Dog training: Phasing out food lures and rewards

Whether you know it or not, what you have been doing here is using food treats in two extremely useful ways: as lures to entice your pup into different body positions and as rewards for the pup to reinforce correct responses when it promptly moves into the desired position. Certainly, food is one of the very best possible lures to entice a dog to perform a variety of responses without coercion, and food is a pretty effective reward for most dogs. A lure/reward training method, especially one employing training treats both as lures and rewards, is without out a doubt the quickest, easiest, most efficient, most effective and most enjoyable way to complete the first two stages of training: 1) to teach the dog the meaning of our instructions and 2) to teach the dog the relevance of our instructions.

Whereas one can never give too many food treats during temperament training exercises, many owners are seduced by the effectiveness of food training when teaching obedience and 1) fail to wean themselves from using food as a lure and 2) give far too many food treats as rewards. Owners quickly become dependent on using food as a lure, since they feel the dog will not comply otherwise. And sure enough, the dog’s compliance quickly becomes contingent on the owner having food in the hand. Similarly, giving too many rewards in training is the quickest way to decrease their value and produce a ‘spoiled dog’.

Food lures and rewards are so valuable in training that it would be unfair to the dog and masochistic for the owner not to use food. However, the number one item on the training agenda is to begin to phase out food as soon as the pup responds correctly, i.e., following the pup’s very first sit! Obviously, no one wants to carry around a smorgasbord of doggie treats for the rest of the pup’s life in order to get it to respond obediently. Food may always remain an occasional ingredient of any training program — as a special reward for the occasional excellent response and always as a lure for teaching any new exercise. However, it is important the dog’s willingness to perform is not contingent on the owner having food or other lures and rewards.

The dog must be convinced that it wants to comply by teaching it the relevance of our requests. Otherwise, the puppy’s initial dramatic learning spree will be followed by an equally dramatic forgetting junket. Just because the pup knows the meaning of our requests does not mean it will necessarily respond. Although food is usually a pretty good reward for most dogs during initial training in non-distracting circumstances, it may not be as effective, if for example the pup would rather play with other dogs. By integrating training into the life of the dog, food rewards may be progressively phased out and substituted with much more valuable and relevant life rewards. Thus, the puppy learns the relevance of our requests and wants to comply.

Substituting food lures with verbal commands and hand signals comprises the first stage of training — teaching the meaning of instructions. Once the pup has learned the meaning of verbal commands and handsignals, it is no longer necessary to use food as a lure to get it to respond, since the word ‘sit’ has become a verbal lure and the lure/hand movement has become a hand signal. Substituting food rewards with more valuable life-rewards comprises the second stage of training — teaching the relevance of following our instructions.

From the outset, alternate sessions using food as a lure only with sessions using food as a reward only. Your dog will quickly learn: 1) just because you have a goodie in your hand does not necessarily mean it will get it and 2) sometimes it will receive a treat when it responds correctly, even if you did not have one in your hand.

Phasing out food lures

With food treats in a shirt pocket and without food in the hand, ask your pup to sit and perform the same hand movements as if you were holding a food-lure. Previously working with food in your hand as a lure has taught your puppy to watch your hand/lure movements, which have now become handsignals. Consequently, your pup will most likely follow the movement of your empty hand and sit. As soon as the pup sits, praise it profusely and quickly pull out a treat to give as a reward. After several trials, the word ‘sit’ becomes a verbal lure, since the puppy now understands the meaning of the word and sits when requested.

Always bear in mind, it is so much easier for a dog to respond to a handsignal than it is to respond to a verbal command. It is both natural and easy for dogs to observe and respond to body language such as ear movements, tail wags, body positions or in this case, human hand movements. If your puppy ever fails to respond to a verbal request, help the dog out, and immediately, give a handsignal. When working with food as a reward only, if necessary use another type of lure to facilitate the handsignal, such as a tennis ball, squeaky toy, chewtoy, food bowl, or anything the puppy values and will follow with its nose, eyes or ears.

Dinner time is another opportunity to practice. Prepare the pup’s supper, and put the bowl on the counter. Give your puppy a hand signal and/or verbal request to sit and/or lie down, and when it does so, praise the pooch, and give it a piece of kibble from the bowl. If your pup does not respond — no big deal, put the bowl back on the counter, and try again later. Your pup will soon get the idea, because we are presenting it with a simple choice: sit for your supper or don’t sit and don’t get supper. Not only does the puppy become accustomed to working when the owner does not have food in the hand, but also it begins to learn the relevance of sitting when requested.

Remember, problems of control — when the dog’s obedient performance becomes contingent on the owner having food in their hand — arise not so much because the dog knows the owner does not have food but because the owner knows there is no food and consequently has no confidence of success and so, does not even try. Just do it! Have the confidence to go cold turkey and work without food lures.

Phasing out food rewards

1) Longer Sequences: Whereas a single sit may be sufficient for a treat during the initial training session, surely it is not sufficient during subsequent sessions? We want the puppy to improve during the course of training, and so with each repetition, ask it to do a little more to receive an equivalent treat. For example, whereas a four-year-old child might justifiably be awarded a gold star for deducing that 2 + 2 = 4, surely a 40-year-old mathematics professor does not deserve a gold star for performing the same calculation. Neither does your dog. Be discerning with praise and rewards.

Once your puppydog sits eagerly and quickly for a treat, the next time ask it to do a little extra — to sit and then lie down before giving a slightly smaller treat. Then, ask it to sit, lie down and sit again for an even smaller treat. As training proceeds gradually and progressively increase the length of the training sequences, i.e., increase the number of responses necessary before giving smaller and smaller rewards. The hypothetical end-point of expecting more for less — the dog will perform an infinite number of commands for nothing.

Perform an extinction test — using food as a lure only, see how many doggy push-ups (alternating sits and downs) your pup will perform before giving up. For example, in their very first training session, Golden Retriever puppies (which I am convinced are whelped in sit-stays with dumbbells in their jaws) average 10 pushups (i.e., 20 responses) for the prospect of a single treat!

2) Stay Delays: In addition to increasing the number of responses, increase the length of time for which your pup has to respond before it receives its reward. Do not be in a hurry to stuff food in the dog’s mouth. The longer you delay giving the food, the more attention you will command. Have your pup sit but delay giving the treat for just two seconds. During initial training, it is helpful to count the seconds as you praise the pup — “Good dog, one. Good dog, two” — before giving the treat. The next time, increase the stay-delay to three seconds before giving a treat, a chewtoy, or a bone and on subsequent trials, to five seconds, eight seconds and so on. If your puppy breaks the stay, there is no need for a reprimand even, since it doesn’t yet understand what we mean. No big deal! Just repeat the request (plus handsignal) to sit, praise the pup as soon as it assumes the correct body position and start the count from zero again. Your puppy may Mickey-Mouse you around as much as it likes, but it will not receive the proffered treat until it stays in the prescribed position for the requisite amount of time. The pup will soon learn that ‘staying put’ is the quickest way to receive its reward. If, however, the pup fails three times in a row, discontinue this phase of the exercise — it is obviously too difficult. It is an important tactic in poker and training not to continue with a bad bet. However, it is vital to always end on a good note and so… sit down, calm down, and then try again, but using an easier body position (e.g. sit) and a shorter stay-delay (e.g. two seconds instead of eight).

Have competitions within the family to see who can hold on to the food lure for the longest time before giving it to the pup as a reward. Persevere. Soon the puppy will be performing bona fide three-minute sit-stays to receive a bona fide bone.

Repeat the above procedure for the down and stand positions. When practicing the sit/down/stand sequences, try to alternate a series of rapidly changing positions with variable length stay-delays in some positions. For example: Sit, Down-Stay (for 15 seconds), Sit, Stand, Down, Stand-Stay (for three seconds) and Sit-Stay (for 10 seconds).

3) Differential Reinforcement: Once the pup performs several responses for a single food reward, say for example, one reward for approximately every 10 responses or one reward for an average of 20 seconds of staying, when is the best time to give the food reward? Obviously, reward the better responses. The power of reward-training depends on only rewarding the pup for better responses and giving the best rewards for the best responses. As Gilbert and Sullivan might have said — Let the Reward fit the Deed. Grade each response, and reward the puppy accordingly. And be a strict judge. Insist on at least an above-average response before even considering rewarding your dog at all. Once your pup realizes only its better responses are reinforced, it will strive to do better.

4) Life Rewards: Once your dog can perform fairly long sequences per food reward, train it for short sessions in which it has to go ‘cold turkey’ on food rewards, and substitute them with other, more valuable rewards. For example, tell the dog to sit, lie-down and sit again, and then lavishly praise and pet the dog, and tell it to “Go Play,” or “Fetch” a ball. If necessary, go back to using food as a lure to entice the dog to comply.

Food is a good reward when initially teaching the dog in the absence of distractions. However, in real life, unless you have a proverbial chow-hound, food loses its effectiveness. The owner wants the dog to sit and receive a food reward, but the dog would rather run and play. In these common life situations, the distractions themselves are the only effective rewards, i.e., life rewards. The deal is simple and one which your dog will quickly come to understand and appreciate: “If you sit, I’ll let you go play. If you don’t, I won’t!” Relevancy training is an amazingly effective, easy and quick way to train.

Excerpted from How to Teach a New Dog Old Tricks, by Ian Dunbar.

Ian Dunbar is a veterinarian and animal behaviorist, founder of the Association of Pet Dog Trainers, and the author and star of numerous books and videos on dog behavior and training. He lives in Berkeley, California with his wife, trainer Kelly Dunbar, and their three dogs. The Dunbars are contributing editors to DogTime.