Training Techniques Last month when discussing the search square, I broke the exercise into many tiny components so I could build the dog’s confidence and ability in every aspect of the exercise before actually doing the exercise. You may have noticed that whilst I gave some tips, I did not really tell you how I train those components. The reason for this is that I am a clicker trainer. As not all readers will be familiar with clicker training, some introduction is needed to enable everyone to maximise on the information given. Many readers will not want to use a clicker and that of course is fine. As they say, ‘there are many ways to skin a cat!’ another appropriate saying is; ‘if it’s not broken – don’t fix it.’ that said, if you are having a problem with your training, now might be a good time for you to consider clicker training. Many clicker trainers – including myself, initially resorted to clicker training when they had a problem that they could not resolve and were desperate to find a way forward that did not include getting a new dog! I won my first award in obedience on 06/06/1966 (I still have the certificate) and have dabbled with training dogs and competing in both obedience and working trials, on and off, all my life. I enjoy training dogs! To me it is the thrill of communicating with a different species that gives me great pleasure – what a privilege it is to have the trust and the bond that we have developed with these furry four legged creatures. Over the decades I have learnt and tried many different ways of teaching my dog what it is I want him to do. When it comes to teaching the dog to retrieve I have used just about every method you will find in the literature and variations of such as practised in dog clubs around the country. These range from the collar twist, the ear pinch, putting the article in his mouth and stroking, tapping the dog on the muzzle to stop mouthing , firmly scratching the sternum, tying a piece of string to the dumbbell to animate it, and clicker training – you name it, I have done it! It is great to have a big box of tools to help fix any problem; to have a host of methods, ideas tips and tricks to turn to when problems occur, but over the past decade when it comes to training a competition retrieve I would not even consider any method over the clicker. That is not to say that I do not want and encourage the innate retrieve instincts of the dog I most certainly do (see parts 1 & 2 of this series in WTM May and June 2012) but for me, there is no better way to train a competition retrieve than using a clicker. It used to be said that if a retrieve is not force trained it will not stand up to the rigours of competition. In those days when force training was the norm, it was deemed that a play retrieve – no matter how good, would not suffice. Now we have the advantage of positive reinforcement training at our disposal I think we can safely change the old adage to ‘if a retrieve is not formally trained it will not stand up to the rigours of competition.’ We can get better results faster using a clicker than we can by using forced methods of training. One of the biggest problems with a forced retrieve is that any kind of stress or anxiety in a dog usually exacerbates the tendency to mouth the article – so whilst it might appear that you get a retrieve much more quickly using force – it will take much longer (if you ever succeed) to eradicate the unwanted behaviours such as mouthing. The main benefits of using a clicker are; a) the feel good factor it brings to the dog, and b) the handler’s ability to mark the good stuff and end the interaction before performance deteriorates. There is a big difference between a clicker trainer and a dog trainer who uses a clicker. The latter is often a successful traditional trainer who has had a problem with one exercise and resorted to use a clicker to overcome a particular problem. Many dog trainers come into clicker training in this way. I did. I knew all the clicker theory and was teaching other people clicker training but preferred my tried and tested ways for my own dog. Then I found a serious problem had developed with my dog’s jumps and I put too much pressure onto the dog to resolve this issue and then performance rapidly deteriorated across the board. It could not have got any worse. The dog looked beaten as I approached the control field – not because he had been beaten – he had not, but because he knew there were jumps on the control field and that where there were jumps there was usually failure followed by disappointment. Our dogs may not see the sense in achieving a qualification, but they certainly know when we are disappointed with them. The choice was simple, either give up trialling this dog or try a completely different approach. When I realised that when using the clicker I could communicate more clearly with my dog, I started using it more and more. As it turned out the improvement in my dog’s performance was so great that I became hooked on clicker training and my next dog (and every dog since) was clicker trained. When you are fairly successful you have to be desperate before you change your method! However, if you are new to dog training it is easier to begin with the newer methods. If you are new to clicker training it is probably best to begin with a brief overview of ‘clicker training’. Correctly stated, a clicker trainer is a positive reinforcement trainer who uses a clicker to mark required behaviours. I apologise if that seems to be a nonsensical statement with a lot of big words but it is worth spending a few minutes to break that down to understand the whole ethos. Positive reinforcement training is simply rewarding behaviours that you wish your subject (your dog) to repeat. For example if you reward your dog every-time he sits, he will begin to sit more often. Not only will he sit more often he will enjoy sitting as he has a pleasant association with sitting. It is rather like when I used to go to my grandmother’s house as a young child I was given lemonade. I liked going to grandmother’s because I had a pleasant association. Imagine if my grandmother spoke to me in a foreign language and pushed me into the chair before she gave me the lemonade! Perhaps I would feel uncomfortable, slightly unsure of what to do. Maybe I would prefer not to visit grandmother, because she seemed a little odd and I could not understand her. How about if grandmother held a glass of lemonade and pointed at the chair? I would quickly work out that if I sat on the chair she would give me the lemonade. If this happened every-time I went to visit, before long I would rush straight to the chair in anticipation of the lemonade. No force, no pressure, no words. I had worked out how to get the lemonade that I wanted. Once I had learnt this sequence of events, if grandmother wanted to teach me a foreign language, she could say the foreign word before pointing and I would quickly learn that when she said that word, all I had to do was sit in the chair and she would give me the lemonade. Before long I would recognise this new word and have a good association with it: grandmother would not need to point to the chair any more. The word meant I had the opportunity to have some lemonade – all I had to do was sit in the chair. After several more repetitions of this sequence of events the lemonade would not need to be visible up front, I would ‘get’ the word and respond by sitting in the chair to wait for the lemonade to appear. In the above anecdote, the child learnt the foreign word in a relaxed and pleasant way without any pressure – only rewards. The child would have enjoyed learning and would probably want to spend more time with the grandmother and learn some more of her language. That is how positive reinforcement training works – good behaviour is rewarded and therefore a good association is formed. When training a dog, as opposed to a small child, we use the sound of a clicker to mark the exact behaviour that we want to encourage. It is easier to pinpoint an exact moment or movement with a clicker than it is to feed at that precise moment (especially if that moment happens to be holding a dumbbell or whilst in the air at the top of a jump). It enables you to convey, ‘Yes that’s what I want!’ as the action occurs (not at the end of the exercise) in a very precise and calculated way. There is nothing magical about the sound of a clicker! Many dogs naturally respond to the novel sound of a clicker when they first hear it, but that is not sufficient for training purposes. I think the sound of the click to a dog must be rather like us being given a £5 note: it has no intrinsic value; it does not taste good to eat, it is not warm and cosy like a scarf; nor fun to look at like video. In order to make the clicker into a useful tool we first have to ‘charge the clicker’. That means you need to pair the sound of a clicker with food. To do this, stand with your ready prepared pot of titbits and your clicker press the clicker and feed a titbit. Pause while dog eats and then go again. This is ‘Pavlovian conditioning’ also known as ‘classical conditioning’. No behaviour is required. The dog does not have to do anything. An association is made. After numerous repetitions the sound of the clicker becomes synonymous with the food. The sound of the click has become a reliable predictor of food being given. It should be the only predictor: i.e. do not reach towards the pot of food before, or whilst clicking, or the dog will only learn that the hand going to the food is the reliable predictor! To avoid this either have the next piece of food ready in your hand and then pause before you click – then feed; or click - and then get the food from the pot or pocket and feed. Before long your dog should be listening for the click, not mugging you! When your dog understands that the sound means food, your clicker is ‘charged’ and ready for use. What the Click means to your dog; It marks the correct behaviour - whilst it is happening, not after the dog has done it (which is commonplace in traditional training). It means you have won a reward you will get something pleasant. It ends the behaviour – the dog may now stop what it is doing and claim its reward. It is unconditional – no matter what happens, if you have clicked you will reward – even if you clicked at the wrong time. The clicker never lies. When clicker training we never use the word (command or cue) until we know we can get the behaviour. So for example if I hold a treat above the dog’s nose he will raise his head to reach the treat and in his stretch to reach it, his bottom would go down towards the floor. I would click to mark the desired behaviour -the almost sit - and release the treat. I would repeat this action many times until I knew that if I raised my hand holding a treat above my dog’s head he would sit. Or should I say; my dog knew that if I held a treat above his head that I would click and give him the treat as soon as he sat. Now I know I can get the behaviour (by holding a treat above his head) I can precede this action with my word (command or cue). Before long my dog will learn that when I say, “Sit” that I follow through by holding the treat above his head and if he sits he gets the treat. Soon he will be sitting as soon as he hears that word – his bottom will hit the floor even before I can get the treat above his head. With lots of repetitions in lots of locations he will soon ‘get’ the word. He will know that “sit” means sit and if he does he will get a click followed by a treat. From now on if he ‘forgets’ what “sit” means and I need to go back to helping him with the hand signal I will not click and treat him. He only gets the click and reward now when he responds to just the word. The above information is enough for me to be able to start clicker training my dog. The process of clicker training is the same no matter what exercise you are about to teach; First ensure your dog is ‘charged’ on the clicker. Find a way to lure or encourage the behaviour, or a tiny part of the desired behaviour. (Pure clicker trainers wait for behaviours to occur naturally and then click – but for me it would take too much time to wait for a dog to offer all the things a trials dog needs to learn!) Click as soon as the dog begins to do whatever it is you are trying to get him to do. Stop to reward the dog. Repeat steps 2 – 4 as often as it takes until you feel that you can reliably get the behaviour with your lure. Say the word before commencing the lure, click when he does it. Repeat stage 6 until the dog begins to recognise the word and anticipate the lure. Only click and reward when your dog does not need extra ‘help’ to remember what the word means. When your dog responds confidently to the word 80% of the time (without needing more help), stop clicking every single time your dog performs this task. The behaviour is now on what is known as a ‘variable schedule of reinforcement’. When learning anything new your dog will learn faster if you click and reward him every single time he performs the new task, so in steps 2 – 8 you should click every correct attempt. In step nine, once he has learnt the word that triggers the behaviour, stop clicking every time he gets it right. You cannot reward every tiny part of the behaviour in competition, or in real life (otherwise you would be constantly feeding the dog). If you suddenly stop clicking and rewarding that behaviour (because you think he knows that now) he will stop doing it. This is why many people think that using food to train a dog does not work. This is not the dog being naughty or stubborn – this is how learning theory works: it is the same across the board not matter what animal you are training or what task you are teaching, if you go from rewarding every time to not rewarding, the animal will stop doing whatever the task is. This is called ‘extinction.’ The process to avoid is changing from a constant schedule of reinforcement (clicking and rewarding every time) to extinction (not performing the required behaviour). The transition from clicking and rewarding every time to not clicking should be a gradual process. This is called a variable schedule of reinforcement – which basically means rewarding on a variable or random basis. So you may progress from rewarding the dog every time he sits to only clicking and treating two out of three correct responses and then click and treat every correct response for three repetitions and then allow two to go unrewarded and click the third. When moving on like this I often use gentle verbal praise to inform my dog he is correct- to help keep him in the game until the click and treat arrives after a another repetition. Gradually you can reduce the clicking and treating, getting the dog to perform longer or to do more things to earn a click. The secret of success is to make this transition so smoothly that the dog never doubts that the click is coming; that he is nearly there. He must trust that the click will arrive in a moment as long as he keeps working. Remember the clicker never lies – you must feed every single time you click. This way you will not lose attitude and enthusiasm and you will have a dog that knows his job and enjoys working with you. Anne Bussey