czwartek, 23 lutego 2017

Try harder

It's funny how the internet works sometimes. There is lots and lots of valuable resources there, lots of crap obviously as well and sometimes for some reason one or the other is dug up and commented upon and shared and then you get this avalanche of responses and shares. A couple of months ago someone posted a short video shot from a drone about dog weaving, accompanied by a couple of wise terms like "sidewise flexion" and suddenly everybody was sharing the video and giving those words of caution: "don't start too early with slalom, don't train it more than... (totally arbitrary number of repetitions) in a session" etc. And I was like, geez, seriously, you needed a video from a drone to know that the dog is bending when doing slalom? Or you really didn't know that ANY repetitive kind of movement if you do it over and over again is not healthy? Or why you say no more than let's say 4 repetitions of the slalom and not 3 or 5? 
Sure, drilling is stupid and unnecessary, think before you train, plan your sessions, but always, always, be critical to what you read or watch. 

Today another articles gets shared on Facebook a lot - here's the link - and it's actually not a bad one.
I do agree with the main thesis of this post, that is that confusion as such is an aversive as well and that if the dog learns to expect it in certain situation, this situation would become conditioned negative stimulus. And yes, more often than not when you see a dog that "shuts down" or looks not motivated, it is the case of the handler not being able to teach the dog properly how to perform the tasks ahead, not splitting it into easy enough parts for the dog to understand it, keeping the reinforcement level too low or being just totally unclear and confusing so the dog is not able to  know what is required of them. 
That being said I'm still a little puzzled by this paragraph in particular:

"There are a lot of falsehoods in dog training, and agility holds its own special mythology. That our dogs must be resilient to failure, that they must work through frustration, that they should experience disappointment and confusion in training in order to withstand these things in life, are just some such fables."

Think of a shaping session. Every time you withhold the reward, the dog experiences some level of frustration. And this frustration causes the dog to try something else, which might bring the reward. This is how I understand teaching the dog to cope with frustration: it's not a "mistake", it's not the end of the world, it just means "try something else". I do think it's essential part of the learning process. Sport apart, teaching a dog to be a well-behaved companion is also about "this brings you the reward, this doesn't, so it's simply not worth doing". 
There are two extremes on this continuum: one would be to present a dog with a task that is too complicated, too complex and too difficult for the dog to grasp right away - and then we get a confused dog that would "shut down". The other extreme is that you feel you need to reward EVERYTHING and every repetition has to be "succesful" for the fear that the dog would "shut down". Now, dogs are really really smart creatures and the latter attitude sometimes leads to a situation where it is the dog training the handler  😏. You see this when someone tries to raise the difficulty of the task a bit, for instance asks the dog to perform slightly more difficult weave poles entry and the dog doesn't do it correctly the first time. Some people will immediately make the task easier again, thus actually rewarding the dog and eventually teaching the dog NOT to try harder. Another example: you're using a particular toy, which the dog is enjoying for the first or second repetition. And then the dog decides the toy is no longer that attractive and again some people would immediately REWARD that by switching to another toy or using food etc. In both scenarios the handler's response could be different - they could ask the dog to try again and if the dog is familiar with the notion of "try something else if you didn't get the reward" there is a high chance they would nail the weaves entry the second time, thus earning the reward. Or the handler might insist that the dog actually plays with the toy the handler offered and THEN reward with something the dog likes even more - thus using the Premack principle and consequently rising the value of the first reward. And so on, and so on.
There is a really fine line here - we should avoid our dogs being confused. We should split the tasks into small enough bits that it's VERY likely for the dog to succeed MOST of the time. But I feel it's equally important to teach the dog that not getting the reward each and every time is okay too. That it's safe to try different things and "make mistakes", as the only thing that would happen is that you don't get the reward this time, but might very well get it the next time. And I think every dog is able to learn this, if you introduce this gradually. When I start shaping with my puppies I would help them a lot in the begining. Things like how you present the dog with the object you want them to interact with can make a huge difference. For instance, if you just place the object somewhere in the room and then let the dog in, it will take them a long time to actually find it and this may be frustrating. But if you put the object in front of your dog, most dogs would naturally investigate what you just brought and click! you have the first thing you can reward. Similarly, where and how you reward can really help and speed up the learning process. And yet still, I would give my puppies the opportunity to do something else and NOT get rewarded from time to time and I would avoid luring in general for the same reason. 

niedziela, 19 lutego 2017

Adult stuff :).

It is an old cliche that time flies, but somehow Mojo is one year old already!

I definitely couldn't wish for a better dog and she is lots of fun each and every day :). Mostly she is also really easy dog to live with. I'm really pleased with all the foundation stuff I did with her, so now that we're starting a little bit more technical training it goes really easy.
I started some weaves training and it was literally three sessions on two and four poles and she was ready to progress to a whole channel, since she was doing so well with understanding various entries even with my movement and one or two obstacles before. Same goes for see-saw training - we're still mostly playing bang games plus some 2o2o position outside the obstacle, but she is such a brave little girl that I have the impression that she is going to master it really quickly. She thinks bang game is hillarious :). 

RC are work in progress and after the winter break (we actually had snow for over a month, which is not so common in this city) the first couple of trainings were not so awesome, but now it looks like we're on the right track again. 

Now actually running sequences is a bit more tricky, but than again, she is so much fun that with all the bloopers you just have to love her <3. 

Yesterday we went for training with Polona Bonac. I really like Polona's training philosophy and great ideas for different games for different dogs, so she is definetely one of my favourite trainers to work with and the one that I really like to observe when working with various dogs. I worked Mojo and Brego in handling and Zelda participated in the puppy training. Brego was awesome, Polona suggested an exercise to help with his collection problem on S-turns, Mojo was totally funny and a bit of a tunnel junkie (I guess it's genetic with spaniels, LOL), but all in all I think she did great, considering that it was her first ever agility seminar and actually first time she ran this kind of sequences. With Zelda we mainly wanted to provide her with positive socialisation opportunity, since obviously she doesn't do any obstacles yet and she did really well as well. No problems with playing in new environment and even less with focusing on a sausage :D. We did little shaping session on puppy sea-saw and a bit of cavaletti work.