REWARD MARKER VS "good dog".......

Discussion in 'Dog Training Forum' started by dr2little, Oct 10, 2008.

  1. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    I'm starting this thread as a carry over from the member introduction forum.

    The thread moved into more of a reward marker discussion, an interesting topic.

    What is a reward marker? Is it the same in theory as a NRM (eh eh, try again) or is it something different?

    A reward marker simply signals to the animal that it will be reinforced when they hear the sound of the marker.

    The difference between a reward marker and "good boy/whale/chicken/dog" ;)
    is that the reinforcer MUST follow or the reward marker is diluted in it's meaning.

    Many trainers absolutely use a marker to say "good job" but do not call them what they are not.....reward markers.

    If the sound is not followed by a reinforcer than it is not correct to call it a reward marker but rather praise or a bridge to nowhere..;)
     
  2. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    I'm glad this has it's own thread now... as intros and 101 were not really very good places for it! :)
     
  3. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    :( I'm rather depressed this thread hasn't taken off... I don't have much to contribute in the way of what works and doesn't... but I was hoping for a good debate and some added knowledge.
     
  4. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    Me too:( Maybe it'll pick up lataaaa!:D
     
  5. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    *keeps fingers crossed*
     
  6. Saje

    Saje Island dweller

    Joined:
    Dec 26, 2004
    Messages:
    23,932
    Likes Received:
    0
    Trophy Points:
    36
    Home Page:
    Oh it will when the right people are online... :D
     
  7. Kayla

    Kayla New Member

    Joined:
    Oct 25, 2006
    Messages:
    1,421
    Likes Received:
    0
    Trophy Points:
    0
    Location:
    Northern Alberta
    OOu good thread! In response to the first question do I think reward markers are the same as no reward markers, yes- BUT they have different end goals. Both in my opinion become classically conditioned: Reward markers, if used consistently and are presented before the delivery of the reinforcer become positively conditioned "Oh yay I just heard mom say blah blah blah Im getting my reward". No reward markers under the same token, if the same word/sound is used consistently and also is consistently presented and followed by nothing might become classically conditoned to predict nothing, for some dogs this might be disapointing and cause them to quit working, for some dogs it might trigger an extinction burst and get them to try double hard. I personally don't train with NRM's so this is just speculation.

    I think anything can be a reward marker, "good dog", "Yes", "Tomato", "Click noise" I think the real question should be what makes an effective reward marker?

    Well something the dog hears every day might not be a very good reward maker because let's face it, by this point that word probably has 100 different associations. So we want our reward marker to be distinct. How do we keep it distinct- easy save it for training. If you want good dog to be your reward marker then it would be better to praise your dog for a job well done with something different perhaps "Oh fido you are so clever, indeed you are, indeed you are".

    For this reason I like to use words that do not fall into my repetroir often so that I do not use it by accident. Duke has two reward markers that have been conditioned thus far- Yes and his clicker. My bearded dragon's reward marker is the flash of flashlight and my Senegal parrot's reward marker is also a click sound from a clicker.

    So I guess to sum up whatever reward marker you go with try and make sure it is distinct, something you dont use every day with your dog, something your dog can percieve ( a verbal reward marker for a deaf dog probably wouldnt get you very far in terms of training) and something you can deliver consistently ( Im not sure how important this is but dogs seem to be good at discriminating so the shorter the word, the easier it may be to keep your tone neutral).

    I liked your comment about the bridge to nowhere Doc, as I do agree reward markers regardless as to what they are for you and your dog should IMO remain a predictor of good things.

    Kayla
     
  8. JoelSilverman

    JoelSilverman New Member

    Joined:
    Oct 9, 2008
    Messages:
    64
    Likes Received:
    0
    Trophy Points:
    0
    Home Page:
    I'm starting this thread as a carry over from the member introduction forum.

    A reward marker simply signals to the animal that it will be reinforced when they hear the sound of the marker. (maybe) Here is where we disagree. I may reward the dog and maybe I won't... By varying this, I create drive... I cannot believe that I am telling you this, and you do not know it.

    The difference between a reward marker and "good boy/whale/chicken/dog"
    is that the reinforcer MUST (Wrong!!) follow or the reward marker is diluted in it's meaning. If you consistently do not reward the animal you are absolutely correct.

    If the sound is not followed by a reinforcer than it is not correct to call it a reward marker but rather praise or a bridge to nowhere. You are right if you are doing it consistently But we are varying the reinforcement. This is nothing more than varied reinforcement schedule. A technique that I, along with every marine mammal trainer and almost all Hollywood animal trainers use.
     
    Last edited: Oct 10, 2008
  9. corgipower

    corgipower Tweleve Enthusiest

    Joined:
    Sep 19, 2007
    Messages:
    8,233
    Likes Received:
    0
    Trophy Points:
    36
    Occupation:
    4 dogs
    Location:
    here
    OK. I was using the clicker to fine tune heel position with Ares. At that point Ares was already trained through AKC utility and competing in novice, so he wasn't a beginner dog. He understood duration and he understood heeling. However, his position wavered a bit and usually lost him a couple points - which is a couple too many for me :p

    So I would click and treat every time he landed himself where I wanted him in heel position. The reason I used a clicker was for the timing. For most people, motor skills are better than verbal skills, which makes the timing of a click more precise, and I had such a small window of time when he was in the right spot that I simply couldn't get a verbal out.

    I started out clicking and treating when he got himself into the correct position. It didn't take long for him to begin leaving heel as soon as he got his treat so that he could be rewarded for returning to the correct place. So I stopped giving him the treat on the first click. He then maintained the position while waiting for his treat. I then would click a second time and give a treat. I gradually increased duration before the second click and very quickly had a dog that heeled almost flawlessly.
     
  10. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    Why use the clicker then. You ARE suggesting that not following with a reward is an actual method to use a clicker. Your CONSISTENTLY INCONSISTENCE IS CONSISTENT ENOUGH TO DILUTE. Call it what it is???? Variable reinforcement schedules were not meant to include REWARD MARKERS. Variable reinforcement is used with the reward marker.....then once the marker is faded variable reinforcement SCHEDULE must occur.


    What you did worked for you and for your dog but it is not recommended as a way of using a clicker. The principles of clicker training were not in place here, you were using the clicker incorrectly as a shaping tool....not that it didn't work but if we're talking about the science, then you made a choice to use the clicker in a different way.
     
  11. Beanie

    Beanie Clicker Cult Coordinator

    Joined:
    May 17, 2006
    Messages:
    14,011
    Likes Received:
    0
    Trophy Points:
    36
    Gender:
    Female
    Location:
    Illinois
    Home Page:
    Question... would the "correct" use of a clicker here have been to require the dog into the position, NOT CLICK, and then click after a moment of the dog staying in one position? And then to increase the duration for the dog to remain in the position before clicking?


    Just trying to understand all of this. =>
     
  12. elegy

    elegy overdogged

    Joined:
    Apr 22, 2006
    Messages:
    7,720
    Likes Received:
    0
    Trophy Points:
    0
    Home Page:
    when i use a clicker, the click effectively marks the end of the behavior. i would simply not have clicked him immediately, since the behavior you were looking for was not "find the position" but rather "find the position and maintain it". then i would have gotten the behavior i wanted and not diluted the clicker.

    i simply don't think it's fair to make the effort to teach the dog that click = reward and then withhold that reward. it just seems... dishonest i guess. varying rewards is important, and varying your schedule is important. there are plenty of other ways to give feedback to your dog that they're on the right track without using your reward marker.
     
  13. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    beanie from what I know of clicker training I would say yes....

    but I'll wait for the doc to explain... I"m interested in that too! :)
     
  14. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    Yes, that's exactly what would be done for a dog that you're CERTAIN knows the behavior. If you're having difficulty, then I would reshape (and even rename) the behavior as if the dog was a 'freshy'.

    For a novice dog, shapping the behavior in small steps (approximations of behavior) will get you there.
     
  15. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    exactly!:)
     
  16. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    that's what I do when they've learned something. They get a "good[command]" and a scratch. I always treat after clicking though... I would feel like I was lying to her if I didn't! :p
     
  17. JoelSilverman

    JoelSilverman New Member

    Joined:
    Oct 9, 2008
    Messages:
    64
    Likes Received:
    0
    Trophy Points:
    0
    Home Page:
    What you wrote here is:
    The click signifies that a reward is coming but not what or how many.

    What you wrote there is exactly what we do, and how we use it 100%. We use it that way with marine mammals and dogs. I am trying to understand where you are coming from, so bear with me. If you are doing something different than this, my question is why would you do anything different?

    I think I can explain better how most dog trainers use it that I know, including myself. The only time we use it with dogs is when we want to bridge a defined time that the animal did something correct. Like with the eye cover.. especially early on in training. Or maybe a dog standing on his hind legs. The sound works so much better than the word "good". We bridge and then reward with food intermittently.

    I hope this explains what we do.
     
  18. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    In previous posts you said that you would click as a reward marker and often NOT follow the click with a reward. This is where you loose me. :confused:
     
  19. xpaeanx

    xpaeanx Active Member

    Joined:
    Jul 10, 2008
    Messages:
    8,387
    Likes Received:
    1
    Trophy Points:
    38
    Location:
    Long Island, NY
    and me... my understanding of your posts was the the click was the reward.... now you're saying the exact opposite... so confused!
     
  20. dr2little

    dr2little Moderator

    Joined:
    Apr 19, 2006
    Messages:
    7,402
    Likes Received:
    0
    Trophy Points:
    0
    Occupation:
    5 dogs
    Home Page:
    This is from the clicker 101 thread - Joel Silvermans states -

    That was a great description of clicker training, but if I could, I would like to make just one correction. Being a former killer whale and dolphin trainer (you can see the pictures on my website), I actually won the I.M.A.T.A. "Behavior of the Year Award" in 1986. I just wanted to mention that I, along with many other marine mammal trainers found that the clicker or whistle did not need to be followed by the reinforcer, as you mentioned.

    I teach people that the clicker or whistle is a form of communication, and nothing else. It only means the animal did the behavior correctly. Will the animal get a reward? Probably, but on a varied schedule, it creates much more drive and excitement from the animal. I teach trainers to stay unpredictable. If you want a great example in dog training, take a look at the agility trainers. They click, but sometimes never reward the animal until the end. That builds drive and attitude.


    This is what I'm referring to..
     

Share This Page