We've started clicker training with the dogs and I'm loving it. Zoey, miss over eager, does extremely well because she will do whatever it takes to hear the click and get treated, so she is real easy to shape. Josie, used to me showing her what to do and then performing it for a treat, gets frustrated easily but more than makes up for it when she does get it (She's a mirror reflection of myself - I'm so used to understanding everything instantly, and hate when I don't get things right away. Such as dog training, I can teach kids how to solve algebraic equations all day but I'm still learning how to communicate with my dogs!). The cues Josie knew from our previous training she has down almost perfectly. But I'm worried about them becoming dependent on treats.
If you've used the clicker training method, was it hard to transition your dogs out of expecting treats, or after so much repetition did it stick with them? What can I start doing differently to make sure they don't get treat-dependent?
My boyfriend's the one mostly critical of training with treats, I'll carry treats around all day long if my dog behaves well for them. I know you're supposed to graduate from this, but hey, I wouldn't work if there wasn't a paycheck coming, either.
Recommended Answer:
Change your reinforcement schedule from a fixed 1:1 ratio to a variable schedule.
First give one reinforcement every other time the behavior is offered. You should find your dogs slightly perplexed at first, but quickly figuring out that two offers are required and they'll probably start offering them in rapid succession. Don't stick with a fixed ratio for long. It dulls the game factor and slows training if continued too long.
Dogs are natural gamblers. Play on that. Don't use a pattern for when you reinforce and when you don't. Instead, generate random sequences of 1, 2, and 3 repeats before reinforcement. They never quite know how many times they'll need to "push the lever" before the "slot machine" pays off. But if you start slow enough (with few enough repeats needed at first), they will actually become more enthused than when you give a treat every time. Training becomes a game of chance, and that's addictive.
This is playing on a concept known as an "extinction burst." Consider what happens when your TV suddenly stops working. You push the on/off button several times because that has always worked in the past. You adjust connections or the antenna, then push the buttons again. You try harder and harder, convinced that if you do the right sequence it will work again. This is an extinction burst. You must be careful, however, not to push it too far, too fast. If you do, you risk them getting disappointed enough to dump that TV and find something else to entertain themselves.
Over time you can gradually increase the range of repeats needed to earn reinforcement. Say 1-5 repeats. Then 2-6. Then 3-7. And so on. They probably won't really notice the change since the schedule is random. This process is called "fading the treats." You still want to periodically reinforce a behavior, but can probably do it with genuine praise and affection instead of treats.
Read More...
If you've used the clicker training method, was it hard to transition your dogs out of expecting treats, or after so much repetition did it stick with them? What can I start doing differently to make sure they don't get treat-dependent?
My boyfriend's the one mostly critical of training with treats, I'll carry treats around all day long if my dog behaves well for them. I know you're supposed to graduate from this, but hey, I wouldn't work if there wasn't a paycheck coming, either.
Dog Training Basics - Some Tips to Help You Start Training Your Dogs
Recommended Answer:
Change your reinforcement schedule from a fixed 1:1 ratio to a variable schedule.
First give one reinforcement every other time the behavior is offered. You should find your dogs slightly perplexed at first, but quickly figuring out that two offers are required and they'll probably start offering them in rapid succession. Don't stick with a fixed ratio for long. It dulls the game factor and slows training if continued too long.
Dogs are natural gamblers. Play on that. Don't use a pattern for when you reinforce and when you don't. Instead, generate random sequences of 1, 2, and 3 repeats before reinforcement. They never quite know how many times they'll need to "push the lever" before the "slot machine" pays off. But if you start slow enough (with few enough repeats needed at first), they will actually become more enthused than when you give a treat every time. Training becomes a game of chance, and that's addictive.
This is playing on a concept known as an "extinction burst." Consider what happens when your TV suddenly stops working. You push the on/off button several times because that has always worked in the past. You adjust connections or the antenna, then push the buttons again. You try harder and harder, convinced that if you do the right sequence it will work again. This is an extinction burst. You must be careful, however, not to push it too far, too fast. If you do, you risk them getting disappointed enough to dump that TV and find something else to entertain themselves.
Over time you can gradually increase the range of repeats needed to earn reinforcement. Say 1-5 repeats. Then 2-6. Then 3-7. And so on. They probably won't really notice the change since the schedule is random. This process is called "fading the treats." You still want to periodically reinforce a behavior, but can probably do it with genuine praise and affection instead of treats.
Dog Training Tips
- Sounds like you've got a great start! I do encourage you to join a clicker training list group such asClickersolutions (a yahoo group) and also to check out the clickersolutions.com and clickertraining.com websites- tons of information, and having others who love clicker training is helpful when you have family and friends who are critical.
I've never found dependence on treats to be a problem. Yes, I use treats a lot when I'm first training something, then I start changing around what the reinforcement is. I still reinforce every (or almost every) correct response, but the reinforcement varies from a quiet "good girl", to being let out of a crate or through a door, or putting a leash on, or a quick game of chase me or tug...the list goes on. Treats are handy when first teaching something, but having a wide variety of reinforcers available is one of the keys to good dog training. I see no reason not to continue to pay my dogs- as you said, you keep getting paid!
Some people do use a variable schedule of reinforcement, as the other poster mentioned, but I don't find it necessary, or desirable. Variable schedule reinforcement leads to variable behavior.
