 partial reinforcement schedule
#388577 - 02/03/2014 08:01 PM |
Webboard User
Reg: 07-16-2011
Posts: 17
Loc:
 Offline |
|
As I progress with my marker training and I've established a solid communications foundation, I have a question regarding not rewarding with food everytime.
As I began marker training by learning from the Leerburg video, it was clearly mentioned to reward every time you mark, no exception, even if you make a mistake by marking prematurely - you still reward.
I recently purchased the EXCEL-ERATED Learning book and I am now at a point in the book called partial reinforcement schedule. This portion of the book now suggests to reward or give reinforcement (food) randomly.
My question is, is the longterm goal not to reward everytime once a behaviour has been learned?
Thank you
|
 Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388580 - 02/03/2014 08:45 PM |
Webboard User
Reg: 05-08-2008
Posts: 1473
Loc: Alaska
| Offline |
|
You will still have to "pay" the dog periodically for him to have any drive to work for you. But you should be able to get him to work consistently for 15 minutes or more at a time, long enough to trial the dog...or get him to respond in situations where you are caught without treats or toys. So there is no point in the dog's training where he never needs a reward again, but you're able to get him to do sequences of commands at a time without a reward.
A tired dog is a good dog, a trained dog is a better dog. |
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388584 - 02/03/2014 11:24 PM |
Moderator
Reg: 06-14-2002
Posts: 7417
Loc: St. Louis Mo
| Offline |
|
Think slot machine!
You keep pulling the handle because you've had luck in the past but if you continue to pull that handle without occasional reward your quit.
Same with the value of the reward. Vary that.
You may love chocolate chip cookies but if that's all you get then they loose a bit of value. Get the occasional Peanut brittle crunch and your interest goes up.
dogs can figure out our pattern of behavior easily so random means just that. Not once every five behaviors or once every third behavior but once every 2nd or fifth, or tenth, and yes even the very first behavior "randomly".
Don't put a pattern to your "randomness".
old dogs LOVE to learn new tricks |
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388589 - 02/04/2014 10:22 AM |
Webboard User
Reg: 09-23-2011
Posts: 2692
Loc: Marrero, LA
| Offline |
|
I' for one, speaking for myself, do have certain behaviors that I eventually phase out the reward altogether. I teach with markers, rewarding every rep in the learning phase. I then progress thru varied and random reward schedules, in that order. For key behaviors, my goal is to remove the reward from the behavior.
One example would be "sit". The only time that I would use a reward for that command is if it is part of a complex behavior that I am chaining together, usually for titling.
Another example is the recall. The only time I reward for a recall is during training for a complex behavior. If we are at home and I say "come", I expect her to beat feet with no expectation of a reward.
I think one of the beefs that some trainers have with reward training is that they don't understand that the goal is to phase out the reward. They think that you've created a monster that always requires a reward in order to produce results.
JMO
Sadie |
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388592 - 02/04/2014 11:33 AM |
Webboard User
Reg: 05-08-2008
Posts: 1473
Loc: Alaska
| Offline |
|
True to what Duane said here...some things like sit, down, in formal heel, stay, informal recall...they just have to do.
A tired dog is a good dog, a trained dog is a better dog. |
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388594 - 02/04/2014 12:00 PM |
Webboard User
Reg: 07-16-2011
Posts: 17
Loc:
| Offline |
|
Very good info here, I'm reading some consistencies in the thread. Varied/random rewards should be incorporated in the training.
I can see this as motivation for the dog too.
For example: behavior - reward....behavior - no reward.....behavior - reward.
This could certainly add drive or motivation to want to perform the behavior - no knowing when or if a reward is given. This could potentially add frustration which would necessarily add motivation (in my experience).
Thank you
|
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388596 - 02/04/2014 07:55 PM |
Moderator
Reg: 07-13-2005
Posts: 31571
Loc: North-Central coast of California
| Offline |
|
Hi, Chuck,
"For example: behavior - reward....behavior - no reward.....behavior - reward. "
Yes, that's variable, which comes before random.
Have you seen this? http://leerburg.com/newsletter/10-07-13.htm
Question: "I'm having a lot of fun training my dog with your marker DVD. Will he be as responsive if we're out and about and I don't have a treat with me or he's not hungry? I am having tons of fun training our Vizsla with your Marker training video. He responds very well to it. I am concerned though about the "end-game."
Will he be as responsive if we're out and about, and I don't happen to have a treat with me, and his hunger drive is low?
I am wondering if we're building in an expectation for him, to always have food, which is logistically difficult."
Answer: "I wouldn’t worry about the end game yet. It takes many months to build the engagement and behaviors to the point where you should be thinking about phasing out rewards.
Right now, you should always have food on you and if you don’t, then I would not ask for any behaviors. You definitely want to build the expectation that you DO have a treat in the beginning. Too many folks get in a big hurry to phase those out.
If this tells you something about me, my dogs are 10, 5 and 2 years old and every time I do laundry I find dog treats in the lint trap of the dryer. Haha. Even when my dogs are fully trained, I randomly reward them and it keeps them guessing (like the lottery). I have food in my pockets all the time.
I'd recommend the video The Power of Training Dogs with Food. Depending on your goals, we also have a video for much later that shows how to wean dogs off food, Finishing Work.
I hope this helps.
Cindy Rhodes"
Also scroll down to the Q&A (mainly the "A") here:
http://leerburg.com/newsletter/03-21-13.htm
Also, see this video clip:
http://leerburg.com/flix/player.php/185/Weaning_Dogs_off_of_Food_Rewards_By_Michael_Ellis/
PS
It's really important, IMO, not to rush the move to variable, and then not to rush the move to random.
I personally usually have tiny Zuke's or similar food rewards on me, as Cindy, mentioned in a pocket (and yes, they are often laundered ...  ). It's a "Wow! Food rewards can come at any time and any place, even on a walk or in the car!" perception for the dog much more than a need to have them to ensure compliance.
|
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388597 - 02/04/2014 03:36 PM |
Webboard User
Reg: 09-23-2011
Posts: 2692
Loc: Marrero, LA
| Offline |
|
Yup, I guess I should have mentioned that, even though I phase out the treats in OB, I never "completely train' a dog to the point where there is no rewardable behavior. One thing that keeps my dog interested in training is that we are always learning new parlor tricks. She also expects a "reward" piece of jerky for finishing a meal, and she gets rewards for "being a good girl" (complying with grooming, baths, medicating, potty breaks, etc.).
Sadie |
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388599 - 02/04/2014 07:59 PM |
Moderator
Reg: 07-13-2005
Posts: 31571
Loc: North-Central coast of California
| Offline |
|
" ... even though I phase out the treats in OB, I never "completely train' a dog to the point where there is no rewardable behavior."
Me too.
One time Ed mentioned that once in a while he even rewards a long-trained adult dog for "good potty" (or he uses "good outside," I think). :-)
|
|
Top
|
|
Re: partial reinforcement schedule
[Re: Chuck Scott ]
#388605 - 02/04/2014 11:19 PM |
Moderator
Reg: 06-14-2002
Posts: 7417
Loc: St. Louis Mo
| Offline |
|
I don't think that rewarding the well trained dog is a problem as far as "randomly".
That may be as simple as Connie's example of Ed reward a well trained adult for "good outside".
How important the "reward" is to the dog depends on the dog.
I can train all day with nothing more the a "Good dog" and an ear scratch with the two I have. That doesn't mean I wont still give them a treat or a Kong toss sometime during the day for complying with a command.
It goes back to your training and the dog fully understanding/believing that pulling that "slot machine arm" will merit a reward sooner or later. That's only created by "random" reward!
old dogs LOVE to learn new tricks |
|
Top
|
When purchasing any product from Leerburg Enterprises, Inc. it is understood
that any and all products sold by Leerburg Enterprises, Inc. are sold in Dunn
County Wisconsin, USA. Any and all legal action taken against Leerburg Enterprises,
Inc. concerning the purchase or use of these products must take place in Dunn
County, Wisconsin. If customers do not agree with this policy they should not
purchase Leerburg Ent. Inc. products.
Dog Training is never without risk of injury. Do not use any of the products
sold by Leerburg Enterprises, Inc. without consulting a local professional.
The training methods shown in the Leerburg Ent. Inc. DVD’s are meant
to be used with a local instructor or trainer. Leerburg Enterprises, Inc. cannot
be held responsible for accidents or injuries to humans and/or animals.
Copyright 2010 Leerburg® Enterprises, Inc. All rights reserved. All photos and content on leerburg.com are part of a registered copyright owned by Leerburg Enterprise, Inc.
By accessing any information within Leerburg.com, you agree to abide by the
Leerburg.com Privacy Policy and Terms of Use.
Previous Topic
Index
Next Topic
