Free Shaping
Working your dog’s brain, building complex behaviours in small steps
Free Shaping is a tool and training technique that I use regularly with my dogs. In fact, most of my dog’s tricks have been free shaped.
Some behaviours especially complex behaviours (IE Dog putting coin in piggy bank) are easier to train if they are broken down into small bits that are easier for the dog to perform. As the dog perfects one bit, you move on to the next.
Free shaping takes time, and you must be very careful to make it easy enough that your dog can succeed and earn the rewards. If you don’t break things down small enough, you will just create frustration in both your dog and you.
Remember, my number one rule... TRAINING MUST BE FUN FOR BOTH YOU AND YOUR PET!
It is worthwhile practicing this on a friend before trying with your dog. You will be able to perfect your skills a little, and cause less frustration for your dog. Your friend can give you feedback and let you know if your reinforcement rate is fast enough and if you are breaking the behaviour down into small enough bits.
Ok, now for some complex language:
“Shaping by successive approximations involves a process of differentially rewarding some behaviours and not others. When you shape a dog to perform a particular behaviour, you are following a procedure that leads the dog to progress by small steps. You accept as your first criterion for reinforcement some behaviour that is less than the behaviour you want to teach (the terminal behaviour or goal response) but that somehow resembles it.”
Reid 1996
So what does that above paragraph really mean?
Free Shaping is teaching an animal to do something (could be a complex behaviour) without speaking to, luring, prompting etc.
It involves breaking the goal behaviour down into small simple steps (successive approximations) and marking and rewarding for these. In doing so you will be working your way toward your final goal behaviour.
“It's especially good for "crossover" dogs with a long history of correction-based training, since it encourages mental and physical flexibility and gives the dog courage to try something on its own”.
(Pryor, 2014)
So what should you start with?
Step 1:
Pick a goal behaviour – what you want your dog to do. Start with something simple – pushing a ball is a good one.
Step 2:
Break this behaviour down into small sections and reward for these gradually increasing the difficulty as the dog perfects the previous criteria. Remember to aim for the goal behaviour, and don’t get sidetracked if your dog offers something else.
My dogs are clicker trained so I use the click to mark the desired behaviour and then reward with a tasty treat.
A handy hint – reward away from the object – this resets them and they can then offer the behaviour again
Behaviours you might look for when training to push a ball include:
When your dog is reliably able to perform the desired goal behaviour add a verbal cue “Push”.
For more ideas on shaping check out:
Karen Pryor’s game - 101 things to do with a box. You can find more information for this at http://www.clickertraining.com/node/167
Below are the steps I took with Humphrey when I free shaped him to put his 2 front legs in a cardboard box. Remember no speaking, luring, or prompting (pointing etc).
Some behaviours especially complex behaviours (IE Dog putting coin in piggy bank) are easier to train if they are broken down into small bits that are easier for the dog to perform. As the dog perfects one bit, you move on to the next.
Free shaping takes time, and you must be very careful to make it easy enough that your dog can succeed and earn the rewards. If you don’t break things down small enough, you will just create frustration in both your dog and you.
Remember, my number one rule... TRAINING MUST BE FUN FOR BOTH YOU AND YOUR PET!
It is worthwhile practicing this on a friend before trying with your dog. You will be able to perfect your skills a little, and cause less frustration for your dog. Your friend can give you feedback and let you know if your reinforcement rate is fast enough and if you are breaking the behaviour down into small enough bits.
Ok, now for some complex language:
“Shaping by successive approximations involves a process of differentially rewarding some behaviours and not others. When you shape a dog to perform a particular behaviour, you are following a procedure that leads the dog to progress by small steps. You accept as your first criterion for reinforcement some behaviour that is less than the behaviour you want to teach (the terminal behaviour or goal response) but that somehow resembles it.”
Reid 1996
So what does that above paragraph really mean?
Free Shaping is teaching an animal to do something (could be a complex behaviour) without speaking to, luring, prompting etc.
It involves breaking the goal behaviour down into small simple steps (successive approximations) and marking and rewarding for these. In doing so you will be working your way toward your final goal behaviour.
“It's especially good for "crossover" dogs with a long history of correction-based training, since it encourages mental and physical flexibility and gives the dog courage to try something on its own”.
(Pryor, 2014)
So what should you start with?
Step 1:
Pick a goal behaviour – what you want your dog to do. Start with something simple – pushing a ball is a good one.
Step 2:
Break this behaviour down into small sections and reward for these gradually increasing the difficulty as the dog perfects the previous criteria. Remember to aim for the goal behaviour, and don’t get sidetracked if your dog offers something else.
My dogs are clicker trained so I use the click to mark the desired behaviour and then reward with a tasty treat.
A handy hint – reward away from the object – this resets them and they can then offer the behaviour again
Behaviours you might look for when training to push a ball include:
- Walking past or near the ball
- Looking at the ball
- Walking toward the ball
- Dropping the head toward the ball
- Sniffing the ball
- Touching the ball (with nose not paw – be careful not to inadvertently reward the wrong behaviour)
- Pushing the ball
When your dog is reliably able to perform the desired goal behaviour add a verbal cue “Push”.
For more ideas on shaping check out:
Karen Pryor’s game - 101 things to do with a box. You can find more information for this at http://www.clickertraining.com/node/167
Below are the steps I took with Humphrey when I free shaped him to put his 2 front legs in a cardboard box. Remember no speaking, luring, or prompting (pointing etc).
References
1. Pryor, K., 2005. Clicker training for dogs. 1st ed. Waltham, MA: KPCT.
2. Pryor, K. (2014). Karen Pryor Clicker Training | The Leader in Positive Reinforcement Training. [online] Clickertraining.com. Available at: http://clickertraining.com/ [Accessed 31 Dec. 2014].
3. Pryor, K. (2014). 101 Things to Do with a Box | Karen Pryor Clicker Training. [online] Clickertraining.com. Available at: http://www.clickertraining.com/node/167 [Accessed 31 Dec. 2014].
4. Reid, P., 1996. Excel-erated learning. 1st ed. Berkeley, Calif.: James & Kenneth.
1. Pryor, K., 2005. Clicker training for dogs. 1st ed. Waltham, MA: KPCT.
2. Pryor, K. (2014). Karen Pryor Clicker Training | The Leader in Positive Reinforcement Training. [online] Clickertraining.com. Available at: http://clickertraining.com/ [Accessed 31 Dec. 2014].
3. Pryor, K. (2014). 101 Things to Do with a Box | Karen Pryor Clicker Training. [online] Clickertraining.com. Available at: http://www.clickertraining.com/node/167 [Accessed 31 Dec. 2014].
4. Reid, P., 1996. Excel-erated learning. 1st ed. Berkeley, Calif.: James & Kenneth.