. So if you need easiest, go for it. I’ll still be around when it doesn’t work for you. DSCN0944

Dog Training Prompts – rewards or reprimands

Very easy to train dogs, learn back chaining quickly too. Photo J Renzoni
Very easy to train dogs, learn back chaining quickly too. Photo J Renzoni

Both a friend and my daughter have had trouble with the use of cues when trying to stop wildness, goofiness, grabbing stuff and general misbehavior. They both are transition trainers having had a past in training traditionally. But some of the stuff we knew as traditional trainers doesn’t work the same way when we’ve turned the training model good-side-up.

So often dog trainers recommend using a known command to stop a dog from doing something that is not wanted. This recommendation was and still is appropriate if the dog was trained with punishment and negative reinforcement, because then the command is tied with a reprimand.

But in positive training most commands (cues) were tied with rewards, so many times the cue becomes a reward in itself. What is rewarded is repeated…meaning the ‘naughty’ behavior gets rewarded by following it with a command like sit or down or come or…

Ah, light bulb moment perhaps?

Know cues are rewards for these two...    Photo by Jeanine Renzoni
Know cues are rewards for these two… Photo by Jeanine Renzoni

Still there’s the behavior we don’t want to occur. Distract (in a low key manner) the dog or Manage the dog so he/she can’t do whatever it is the next time or Fix the problem by training.

How to distract – this depends on what is happening, of course.

But here are some ideas –

  • slow down
  • stop or hold completely still
  • replace with toy or other activity – smoothly and with no excitement
  • turn face away or look away removing attention
  • slowly turn away
  • leave
  • shuffle your feet
  • sigh or yawn
  • move your hand or body
  • relax your shoulders
  • cough
  • use your ‘non-reward marker’ phrase or word
  • let’s go cue
  • light touch to hip or back (like a little tap)
  • bang/noise (if your dog is sound sensitive keep it softer – and hide that you’re the source)
  • collar grab (if you’ve worked on desensitizing collar grabs, or if it’s an emergency)
  • pull dog away (this can trigger an outburst, which is not a desired response) but we’re getting down to the have to remove zone

Then, depending where you are in your training or re-training a low key reward is offered for the dog’s right choice. You have to decide how soon or how much of the wanted behavior has to be given. Initially, just stopping whatever they were doing and beginning the wanted action should be good enough to get the offer of a low-key reward. The hard thing for most people is their need to command the action when it’s best for the dog to make the choice to manage themselves. The other hard thing is holding back the joy when their dog chooses the right thing (I mean when there’s only one step between the naughty and the right thing. Why? because some dogs chain things together and will start doing the naughty, then the right thing to get the super joy). Just get a bit more space in there and you should be fine and can be a happy maniac…

I think the back chaining phenomenon is more likely in these instances because the dog is already doing it … reward, cue, ‘naughty behavior’, cue, reward and that’s why the handler’s are so frustrated.

I saw a recent post about always rewarding your dog when he comes to you no matter the amount of time or detours. Hmm. I aim to reward only average or better. On less than average I aim to be neutral, but certainly not punishing. If I reward less-than-average performance then I will get more less-than-average performance. This is true of all the behaviors we would like to see our dogs do for us. And dogs learn the back chain on recalls too – go out a little too far, get called, whoopee! Go out too far again… Be aware and you’ll spot the shaping strategy they may be using on you. Then you can turn it back around so you’re the shaper and they’re the shape-e.

Good luck training and lucky you if you’ve got a back chaining dog … think of the chains of behaviors you could get. Fun.