Dexter loves learning through shaping!
Free shaping is a type of animal training where you teach the behaviours in gradual steps using a marker, like a clicker, and rewards. Shaping can be a great way to teach some difficult behaviours, expand your animal’s capabilities, exercise their brains, and to build your chops as a trainer. There has been a trend pushing towards free shaping as much as possible, and while it is a powerful training tool, it can also be frustrating for the learner if done poorly and sometimes not the most effective training option.
If you want to be successful with building behaviours with shaping you will need to have good timing. You will probably need a more precise marker than a verbal marker word – I suggest a clicker or using your tongue to make a click noise. There are lots of games you can do to practice timing with a clicker. Try bouncing a ball and clicking every time it hits the ground. Or, while watching TV, take a few minutes and click every time the camera angle changes. If you don’t have good timing, you’re not going to be able to click what your target behaviour is and you might end up shaping some pretty bizarre things.
Before beginning shaping sessions you should have a plan of what the probable steps of the behaviour should look like. You want to start with something the dog can and likely will do easily, and build up in logical steps to the finished behaviour. For example, if I wanted to train my dog to bow, my steps might looking something like:
- Dip the head in a standing position.
- Head halfway to the floor.
- Nose close to touching the floor.
- Elbows bent.
- Elbows touching floor, rear in the air – a bow!
Check out this video of my 13 year old Siberian husky doing some shaping. What do you think my shaping steps might have looked like?
A common misconception about free shaping is that there is a lack of information provided to the learner. If you are free shaping well, you will provide the animal with plenty of feedback. Your goal should be about 15 clicks a minute – that is feedback an average of every 4 seconds. With that rate of feedback and reinforcement your dog should be having lots of success, understanding what you are looking for, and working eagerly for you. If you notice your dog is getting frustrated, then you are probably asking too much and need to adjust your criteria. When you are consistently getting the 15 clicks a minute after a few training sessions, then it is time to wait the dog out to offer the next step of your shaping plan. Keep your training sessions short, only a minute or two, and track how many treats you go through so you know what your rate of reinforcement is. I would count out a certain amount of treats before the session and count what was left afterwards to know how many clicks per minute I was at.
Clicks per minute is also know as Rate of Reinforcement (ROR). What is my ROR in this session with Dexter?
Another objection to free shaping is that is causes dogs to be frantic and continually offering behaviours. While this definitely can happen, I don’t feel it is the fault of free shaping itself. I believe that dogs get this way because their trainers do not add cues early enough. It’s commonly accepted in the clicker training world that you do not add a cue until the behaviour is perfect – however, this gives the animal plenty of rehearsals of the behaviour without a cue attached. The more times the dog does the behaviour without being cued and gets reinforced for it, the more likely it will be that the dog will offer that behaviour when it hasn’t been asked (cued) to do so. So, when do you add a cue? As early as possible! When you can predict with relative certainty that the dog will do some form of the behaviour, start attaching a cue. Once you add a cue, do not reinforce un-cued responses afterwards. You can always change your cue once you get the behaviour exactly where you want it, so your final cue is not attached to the imperfections associated with training.
Here is Brew’s second session with this behaviour. As you can see as soon as I’m relatively sure he will come to the platform I start adding a cue.
Provided that your dog has a good understanding of the concept of cues and you are diligent about getting behaviours on stimulus control, this will help avoid frantic offering of behaviour. Your dog will know the difference between when it is time to experiment (shaping) and when it has been asked to do a specific behaviour. Another trick I really like to do that seems to help dogs have a ‘shaping off switch’ is to use ‘game on’ and ‘game off’ signals to tell them when we are going to start shaping and when we are done. I will use “are you ready” to mean we are going to start training and “all done” to tell the dog our session is over.
One of the single most important things you can utilize to speed up your shaping sessions (and training in general) is to utilize placement of reward. You can deliver your reinforcement in a physical location that will jump start your dog to offer the next repetition. This comes down to planning, but also thinking on your feet. Where do you want the dog to be positioned to set up for the next rep? If you want the dog to stay in position, deliver the food directly to them. If, for example, you are trying to train a dog to go around an object, you can click for just moving beside it and toss the food so that the dog has to move even further around it. Instead of having the dog return to you to get the food, you can jump start the behaviour of moving around the object by using your food reward placement to get them there. If you want to set the dog up to repeat an action, say go to a platform, you can toss the food away after you click so that the dog moves off, and has the opportunity to return to the platform.
Can you think of other ways you could utilize placement of reinforcement?
I find many people seem to think that have to be extremely sterile during clicker training and shaping in general. Put some heart into it! If you are engaging, your dog is going to enjoy the process so much more. It should be a game that both of you enjoy. While you should remain quiet before you click, there is no reason why you cannot praise the heck out of your dog after a click , at a big breakthrough, or the end of a session. Relax, have fun!
Working on shaping four feet on the stool.
While shaping can be a really cool way to get some behaviours, it is sometimes not the most efficient or effective way to train a skill. This is why it is not something I will use to train all the time, unless I cannot get the behaviour easily in another manner or I want to challenge myself and my dog. To avoid frustration and make training go smoother, I suggest that you pick a method that will get the behaviour started as quickly as possible. Often, this will not be shaping. Utilizing prompts such as targets or lures, setting up the environment or even just capturing may be a much faster means of training. You can even mix a combination of luring, targeting, shaping etc. – whatever works best to explain to the animal what you are looking for.
A great benefit of shaping if that there are no prompts to fade since the training process is completely based on the dog offering behaviours. It is important if you are going to use a prompt, that you do not use more than necessary. For example, if the animal will just do the behaviour naturally, you don’t to set up the environment. If you can set up the training area to easily get the behaviour, try not to use targets. If you can get the job done with targets, avoid using a lure. The less you prompt, the less you have to fade. Always remember if you are using a prompt you want to fade it as quickly as possible to avoid the animal becoming reliant on it. Get it out of the picture as quickly as you can.
It’s important to understand some dogs will prefer shaping more than others. My 5 year old Belgian shepherd, Dexter, loves to shape and is really fun to work with. I will do shaping with him often just because he is so enjoyable to train this way. In contrast my young Australian shepherd, Brew, finds it fairly a frustrating process. For Brew I limit the amount of shaping I do and try to use other methods to prompt behaviours with him. I do sometimes still practice shaping with him, as it is good practice to get him to think and use his brain in that way. Sometimes shaping really is the best way to accomplish my goals.
If you or your dog doesn’t really like shaping, that’s just fine. There are plenty of other manners of training your dog to do many wonderful tricks and behaviours. Shaping is commonly misunderstood and difficult to do well, but if you have good timing and planning skills with some practice it can be a good addition to your toolkit. There are some things you can shape which would simply be very difficult to train in another manner. It’s also a great way to enhance your skills as a trainer and mentally exercise your dog. Until next time, happy training!