I don't think ANY dog will perform reliably if it honestly believes that it has zero chance of being rewarded (and that includes anything THE DOG finds rewarding, including praise, petting, social acceptance, etc in addition to the standard toys or food) and zero chance of being corrected for failure to perform. Dogs are selfish creatures - they do what is best for them. If there is nothing to be gained and nothing to be risked, they won't perform reliably.

Sch3FH2,

I agree with what you just wrote. Both ends of the spectrum are important. I think Ellen's question is which one plays a bigger role in creating reliability.

...and wouldn't that depend entirely upon the dog?

Good point Pete. That's why I said "most of the time".

I'm not sure I understand Ellen's question. Is it "which one plays the biggest part in creating reliability", or, "what *other* factors are there in the absence of either a reward or a correction"? If it's the former, I'd have to say it depends upon the dog and the trainer. If it's the latter, then I'd have to say that attaining an automatic/reflex response through enough repetition would have to occur.

If the question was which method produces more reliable results, positive reinforcement or escape training, then my answer would also be: it depends entirely on the dog. Positive reinforcement only produces outstanding results IF the dog has some goal which it will work very hard for. Ex: many of the sport GSDs that are being bred these days are so ball crazy that they will work til they drop, trying to figure out how to earn that stupid ball, same with border collies, etc. With that type of dog, positive reinforcement is sufficient to produce awesome and reliable work. But when the dog doesn't have any high internal motivation, nothing that he will work hard for, then avoidance training will produce the more reliable picture.

I guess my question wasn't clear. Sorry <img src="/ubbthreads/images/graemlins/smile.gif" alt="" /> . Let me try again.

Is it possible to produce a dog that will reliably perform on command when it knows it will not be rewarded with food, toys, etc. and/or when it knows it can not be physically corrected(compulsed) or punished if it does not perform? AND, if it is not possible (as some indicate), how does one develop a reliable working dog for real-life work, or can't one?

Ellen,
As I understand your question, No I don't think you can. <img src="/ubbthreads/images/graemlins/frown.gif" alt="" />
For what reason should the dog do anything, if other then to please the Handler only. This is what we all strive for, in the real world I haven't seen it. The desire is often there, the results not. <img src="/ubbthreads/images/graemlins/frown.gif" alt="" />

Would our children make all the right decisions if we did not instill values in their upbringing.
Through Positive motivation and compulsive corrections, it's the same with our K-9's.
Nothing hurts more then to have to administer a correction but we do it for their betterment. A firm correction will be remembered where a continious threat not acted upon gets very little results. Rewards for doing well are also remembered. The use of BOTH MUST BE USED in a consistent manner. <img src="/ubbthreads/images/graemlins/smile.gif" alt="" />

Our K-9's are like children, they need training,guidence, praise, correction, and more praise, and continued training. Just my thoughts! <img src="/ubbthreads/images/graemlins/wink.gif" alt="" />

 Quote:

Originally posted by Ellen:
if it is not possible (as some indicate), how does one develop a reliable working dog for real-life work, or can't one?

"Real-life" working dogs are similar to sport dogs in one respect - they shouldn't know the difference between the training and the work. Let's take positive reinforcement first: if you've trained it well, you've reinforced on a variable ratio or variable interval schedule. The dog doesn't know when the reinforcement will come, but he does know that it WILL eventually come, so he keeps working, hoping the next behavior will trigger the reward. That's why the "real-life" narcotics dogs and S&R dogs are rewarded with a find that is planted by the officers if the dog's real searches come up barren. Eventually, the dog's behavior would weaken and even become extinct if he wasn't occasionally reinforced, so they ensure that he does get a reinforcement, even if they have to arrange it for him. so the training and the real work are no different.

Now for avoidance conditioning (or compulsion): the beauty of avoidance conditioning is that the dog learns to offer a behavior BEFORE the aversive stimulus (the prong, shock, etc...) is presented. If he offers the behavior prior to the aversive stimulus, it will NOT be presented; he has successfully avoided it. So he keeps offering the behavior out of fear of the impending aversive. He is working to avoid something that actually is NOT presented, and the behavior is very persistent if the training was done well (the aversive is unpleasant enough to make the dog want to continue to avoid it long after the last repetition of it, it is consistently presented if the behavior is not offered in a timely manner, etc). Since the aversive is not normally presented, how does the dog know when the avoidance paradigm is no longer in effect?? He doesn't. So he keeps working even when the handler does not have the ability to present the aversive (this could be during real-life arrests OR it could be in the middle of the schutzhund trial where corrections are prohibited. No difference. The dog believes the paradigm is in effect, so he works reliably.

IMHO, reliability in performance comes from a consistency in your training method. Consistency is the result self-dicipline and a belief in ONE training method and philosophy. When the correct relationship is established between the handler and the dog, enthusiastic praise should be enough for a dog to lay down his life for you. Corrections should be expected, accepted and understood when required. Self-doubt with regard to whether or not we are implementing the perfect training method only leads to confusion and lack of consistency. Reliabilty and overall performance suffers, then it's time to sharpen the prongs.....only kidding..