Improve my training! Episode 1

[snowbunny]

Welcome to the first episode of a new recurring thread. I thought it would be good, alongside the observation skills threads over on the Behaviour Geeks section, if we posted videos every now and again that would help both ourselves and each other assess our training skills. Those who know better, do better, and all that!

So I shall start with a video I took of Shadow with Alex-who-became-Dante-who-became-Alex-again.

Right, a little background to this video.
Firstly, it was only ever intended for my own use and maybe to send a clip to @Aitch to show how ambivalent Alex was around Shadow. So, production quality is lousy, it was filmed on a chest-mounted GoPro, so you can't hear most of what I'm saying to Shadow, only when I talk to J. You also can't see one or both dogs for a lot of it. Still, I think there's some good stuff and I learned a lot by watching it back.
Most of you will know that Shadow is lead reactive to all dogs and reactive off lead as well to other male dogs. This normally demonstrates by him running up to an unfamiliar dog (if given the chance), standing really rigid and on his toes before having a go, unless the other dog is ambivalent and moves away, in which case, he tends to stand there looking a bit of a dick until he relaxes
My ideal would be to get him to a point where he can meet another dog and immediately disengage himself, to teach him it's OK to walk away. We're a long way from that, but it's a little fantasy I have!

The video is 6 minutes from about a half hour session, with lots of breaks to get water, shade and space.

Here goes:

I'm going to give my notes first, feel free to comment, to disagree, to highlight anything else you notice.
0:00 We're in our local "dog park", where you are guaranteed to never see another dog. It's neutral territory. I start with Shadow at one end; I have been playing pattern games with him for ten minutes or so to ensure he's nice and mellow before Alex comes in at the other end. I use a cued LAT to start with; Shadow is happy and relaxed.
0:19 We start to move closer to Alex. I continue clicking Shadow for relaxed looking at him, or for looking at me.
0:29 Shadow starts to pay a bit more attention to Alex. I should have stopped at this point rather that continuing forwards and worked on counter-conditioning. This is the threshold of his "pre-worry" zone.
0:34 Shadow doesn't respond immediately to the click. He is definitely starting to worry. I went too far.
0:44 I eventually notice and move him away.
0:56 Change of exercise - parallel walking. Alex is on our right and walking on J's right side, so there are two humans and plenty of distance between the two dogs. Shadow remains relaxed and engaged in the game.
1:16 Change of exercise - 1,2,3 pattern game around Alex, who remains in one place. She finds it easier to be fed on the move rather than stopping after every third pace, so we use that variation. Movement helps Shadow remain calm; it is his friend.
1:28 Shadow move to the end of his lead, interested in Alex. He adopts uncertain but soft body language, and I let him do it. He chooses to move away and re-focus on me; he gets a few pieces of food for this and we continue to move away, using functional reinforcement - turning to me get him distance from the thing he's unsure of.
1:36 Same game, different location (other end of the "park").
1:46 Shadow pauses to look at Alex, but is relaxed and continues the game. We're right on the threshold of his "pre-worry" here, the perfect place to be giving him good feelings.
1:52 Alex starts to move towards us and look at us. We switch to LAT and Shadow plays the game, although he is starting to cross the line to where it's less than ideal learning conditions. He struggles to look at Alex a couple of times; this can be a sign that the dog is totally relaxed about the trigger so it's not even worth looking at, but in Shadow's case, I believe he's not looking because he doesn't want to - he finds it too stressful. When he plays the game, it's "look at the nasty thing and get a reward for it" which is far less effective than "there's a thing which is at a great enough distance that it's not nasty, and I'm getting this good stuff!". The former is coercion, the latter is counter-conditioning.
2:14 We start to walk behind Alex, almost parallel walking but where no eye contact is possible. Shadow easily able to remain in the game without worry, even at a relatively close distance.
2:35 Doing LAT. In hindsight, Shadow is making it abundantly clear that he doesn't want to look at Alex. At the time, I took that to be not being concerned, but, as above, I now believe it to be because he's worried about looking. I should have increased the distance.
3:26 "Fly bys". Playing the 1,2,3 pattern game while walking with both dogs on the inside. Alex is quite interested here, but Shadow does well to stay int he game and ignore him.
3:47 Alex has a sneaky sniff of Shadow, who luckily doesn't really register and takes it in his stride. Pattern games are hugely effective for this sort of scenario.
3:53 Starting to work on a "sniff and retreat" where Shadow learns to disengage before squaring up. One of Shadow's problems when greeting is that he stays at the head end rather than doing the polite thing (as far as dogs are concerned!) and moving to the rear end to sniff, so teaching him to be able to move away from the head is key, I believe.
4:17 Shadow didn't really want to say hi. I've tried to fit too much into a single session because of logistics, but he's not gaining anything from this interaction - except that he had an encounter with no nastiness, which is a great thing, BUT didn't achieve what I set out to because it was too much, too soon for him.
4:31 Shadow looks over when I say "Say hi", but doesn't engage at all. He's had enough!
4:40 Alex cones through J's legs and gets a bit closed (or, rather, "unmanaged") than I had hoped, Shadow was definitely more interested but disengaged easily. I think by this point it could easily have gone the other way, though, and I'm cross with myself every time I watch this video
4:47 Shadow sniffs the ground. Displacement? No, cheese I dropped earlier Not such a bad thing - unfamiliar dogs around and mystery cheese appears from nowhere. That's an OK association to make.
5:08 AT LAST! SHE STOPS!

All in all, I'm overwhelmed with how good Shadow was. I pushed too far and he took it on the chin like a complete legend. A person on the street watching this video probably wouldn't have the slightest clue that Shadow is aggressive towards unfamiliar male dogs. I learnt a huge amount by watching it back, so I think it's a really worthwhile thing for us all to do.

One other thing I would adopt now, having done my course with Amy Cook, is to use a counter-conditioning party rather than just posting food. That has its place too, but when we're trying to replace a strong negative emotional response with something at least neutral and hopefully even positive, we need to bring out the big guns.

Anyhow, let me know what you see, give me your comments, feedback, what you might have done differently. The whole shebang.

 

[Joy]

This was really interesting to watch. I picked up Shadow's uneasiness near the beginning where he's showing less interest in the treats and is beginning to stare at Alex (but if I'd really been there I'm sure I wouldn't have noticed any sooner than you did.)

He struggles to look at Alex a couple of times; this can be a sign that the dog is totally relaxed about the trigger so it's not even worth looking at, but in Shadow's case, I believe he's not looking because he doesn't want to - he finds it too stressful.

I missed realisng this. I saw relaxed ears and normal / low tail carriage and thought he was just more interested in the treats. However I can also see your interpretation is almost certainly correct because you know Shadow so well.

 

[Charlie]

Shadow did very well considering how reactive you say he can be One thing I learned is that you can cue LAT. I thought the dog had to look uncued. So I will remember that, thank you. Frankly I don't know how you remember all of the training you do, I would be in a tailspin! x

 

[Beanwood]

Really interesting!

I have only had a quick skim through, but my initial thoughts are like you at the beginning Shadow is uncomfortable with LAT. His body language and interaction with you seems a little "automatic". I wonder if "look and dismiss" may have worked, but not on cue. So if Shadow looked, then looked away, I would have clicked and given 2 or 3 really high-value treats.

I agree, being on the move is a good thing, as you can see Dante looks a bit pensive, and there is a bit of tension in his body. I do wonder if both dogs were doing something, and not focussed on the fact that they were both the only moving objects in the park, then it may have been more comfortable for Shadow, failing that I think I would have given Shadow more space, and choice (as above).

 

[snowbunny]

I have only had a quick skim through, but my initial thoughts are like you at the beginning Shadow is uncomfortable with LAT. His body language and interaction with you seems a little "automatic". I wonder if "look and dismiss" may have worked, but not on cue. So if Shadow looked, then looked away, I would have clicked and given 2 or 3 really high-value treats.

What you can't hear through most of this is that I'm using Leslie McDevitt's pattern games, primarily the "1,2,3" game, so that he is being "contained" by the pattern and the predictability of it and can therefore take note of the trigger without it causing as much of an impact as it would normally; so, yes, he is behaving quite automatically but that's cued by the game. I did faff around a bit in the beginning working out what was going to work best for him, for sure - when I used 1,2,3 I had far more success than with LAT.
For him, I don't think look and dismiss is where we want to go, or even using the clicker at all, to be honest. I need to be clearer in my own mind that I'm working on counter-conditioning NOT operant learning, and the clicker has no place in CC. He has learnt the behaviour patterns of LAT so he will offer them, as we saw in the video, but I think it's gone past the point of usefulness because I've poisoned it into him looking even when he doesn't want to and looking away even if he's not relaxed. This means it's not a useful form of him communicating with me, which is what it should really be about.

The analogy for the pattern games is this: if you are scared of heights and are asked to cross a rope bridge across a deep ravine, you would likely be frozen from fear, have a panic attack or some other combination of behaviours caused by your negative emotional response. Someone might try to get you to edge closer to the bridge, one step at a time, but with each step you only get more scared. It is highly unlikely you will make it across.
However, if you had a friend with you, giving you some simple instructions: count 1,2,3 and on "3", take a step forwards, then the rhythm of the "game" will absorb you, give you a pattern in which you understand the rules, and will make you far more likely to be able to achieve the goal.
The rhythm is almost meditative and gives very straightforward rules for a game. The dog understands it very easily and so their stress is lowered. Rather than having to make any decisions, they just keep playing the game. Stuff goes on around them, but the game is a constant.

That part is desensitisation with a teeny part of CC because of the food, but the main focus is DS. When the dog knows his trigger is there but realises he doesn't have to react, then he will be more likely to be able to do the same again in future. He may still need to play the game for weeks, months or years, but he can cope; the trigger becomes a cue that the game will start and he understands the game. The game is a safe place where scary things can be handled.

For those who are interested. Leslie McDevitt's pattern games video is available from Tawzerdog to buy or rent. It's old and has low production quality, but I think it's definitely a worthy addition to anyone's library: Tawzer Dog | Pattern Games: Clicking for Confidence and Connection

What I should be do is taking these DS exercises and layering them with real CC, and that means more than a few pieces of cheese or frankfurter. It means investing in his CC party and pulling it out whenever we have the chance to work around other dogs. Next time we come to the UK (when I can next face the drive!) I might call up some trainers with stooge dogs and see if we can put some stuff together for him. I'm also tempted to put him on fluoxetine, too. I'll talk to my vet about it, obviously, but I really do think that it would provide him with the best chance of a training programme "sticking". Hmmm, things to think about. I love "chatting" about this stuff!

 

[snowbunny]

PS I realise I've been a bit confusing there about talking about the clicker. When doing some of the pattern games (including LAT), the clicker is definitely useful! But the pattern games are not primarily counter-conditioning, they are desensitisation. When using "pure" counter-conditioning, there is nothing to click because we are not working with behaviours.