polmaise wrote:DoubleBarrel GunDogs wrote:Nearly everyone reading this has heard about clicker training. Some have successfully experienced good results, and some have decided that it's just an extra step in training. Others think it's hogwash, and it has no place in gun dog training. This post isn't to convince you to start clicker training, but to clear up some of the misconceptions of a very misunderstood training tool. I apologize for the choppiness and repetition. I chose to break this down into a Q&A format, so that it can be easily picked apart for discussion.
What is a clicker?
A clicker is a simple box type device with a button, when pushed it makes a "clicking" sound. There are many shapes and sizes, but they essentially all work the same as a reinforcement.
Like a Rabbit?
What is a reinforcer?
A reinforcer is something that when presented following a behavior increases the likelihood of repeat behavior in the future. To be a reinforcer it needs to be perceived as something positive to the recipient. I you use a primary reinforcer (anything the dog inherently likes such as food, social contact, play or visual stimuli), paired with a secondary reinforcer (such as a click), a positive reinforcement and association will occur. We then use this positive association to reinforce desired behaviors, which can then be cued by a command. Any audible or visual signal can be a cue to the dog.
Like a Rabbit?
Why is a clicker different from other reinforcements?
The clicker is very consistent, and timeable. You could substitute the sound of the clicker with your voice or snapping your fingers, but the consistency and timing could never be compared. Your voice carries baggage whether you realize it or not. Also, the tendency to repeat a cue or talk to the dog, is virtually eliminated with clicker work.
Like a Rabbit?
On what principal does positive reinforcement "clicker" training work?
The is more of a general question, but it's important to understand. Positive reinforcement training works because your dog is far more motivated to please itself than it is to please you. If the dog anticipates a reward for a behavior, it is more motivated to exhibit that desired behavior than if it were forced to do it. That said, there must be a balance to reward for compliance and an understanding of how to avoid pressure. Just keep in mind that you can't build style in a dog with pressure. That's why I feel it's best to build style with the clicker, teaching and shaping the behavior first. When the behavior is in place and the cue is beginning to generate a conditioned response, you know the dog can handle and learn to avoid pressure.
Like a Rabbit?
How do I make the avoidance transition?
Making a transition to avoidance training is quite simple when the dog already fully understands the cue. By avoidance training I'm referring to teaching the dog to turn off, and avoid stimulation with the e-collar. Once you've established the lowest stimulation level the dog will respond to, you can begin to bridge the e-collar and the clicker. I use a training method called pre-cue / cue. This simply means the new thing before the learned thing. If you were to teach this kennel command with the clicker, you would first lure the dog to the kennel and click and treat every baby step of the way. Once the dog is enthusiastically entering the kennel, and gets rewarded every time he goes in then you generalize the command. Turn the kennel 90 degrees, stand further away from it, place it in various locations, etc... This is all done in very small steps that the dog can understand, and over the span of several months. (I don't do any e-collar work before the pup is at least 5 months old, but begin shaping behaviors with the clicker at 8 weeks. I also try to have introductory field work (birds and guns) completed by the age of 5 months. Begin with the dog on a check cord, and the e-collar receiver on the neck. (The dog should already be familiar with both the check cord and wearing the collar on his neck.) Give the dog the cue kennel and each time he goes in, click and treat. Set your pre-determined e-collar level. Stimulate and say kennel, when he goes in click and treat. When he's consistently going in on the stimulation before you say the cue, stimulate without the cue, and click and treat. At this point I give the cue and follow it with the stimulation, never again giving the dog stimulation before the cue for this command. Generalize the command and continue to click and treat when the dog goes in the kennel to keep him upbeat. With enough consistent repetitions you can eliminate the stimulation and the clicker. A conditioned response to the cue will eventually be made. Keep in mind that this doesn't happen right away, but over an extended period of time and with generalization to the command. Teaching the behavior with the clicker before teaching avoidance will result in a shorter learning curb, less pressure and less frequency of stimulation.
LIKE When it chases a Rabbit?
Nate
How do you stop the dog from chasing rabbits with a clicker?..or Clicker programme?
Ok, here's a bash at the question!......
Obviously a clicker in itself is not going to “stop” any behaviour but through a process of marking and rewarding desirable behaviours the desirable behaviours become more likely and the undesirable less likely….
With my lab I have not really done any ‘specific’ training for ignoring live running rabbits or running/flying birds – most of this was achieved as a ‘by product’ of our early and continued training environment.
I have never told her ‘not’ to chase live game – I have given her rewardable things to do in the presence of live game.
Early work on recall, heeling and sitting was carried out within a game scented environment (pheasant and rabbit). For the next stage in our training, I moved to a field of which the top end was about 50 yards from a pheasant pen of 2000 birds, this area also had an abundant population of rabbits. I would start by walking the field myself to clear it of pheasants prior to bringing my pup in, where we would work on positively training heeling, sitting, recall, etc. and a bit later direction work. Although the actual field and ‘training space’ was clear of pheasants and rabbits, they were all around, (the scent of them, the sound of them and the sight of them moving on the steep banks at the side of the small field.)
Later I would only clear a bit of a space in the field, so that the pheasants were a bit closer to ‘our space’ and so on and so forth….so the up-shot is that I am able to sit her up in a pen of poults – without an issue, I can throw a dummy into a group of older pheasants and she will run through putting up the pheasants and retrieve the dummy without an issue.
So basically I treated live game as a distraction against other behaviours I wanted to see and desensitised her to their presence by gradually exposing her to their closer presence and movement.
The reinforcement history that you build up – where the reward for doing other stuff in the presence of rabbits/other live game rabbits becomes 'accumulated' and becomes far more than just a piece of cheese or hotdog or whatever.
This is one way to use a ‘positive’ or clicker style programme to work in terms of “stopping a dog chasing rabbits”
So... I can see what Polmaise is getting at...Using the high value reward, such as the ‘rabbit’ to reinforce the behaviours that you want. Teaching the dog ‘how’ to get the rabbit. i.e. sit, and wait to be sent to retrieve the ‘rabbit’, from the
outset means that the ‘reinforcement history’ that is built up is the correct one. Any attempts to self reward on the ‘rabbit’ in terms of a chase will be less likely to end with the reinforcement of the rabbit (this will have to be set up carefully so that you know that the chase in unlikely to end with a ‘catch’) and so any chasing behaviour should quickly diminish in favour of the behaviours that have proven a reliable way of getting the rabbit. i.e. the ones you want.
Wendy