No Reward Markers

sara

Moderator
Staff member
Ok so since this came up in another thread, I think it'll be our first theory discussion topic! and if you haven't read the sticky "The Purpose of This Section" please do!

NRM's are a word or a sound that's considered by users to be completely neutral, kind of like the "Hot/Cold" game we played as children. However here's the controvercy:

Advocates believe it's giving constant feedback, and that's all it is. It is a neutral word that means that was wrong and try again (supposed to help extinction)

Those opposed to NRM's believe that they are, infact, adversive (however mild) and it is unnecessary and can actually be detrimental to use them in training.

Here's an article from Karen Priors clickertraining website:
http://www.clickertraining.com/node/2848

I do use them with Oliver, as he seems to need constant feedback, but I wonder if it's because he's my crossover dog, and I didn't know enough about it, and we've both become dependent on NRM's

I do not use them with Mouse, as she's way happier throwing out behaviours at me and doesn't need a "cold" button LOL

Ok so the point of this thread, we will discuss the finer points of NRM's, (for and against) and the more articles, the better!
 

MaryK

Honored Member
If Ra Kismet doesn't 'get it', it's usually MY fault for not giving a clear enough cue. What I do is to say 'let's try again' and make sure I give a clear cue!!!!!! If he offers me say, sit pretty, when I wanted something else I reward for the sit pretty telling him that's a good sit pretty and move on to what I originally wanted making VERY sure my cue is correct. Sometimes I take a deep breath, which gives a pause of course, before continuing, as that gives me time to get my cue correct. I can always see by his expression when I've goofed and he's trying to please me. I would never punish him for trying to figure out what on earth I wanted, he's thinking it through, and to me anyway, that shows he's a smart dog and at times the trainer (me) isn't quite as smart as their dog:D:rolleyes:

On the odd occasion when he just plain 'runs off' to play, I call him back, click/treat and then allow him to have a break using the appropriate cue. I know he's been working hard and we all need a break. I do it this way because I want him to learn to focus for a fairly long time, as we are moving into Dancing With Dogs, and he needs three minutes or more of total focus.

We also work on focus using a platform, increasing the time slowly, bit by bit and that's working just fine. I've noticed a mark increase in his focus since starting work on his platform and have not had to use any negative reinforcement at all.

My personal view is that, whilst all dogs are different, I believe that using Positive Reinforcement Training virtually negates the need for any negative reinforcement. It does put more on the trainer, as they have to ensure they consistently use the right cue and same body language. It as if we learned a particular dance routine to a certain piece of music, always played at the same tempo and then suddenly the music was played at a different tempo, we would be thrown completely, at least in the first few bars, and would find it hard to follow our routine correctly.

 

Mutt

Experienced Member
On one hand it seems like it could help a dog out, because you are not letting him 'swim'. the dog gets feedback that what he is doing isn't the behavior you would like to see and would than know that other behaviors are desired and should be offered.
On the other hand you are constantly telling a dog what he is doing wrong, without helping him figuring out how it should be done (like telling a kid that the awnser he came up with is wrong but not helping him what is right and that way leaving the kid in doubt/uncertainty/indistinctness). So that the dog can completely shut down.
So I think it may work for some dogs (like other training methods will work for some dogs and won't for others), as long as the NRM is neutral and not negative for the dog.
It reminds me of the xxxx method (x x x x x meaning the dog is doing what you want him to do and should continue doing it, so you are saing 'warm'), only then the other way around.
For Mazzel I don't think it would work as it would frustrate him that I'm constantly telling him that he is doing it wrong without helping him out. Like MaryK I also treat mazzel for something he offers although it isn't what he offered (though I don't click), so that he knows what he did wasn't wrong only not what I wanted.
For Boef it might be an option, although I think it helps her more to show here what I want and going a step back.
I think the 'buzz game' given as an example in the article is exactly why I wouldn't choose for it for my dogs.
Other than that I think training should be a fun and positive experions for dogs and there shouldn't be a down side expressing itself in negative feedback.
 

Anneke

Honored Member
I don't think it is negative feedback. More like: go on, you're on the right track. Just not quite there.
I am still trying to get Jinx to do Shame. She knows what I want her to do and she brushes her nose wit her paw. BUT I want her to hold her paw there, not move it. So now I don't reward a brush anymore. I say: do it better.
Cooper sometimes needs help when we train. So I have to tell him he is doing good and to keep going. He is traditionally trained and never learned to figure things out all by himself. I always showed him what to do, by luring him. So when I ask him to figure something out himself, he needs a lot of encouragement. If I just stand there and say nothing, he will lay down and shut down. When I tell him: go on! he knows he is on the right track and he will keep going.
If my dogs goes completely the wrong way, I don't use a word for that. I take a little playbreak and then start over.
For instance the boxtrick. I put the box down and wait. The dog looks at the box, but doesn't go towards it. At first I will reward just looking, as I want him to know, he needs to do something with the box. But if he doesn't go towards it, I will stop rewarding for just looking. I will use: keep going! To make him think of another behaviour.
So I am not telling him he's wrong, I just want him to do something else.
So, as soon as he takes a step towards the box, I will reward that.
I guess it is like a half filled glass;) One would say it is half empty(nrm is telling a dog he is wrong), another would say it is half full(nrm is encouragement to keep going and offering a different behaviour):)
 

648117

Honored Member
When Holly is learning something new I don't use a NRM, we just try again. Or do something easy and then try the new thing again.
I guess I sometimes say "nearly got it" or something if she is close, just like you would with a child, but it's not really a NRM because it usually means she is starting to figure out what I want so she does get a treat :) . If she is getting further away from what I want her to do then we go back a step. I don't really think that NRM's are really needed during the learning process.

If we are going over things Holly knows (and I know she knows it very well in a variety of situations) and she does somethine wrong, then I'll use a NRM in the form of me shaking my head (sometimes I whisper "no" at the same time, but often I just shake my head) and we try again, I'd have to film myself to figure out all the NRM cues I give without realising (maybe I sometimes sigh too) but they arn't harsh and I usually only do it once or twice and if Holly still isn't doing what I want (for whatever reason) then we do something else and then come back to it.


I just read that article and it made me think of the interview with Ian Dunbar that was posted in another thread. He said that it can make training too robotic and like a science experiment if the only feedback that the dog gets is a "click/treat".
And it does almost seem like the article linked to on this thread is suggesting that is how training should be (I'm sure that's not the intention and the author might encourage talking to dogs and not just "clicking" but it is not addressed in the article and it does seem to suggest that the click is the only response from the trainer that is needed).
I felt that the article was rather harsh and made assumptions on how most people would use NRM, like the example given in the yellow box. I didn't think that many mostly positive trainers would use NRM in that situation anyway so I don't think the example was that valid (but I don't have any experience with using aversives so maybe some trainers would use NRM or aversives to get the dogs paw out of the bucket? I don't know, I can't really see how aversives could be used in that situation at all).

Holly is the first dog I've ever really trained and when she learns new stuff I mostly use luring (with food, my finger or a target) and only sometimes do free-shaping (which sometimes involves luring with my eyes, I find it hard to not give Holly a hint when she seems lost :sneaky: ) so maybe the NRM is used more in free-shaping where the dog does need to figure out what you want rather than being shown, idk, I don't have much experience with that.

The article also seemed to exagerate how often a NRM would get used, I'm assuming that if you used a NRM you arn't going to just sit there giving the NRM until the dog does what you want, that doesn't make sense to me. I thought it would only be used after the dog has learnt what to do (or mostly learnt) and then does it wrong so then it shouldn't really lead to frustration. But maybe I'm wrong and that's not how NRM are usually used.

I've only experienced positive only trainers so maybe some of the assumptions I make about trainers in general is wrong :).
 

jackienmutts

Honored Member

MaryK

Honored Member
Good articles Jackie. Today I worked with the hoop with Ra Kismet. He jumps through it really well when used for his chair routine but LOL when I just hold it up away from the chair, he tries to clamber through it, run under it around it or even jump up and try balancing his front paws on the hoop. How could I though use a NRM when he's so funny? I have turned that part of 'training' into a game, as I know full well he can jump - I'm not holding the hoop as high as my dining room table and believe me he can jump on that when he feels so inclined. And no he doesn't get into trouble, I have to laugh at him as he looks so chuffed and he NEVER does it while there's food on the table, I just ask him to jump down, nothing more. I find laughter is the best way, he's happy he's made me laugh and is ready, willing and happy to continue training. I don't feel stressed or frustrated either which of course also communicates itself to the dog.

I feel that keeping training happy, filled with fun and laughter, does pay dividends in the long term. And really, when you think about it, it's not the end of the world if it takes your dog just that little bit longer to 'get it'. So long as you both had a lot of fun along the way. It's not the destination which is the end result, it's the fun you had along the way. Training should be a fun journey, not a boot camp, and the end result is a happy, well mannered dog, who loves to work with you, play with you, sleep with you - so long as he has most of the bed - and just enjoy being around you. Isn't that what life should be about?

If I had used any form of NRM with Ra Kismet whilst working to help him over come his fear of other dogs after the dog attack, I know he would'nt be as far along the path of total recovery as he is today. Even if he does swing out a little when a particularly rowdy dog is madly running alongside his fence line, barking like crazy, Ra Kismet will now do so with a tail wag and re-sets himself very quickly, looking hopefully for that treat which is always on hand. Had I used NRM's it would have increased his fear and caused him to zone out on me. What I wanted was for him to know I 'had his back" and he could rely on me to make passing that scary dog something which was fun, not something to fear, or worse, fear my reaction to it all. I even joke with him about the rowdy dogs or agree with him that they're rowdy and I swear he understands, if not my actual words, the tone of my voice and my expression.

Oh and a wee correction in the article LOL it's whether people......... not weather people I don't think the Meteorologists are into dog training blogs. Sorry I know slight derailment but had to point that out.:rolleyes:
 

MaryK

Honored Member
Sara it's late, will read the article you posted tomorrow. Somehow I worked backwards on the posts in this thread:rolleyes:
 

rouen

Experienced Member
Going from my own experience, when working with Dasy if I use NRMs she tends to shut down. If I let he figure it out on her own she does much better.
Like Oliver, Dingo is my crossover dog. He does do better sometimes if I use a NRM when he's hit a wall.
I think dogs who've been on the other side and helped us cross over have their brains wired a bit differently than those who were raised surrounded by +R. Dasy also tends to be sensitive to any sort of correction so this also plays a role in her response to NRMs. If we're shaping something and she offers something that's not part of what we're working on and I respond with a NRM she will look at me, bark, offer a couple things she knows then walk away unless I can get her focused again. With Dingo I have been slowly working on weaning him off of the NRM.
 

MaryK

Honored Member
Read the article Sara. I really cannot comment much as we do not do nose work. However we are moving into DWD and as such no treats etc. are allowed, just subtle cues (which I need to work on myself). I still cannot see that NRM is needed. Jackie, who attended a recent seminar of Michelle Poilit, said that Michelle spoke of many times when the dogs 'changed to choreography' and actually made it better. And this was right in the middle of a competition not just rehearsing at home. That's good enough for me coming from such a champion who has earned more perfect 10's than anyone else.
 

JazzyandVeronica

Honored Member
I haven't read the articles yet...so I'm going to tell you what I think now; then read the articles and see if I change my mind :p

I have always had a dog; and I have trained "the basics" - but Veronica is the first dog I have had where I have been learning about dog training and making an effort to train something more advanced than potty outside and don't maul the guests.

I didn't know enough about clicker training when she was a puppy to do any free shaping with her; and it is still something I struggle with. Free shaping is not something I personally have a natural affinity for; and maybe that's why Veronica doesn't seem to take to it. More likely because she didn't have the experience of learning that way from a young age. When I tried it later with her - her response has universally been frustration. If she could talk she would say something like "Look Lady, clearly you have no idea what you want me to do. So...instead of wasting both of our time...why don't you figure it out...I'm going to scamper off and take a quick nap...and when you've got it worked out...call me".

It may be my earlier failings, my lack of skill or just her personality and learning style preference. I do think all dogs have different styles and preferences and the best training would be geared towards the style and needs of the individual dog. That said, I very often use a "keep going" marker (the "xxxx") with Veronica, as she seems to crave the reassurance that she is on the right track, and that keeps her active and engaged.

I also use a NRM in 2 instances. I try not to use one if we are learning something new; but if she is continually way off and getting frustrated then I will give her feedback. I never really thought of feedback as adversive, and I guess I have naturally assumed that feedback (both positive and negative) is part of the learning process.

I also use a NRM when working on cue discrimination. For example, Veronica knows spin, twirl, heel position, front, back up, etc.; however sometimes she will get mixed up when presented with just the verbal cue. So if for example, she knows twirl and gives me a twirl on cue 80% of the time and this time I ask for a twirl and she gives me a spin I will do some version of "eh-eh, try again". I don't know what else I could do. I couldn't treat the wrong behavior; and with Veronica if I just did nothing and waited for her to do something she'd get really frustrated...so actually with her I think my doing nothing would be adversive...as opposed to giving feedback and then encouragement to try again.

However this weekend I will read the posted articles and see if I change my mind. :D
 

Dioritt

Well-Known Member
The only NRM I use is a polite "wrong" said with a happy voice when Alf gets confused over cues. He's just like JazzyandVeronica describes Veronica.... if he doesn't get any feedback he'll just stand there looking at me like "so I did something and you're just sitting/standing there doing nothing...what's that all about?". When I tell him "wrong" he knows to either try something else or focus more and actually get it right so that he can get the treat he so desperately wants.

I can see that certain forms of NRM would be detrimental in use, but I can't see how a simple "wrong" would be as long as the dog knows that it means "try again, sweetie, and focus a little better this time" rather than "you stupid hound, why don't you just listen to me!".
 

Dogster

Honored Member
If I use NRMs with Shivon, she'll get upset and shut down. I have to use a lot of enthusiasm and the happiest tone of voice possible when I'm training her. Sometimes I say "try again", but I don't think it makes a difference so I usually don't use it.:)
 

DaniG

Well-Known Member
I use NRMs in training, most of the time its when we're training agility, and one of my kids jumps off an obstacle before they hit the contact point, go around a jump, exit the weaves too soon, etc. The one I use is "UH OH! What happened?!" in a happy sing-songy voice, and then we try it again. 9 times out of 10, when we repeat they get it right. So I would have to say that the NRM, in that instance, is highly effective.

I also tend to use it when I'm working with Lin-Zee on new behaviors, because she tends to get way overexcited and start offering me behaviors left and right. If I don't give her a NRM, then she will continue to offer behaviors that she has previously gotten rewarded for, in rapid succession (it really is quite cute). It tends to calm her down and make her think a little bit more about what it is I want. If I don't give her any kind of feedback, I run the risk of her getting very frustrated as well as the potential of making one of her behaviors go extinct all together.

It's also important that the NRM does not take up the majority of the session. It's easy to fall into the "Cold Cold Cold Cold Cold" and never telling them/failing to see when they are "warming up." If you are going to use them, it's important to understand how to balance them so that they are infrequent, and the majority of the session is +R.

I think it all depends on the dog. Too often we try to make all dogs fit into one tiny little training plan, do this, don't do that, this works, that doesn't. But the thing is, is that no one method will work for all dogs. Just like not all dogs prefer food over their toys, a NRM may work fabulous for one dog while it being detrimental to another. It's important to know and understand to know your dog, and develop a training plan/style that fits THEM.
 

JazzyandVeronica

Honored Member
//I also tend to use it when I'm working with Lin-Zee on new behaviors, because she tends to get way overexcited and start offering me behaviors left and right. If I don't give her a NRM, then she will continue to offer behaviors that she has previously gotten rewarded for, in rapid succession (it really is quite cute). It tends to calm her down and make her think a little bit more about what it is I want//

That also describes Veronica. But with her, it isn't so much that she is throwing out all sorts of new behaviors...but she is so ready to "do it" that she starts to act before she knows exactly what "it" is...and if you don't help her calm down and collect herself, she'll just keep going doing whatever, or she'll get frustrated and bark really loud in your face, non-stop.:rolleyes:
 

jackienmutts

Honored Member
Am finally getting caught up. Read all the articles and posts. I can certainly see both sides, it's an interesting topic. I do use them, altho I never use them in a negative or harsh manner. When we're training and I get something besides what I'm looking for, I'll use a very soft sing-songy voice and say something like try again, or noooo (with my voice going waay up at the end), c'mon, keep going ... more a way to encourage them to keep going, keep trying, and not just quit.

I think, as a few have pointed out, that we as trainers and 'pet parents' have a responsibility to our dogs to truly get to know them and figure out what works best for them, figure out how they learn, how we can best help them learn. When working with Makena and shaping something new, she'll work and work, but if she starts hitting a wall or getting frustrated, and I just sit there and say nothing .. she'll quit. If I give her feedback ... work with her ... she'll keep trying. I think again, so much depends on how the NRM is given. In one of the articles (was it the one from KP's site?) the example was given of the game show contestant hearing a buzzer for a wrong answer - the body language changed, the person slumped, and looked ready to give up. Well, what might be different for that person if, instead of a harsh buzzer, they heard a kind sing-songy voice say "oh, well, that wasn't quite right, how about if you try again?" with no harsh buzzer, and no danger of their game being over? Might it be a different feeling and outcome?

I think one must also judge the individual dog's body language. I can only judge my own dog(s). I look at my girl, trying and trying, looking bright and happy, all excited, and when things aren't going quite right, watch her look to me for something, anything, some sign to tell her if she's on the right track. I also can tell quickly when she's getting too frustrated, and I do my best to immediately switch things up and ask for something else so she can be successful. I think we first and foremost have to get to know our dogs, then learn what tools we have in our toolbox, when to use/not use them, and how best to use them. And always remember that when used incorrectly, even the best tools can be bad.

As for using NRMs in Nosework (I'll only speak for Nosework, as it's a bit different from S&R scentwork) - we do lots of searches. Yes, we do some blind, but most are not - meaning, we (handlers) know where the hides are. We know when our dogs are wrong, if they're alerting on a location - and perhaps they're just dead wrong, or maybe they haven't sourced it close enough (we missed getting our NW2 title because my girl was literally 5-6 inches off when she alerted on an odor - but inches count - and that was blind). When our dogs alert incorrectly, we do have to give them feedback, we can't just stand there. They've done what we've asked - they've found what we've asked them to find (or at least they think they have), and they've "alerted". Now what? They don't get rewarded if they're not correct - but they do get feedback, they have to get feedback. We usually shrug our shoulders and put our hands up, in that universal "now what, I don't know" gesture. The dogs catch on, and they go back to work, and resume searching til they get it right. Is it punishment? Nah, I don't think so. It's just not a reward. But - as soon as they really find the odor, we call the "alert", it's verified as correct - then they get their reward: food, toy, praise, or a combo of the above. Again - no negativity, only feedback.
 

JazzyandVeronica

Honored Member
I read all the articles; very interesting.

So at first my initial reaction to the idea that a NRM is positive punishment was that that was (in my opinion and from my perspective) a bit over the top; however I will concede it is dependent upon the perception of the learner and for some very sensitive dogs it could be perceived as +punishment.

However I find myself realizing that to some extent this is going to come down to trainer perception and worldview and that I tend to agree more with the perspective of the second article. Particularly these 2 points:

//I don’t believe that it is possible or healthy to eliminate all stress in any being’s life, human, canine, or otherwise. Just because modern science has shown us the negative impact of too much stress, does not mean that it’s opposite – no stress – is desirable. Stress will occur in the ring and elsewhere, and the only way to learn to deal with it is through experience//

and:
//my chief complaint has been the lack of real guidance on how to go from rewarding every correct response to prolonged work, time after time, in an environment that the dog knows is devoid of primary reinforcers. //

(and I have no idea why I am now typing in huge letters and I can't seem to fix it...sorry)
 

JazzyandVeronica

Honored Member
However (hoping for normal sized letters in new post) I also very much agree with Garret's point in the third article about building a history of reinforcement before introducing NRM's.

And I was going to type more but I was just informed we have to leave NOW for Veronica to get her allergy shot!
 

DaniG

Well-Known Member
"So at first my initial reaction to the idea that a NRM is positive punishment was that that was (in my opinion and from my perspective) a bit over the top; however I will concede it is dependent upon the perception of the learner and for some very sensitive dogs it could be perceived as +punishment."

TECHNICALLY, NRMs are a form of +P. +P does not mean that it scares the dog or causes any kind of emotional or physical damage. The actual definition of +P is something that is added such that the frequency of a given behavior is decreased. So, by adding in the NRM you are decreasing the frequency of the wrong behavior being offered. No-Pull harnesses work by using +P/-R to get a dog to stop pulling, yet they do not cause pain or intimidation. Nowhere in the definition does it say that punishment has to be mean, intimidating, scary, or cause damage (be it emotional or physical) (paraphrasing the words of Dr Ian Dunbar). So just because something is +P doesn't make it harsh, it just means it fits into that happy little definition :).
 
Top