question for clicker/marker trainers - Page 2

But if you are "marking and rewarding" a dog for putting his mouth on the dowel, aren't you releasing the obligation of putting his mouth on it to get his reward? Why is the click allowed to be the release while shaping but not in other contexts? Isn't this then unclear and inconsistent?
Or maybe you are using the clicker when you are at the stage of having not named the behavior? And then when you name the behavior, you don't use the clicker but use a release word?

That was more of an example of not using the clicker for a command or at the end of a behavior.

Say I am using a clicker to build a down-stay.  The dog already knows the down/platz but needs to learn to stay.  Each time I up the ante I will click and reward but the dog cannot get up out of the down until I say "OK".  So, if the dog can do a down stay with me standing 5 feet in front, if I move 6 feet off to the side, I click and then walk back and drop a treat or toss a treat and keep working.  Another example would be heeling.  I don't use a clicker for heeling but if I did, I would expect that eventually I can heel along clicking and treating but the dog is expect to continue heeling until I say "OK" or give another command.

no .. you click when the reqwuirement has been fulfilled .. that is .. when you have gone back ... and then IMO i await a sceond or two .. to teach myu pup not top anticipate the treat immdeialtely upon my return nor to anticipate a release upon my return !! RE: heel .. i am teaching this to my pup .. why the word heel ? where did it originate .. a heel is part of a foot , yes? that is the concpt .. heel is a p;ositon ... i am teaching my pup thge postion ,: beside me , left side toe to toe. in a few weeks it is irrelevent if i move or add movement - heel means exactly what i just described .. still or moving will make no difference .. if the dog is away from : heel still means one thing - the dog must move to get into that potion "heel"

changer ..like a word - having  1 menaing , sit means sit and down means down ....,, the clicker best mean one thing .. BUT as you wish .. i make no judgement.. i write for the many that read the posts .. ok .. do what  you think .. that is what i dezsire : for people to read and THINK for yourself. value my posts as you wish ..

But if you are "marking and rewarding" a dog for putting his mouth on the dowel, aren't you releasing the obligation of putting his mouth on it to get his reward?

Yes and no- that's where differential reinforcement comes into play. It's crucial for shaping a behavior which isn't likely to occur on its own. You click and reward for successive approximations to the desired end behavior and "surf the extinction bursts" to get more out of the dog each time. It really depends on how clicker savvy the dog is. When a behavior that has been marked several times ceases to generate a marker/reinforcer, the dog begins to throw a 'tantrum' of sorts called (as you probably know) the extinction burst in an attempt to generate a mark/reinforcer. On a truly clicker savvy dog, you can shape relativley complex behaviors in a short period of time. The big thing is to keep the rate of reinforcement at about 85%- any lower, and the dog is being pushed too hard and may become discouraged, and any higher, and the dog will be reluctant to up the ante to generate a reward. 

Why is the click allowed to be the release while shaping but not in other contexts? Isn't this then unclear and inconsistent?

IMO, it is. I use the click to end the behavior. 

http://www.clickertraining.com/node/303


Or maybe you are using the clicker when you are at the stage of having not named the behavior? And then when you name the behavior, you don't use the clicker but use a release word?

I can't speak for the poster you're addressing, but I personally condition the click as a release first, then when I feel the behavior is truly under stimulus control, I introduce a release word and fade the clicker. 

I use the click to end behaviors. Obviously the first step in training any behavior is compliance, so in the early stages any compliance with the request generates a mark/reinforcer. When capturing, of course, I mark/reinforce what I'm capturing, then later when it's named I mark/reinforce only when the behavior is asked for. 

I don't worry about time (holding a behavior) until I have compliance. When the dog is complying 85% of the time or so, I will withhold the marker/reinforcer for a few seconds and build up from there. I don't build up progressively, though. I may wait two seconds, then five seconds, then reward instantly for compliance, then wait ten seconds, and so on. I've found if I withhold the mark/reinforcer for progressively longer periods of time I end up with a dog that breaks commands because he's lost interest. If the mark/reinforcer could come at any time, I find I have the dog's rapt attention. 

Only after proofing complaince and time do I introduce the release word and fade the marker. 

I find that I can get a lot of training done in a very limited amount of time using this method. I shoot for five reps of a behavior, five times per day and usually get great compliance the first day I work a behavior, and proofing usually takes a couple weeks at most. I am a huge believer in differential/progressive reinforcement because it gets solid behaviors FAST. 

I do use a clicker for heel, too, but it works slightly differently. With sit, the dog sits, I click-treat-reset-repeat- I click the clicker, throw a treat, the dog gets up and gets the treat, and I ask for a new behavior. The behavior ends because the dog gets up to get the treat. 

With heel, since I'm a group obedience instructor, normally I start off with dogs who pull. I use the click to mark returning their attention to me- so that behavior is complete. However, the dog wants to continue to generate treats, so the dog sticks with me. When pulling happens again, I stop, wait, and repeat. 

I click to mark direction changes, pace changes, and eye contact. I usually toss the treat on the ground in front of me and the dog snatches it on the run, then repeats whatever it just did to get the reinforcer. If I'm building eye contact, I wait longer and longer before delivering the reinforcer so that I get more and more eye contact. By tossing the treat, I make it easier for the dog to figure out what generated the reward by doing something (throwing the treat on the floor) to cause the dog to look away so it can look back again for a reinforcer. 

Heel/leash walking is also slightly different because for most dogs, going forward is a primary reinforcer, so you can build a good heel/leash walk with a continual reinforcer and R- when the dog breaks complaince (standing still, not letting the dog go forward or sniff if they're pulling or straying). Walking properly can easily become self-rewarding when the dog doesn't get to go forward at all when s/he's not walking properly!

did i mis read the post?? i don't get  that ?? you are suggesting everything you do not want to teach the dog !! i want people to understand what the dog is learning ./   all reinforcers should be from you ... and directly from you . secondly  ..  outside from the fact that a clicker is to shape behavior .. you decide when the behavior ends , not the clicker. case in point .. i have my pup sit .. i click and treat .. that does not end the exercise .. it tells the pup " : yes 'sit' , that's  right , good sit.. i will click and treat 10 times - the dog is to stay sitting still . the sit ends when i say so .. stay is implied always with sit , down , stand.. [ there is in reality no stay , either the dog learns to holds it's postion / request or it does not]the dog will not learn to anticipate nor undertake another decision unto itself . .. i am also teaching that the pup it never, never , ever  breaks the position .. not even to get a reward or a treat .. nothing , nothing will ever overide ' sit ' unless it is me by a release cue or by giving the dog another request. :.. sit is an absolute [ or should be taught that way, unless you teach 'sit and get up when you so please']. the best scenario .. sit [ click treat] , sit [click treat], sit [click treat]..the release cue:  " ok all done  , good  dog .. "!!  eventually when the sit is rock solid , you can /may leave the dog , 1 /2 step for starters,  leave the dog and return , click treat and leave agin , return . .. such as sit , click treat , leave , return , click treat , leave , return , click treat .- the 'sit ' does not end when you leave nor with 'click/treat' scenario ', then state " ok all done, good job" ... which brings me to other points ..
            why when a dog is sitting you would encourage it to break the sit by telling it to go pick up the food. , even more so go away from you in thought or action !! why would you throw the food away when you return and /or why would  have the dog think : " not everything comes from you " .. why when a dog is looking at you the food is thrown away and use that as a way to get's it's attention off you  .. no .. shuld be the opposite in life... : ' you looked at me , good , reward, and now dog ,  keep looking at me .that will bring the next reward .. nothing  else. if you want to give food  to the dog , then from your hand , give it ...... even though that treat leaves your hand opun being thrown [ away from youi , throwhe dog is different].. it is still yours.. so you drop a piece of food on the kitchen floor .. who does it belong to ?? it always belongs to YOU ..not the dog !! be consisitent in what we teach our dogs ..

did i mis read the post?? i don't get  that ?? you are suggesting everything you do not want to teach the dog !! i want people to understand what the dog is learning ./   all reinforcers should be from you ... and directly from you . secondly  ..  outside from the fact that a clicker is to shape behavior .. you decide when the behavior ends , not the clicker. case in point .. i have my pup sit .. i click and treat .. that does not end the exercise .. it tells the pup " : yes 'sit' , that's  right , good sit.. i will click and treat 10 times - the dog is to stay sitting still . the sit ends when i say so .. stay is implied always with sit , down , stand.. [ there is in reality no stay , either the dog learns to holds it's postion / request or it does not]the dog will not learn to anticipate nor undertake another decision unto itself . .. i am also teaching that the pup it never, never , ever  breaks the position .. not even to get a reward or a treat .. nothing , nothing will ever overide ' sit ' unless it is me by a release cue or by giving the dog another request. :.. sit is an absolute [ or should be taught that way, unless you teach 'sit and get up when you so please']. the best scenario .. sit [ click treat] , sit [click treat], sit [click treat]..the release cue:  " ok all done  , good  dog .. "!!  eventually when the sit is rock solid , you can /may leave the dog , 1 /2 step for starters,  leave the dog and return , click treat and leave agin , return . .. such as sit , click treat , leave , return , click treat , leave , return , click treat .- the 'sit ' does not end when you leave nor with 'click/treat' scenario ', then state " ok all done, good job" ... which brings me to other points ..
            why when a dog is sitting you would encourage it to break the sit by telling it to go pick up the food. , even more so go away from you in thought or action !! why would you throw the food away when you return and /or why would  have the dog think : " not everything comes from you " .. why when a dog is looking at you the food is thrown away and use that as a way to get's it's attention off you  .. no .. shuld be the opposite in life... : ' you looked at me , good , reward, and now dog ,  keep looking at me .that will bring the next reward .. nothing  else. if you want to give food  to the dog , then from your hand , give it ...... even though that treat leaves your hand opun being thrown [ away from youi , throwhe dog is different].. it is still yours.. so you drop a piece of food on the kitchen floor .. who does it belong to ?? it always belongs to YOU ..not the dog !! be consisitent in what we teach our dogs ..

did i mis read the post?? i don't get  that ?? you are suggesting everything you do not want to teach the dog !! i want people to understand what the dog is learning ./   all reinforcers should be from you ... and directly from you .

A reinforcer is something that causes the likelihood of a behavior to be repeated in the future to permanently increase- when dealing with operant conditioning, to permanently increase in the presence of a stimulus because of a history of that behavior being differentially reinforced in the presence of that stimulus. Ergo, there are reinforcers that come from me (primary reinforcers such as food and conditioned secondary reinforcers such as praise, clicks, and so on) and there are reinforcers that I control (such as the ability to go forward) but the dog decides what is and is not reinforcing. 

secondly  ..  outside from the fact that a clicker is to shape behavior .. you decide when the behavior ends , not the clicker. case in point .. i have my pup sit .. i click and treat .. that does not end the exercise .. it tells the pup " : yes 'sit' , that's  right , good sit.. i will click and treat 10 times - the dog is to stay sitting still . the sit ends when i say so .. stay is implied always with sit , down , stand.. [ there is in reality no stay , either the dog learns to holds it's postion / request or it does not]the dog will not learn to anticipate nor undertake another decision unto itself . .. i am also teaching that the pup it never, never , ever  breaks the position .. not even to get a reward or a treat .. nothing , nothing will ever overide ' sit ' unless it is me by a release cue or by giving the dog another request. :..

I agree that there is no 'stay.' If there is 'stay,' what does sit mean? 

However, I DO use a clicker to end a behavior, which is the way clicker training pioneer Karen Pryor says to use it. You can choose to use it differently, but that does not make my method incorrect. I STILL choose when the behavior ends- again, as stated, with a click in the early phases and with a release command once the behavior has been proofed for compliance and duration. The dog is NOT making the decision on his own. The dog has been taught that the click ends the behavior. 


sit is an absolute [ or should be taught that way, unless you teach 'sit and get up when you so please']. the best scenario .. sit [ click treat] , sit [click treat], sit [click treat]..the release cue:  " ok all done  , good  dog .. "!!  eventually when the sit is rock solid , you can /may leave the dog , 1 /2 step for starters,  leave the dog and return , click treat and leave agin , return . .. such as sit , click treat , leave , return , click treat , leave , return , click treat .- the 'sit ' does not end when you leave nor with 'click/treat' scenario ', then state " ok all done, good job" ... which brings me to other points ..

Sit IS an absolute. I can have my dog sit and leave the dog and go out of sight without the dog breaking sit- all of my dogs. 

            why when a dog is sitting you would encourage it to break the sit by telling it to go pick up the food. , even more so go away from you in thought or action !! why would you throw the food away when you return and /or why would  have the dog think : " not everything comes from you " .. 

It's a pretty standard clicker training practice. Again, it normally goes "click, treat, reset, repeat." You click to mark the behavior, toss the treat so that the dog RESETS (is no

Look, I'm not just making this up. I clickertrain dogs for a living. So, because it's easy to miss a link, here are some excerpts from the link I posted above- again, from Karen Pryor's website (Karen Pryor being the person who popularized clickertraining as a method of training companion animals.) 

It seems straightforward: we click to mark a desired behavior, and then we reinforce. The act of reinforcing the behavior necessitates a change in action: the horse eats the treat, the dog plays with its favorite toy, and the animal in the process of being reinforced no longer performs the behavior for which it was clicked. The click, therefore, ends the behavior. The phrase has become a widely-repeated tenet of clicker training.

How can you clickertrain with a toy if the click doesn't end the behavior? 

Karen Pryor explains: "In shaping new behavior, we often use the click to say 'That's what I want, and I'm going to pay you for that.' It's true that the animal does stop what it's doing to collect its treat. Then, however, it repeats the behavior in order to get another treat. So, yes, the click ends the behavior. But it also promotes repetition of the behavior. Once the behavior has been repeated to the point of being learned, and is on cue and being performed with confidence, we don't really need to mark it any more. We can use words or gestures to either signal for more behavior or to signal a halt, depending on what we want.

Pretty much exactly what I've said. 

Studies at the University of North Texas have demonstrated that to maintain the strength of the association of the click with the reinforcer, you need to keep those two items paired and associated in real time as much as possible, or the power of the click will be reduced until it is meaningless to the animal.

Sometimes, people click and also treat repeatedly during a long-duration behavior, but don't let the behavior stop. In this case the click continues to be paired with the reinforcer, but the click actually has no information in it at all, since it is not marking a clearly identifiable behavior. You have eliminated its main purpose, as a marker signal.

So as you can see, I'm not just pulling this out of a hat here. It's industry standard among clicker trainers. 

Here again is the link to the article: http://www.clickertraining.com/node/303

 And it's not immutable. I've known people who click to continue a behavior and have success. What I'm trying to communicate is that my way is neither wrong nor unheard of and is in fact the most mainsteam method of clicker training. It's also very effective.