table games

by

The Goat Diaries – Still Catching Up

The Goat Palace – Foot Targets and Puzzle Moments

This will teach me to take time off from these Goat Diaries for the Holidays.  It feels as though I have entered Alice in Wonderland territory – always running and never getting anywhere.  By the time I catch up, I will be behind again.  Oh well.  We’ll see how far I get today.

I think the easiest approach is to describe where we are and not worry too much about the details of how we got there.  In my previous post I wrote about stillness: in stillness comes understanding.  I needed stillness, especially with Thanzi.  She’s so used to getting what she wants by pushing her way in.  Touching a target was easy.  That part was great, but after the click she was pushing in trying to go directly to the treats.

Thanzi pushing for treat.png

Thanzi has just touched her target, but now she’s trying to go directly to the treats.

In loopy training both sides of the click have to be clean for a loop to be clean.  The food delivery was a long way from being what I wanted.  I could move on with the targeting, but I would be dragging along with it all the unwanted food manners.  So I changed tactics and introduced both ladies to platforms.

Platforms give the feet a place to be.  I could click them for being on the platform. That was the behavior side of the click.  On the food delivery side I positioned myself so they remained on the platform while they got their treat. That meant staying very close to them which made my treat bowl even more of a target, but at least I had their feet more or less planted.  It was funny watching how much they could stand elephant-on-a-drum at the edge of the platform.  They would look as though they couldn’t possibly keep from tumbling forward, but they spread their toes and somehow managed to stay on.

When I first set out the platforms,  Trixie predictably was suspicious of the platforms.  She able to be by herself in the hallway, but she was uneasy about moving too far away from Thanzi.  She certainly wasn’t ready yet to go as far as the second platform which was set further down the hallway.  Both platforms were initially something to avoid. She was the most horse-like of all the goats.  It wasn’t that she was afraid to step on raised surfaces, but she was definitely suspicious of anything new in her environment.

Thanzi was the complete opposite.  When I opened the gate, she would dash out.  She would then stick to me like glue.  If I went to the far end of the hallway, she dashed right along with me.  She certainly had no concerns about being away from Trixie.  Instead it was hurry, hurry, hurry to stay glued to me.

The platforms, however, were also something to be avoided.  At first, I just worked on targeting and ignored the platforms.  Thanzi was getting pretty solid with the targeting so it was time to make the platforms part of the lesson.  I had my green target stick with me. I had her orient to it a couple of times. Click.  I fed so she had to back up away from me.  Then I kept changing my position.  To get to the target, the most direct route was to step up onto the platform.  This is so very different from the way I would do it with horses!

Thanzi put her front feet up on the platform, and then it was easy.  She stepped up all four onto the platform.  Click then treat while her feet were on the platform.  Finally, with her feet still, I could work on grown-ups. I held the target stick straight down in a neutral position and waited for her to take her nose away from my pocket. Click – treat.  It helped so much having her feet anchored on the platform.

In stillness comes understanding.

The platform is a foot target.  Keeping her planted on one spot helped her to notice what the rest of her body was doing.  I was standing right beside her, close enough for her to easily reach my pocket.  If I had stepped away even a little bit, my movement would have drawn her off the platform.  When she was pushing at my pocket, no treats, but as soon as her nose moved away from my pocket, click, I fed her.

When she wasn’t on the platform, her feet moved along with her nose so it wasn’t clear what she was being clicked for.  Here it was so much clearer.  I clicked because you moved your nose.  That sound you just heard is a predictor that treats are coming.  Thanzi was putting two and two together fast.

With people we often find rules restricting.  The more rules we have, the more we feel caged in.  But here rules were liberating.  Stand on this platform, look straight ahead, and you are in control. You can make this very odd person reach into her pocket and hand you a much desired piece of squash.  For Thanzi understanding created even more of an eagerness to play the game. That eagerness was expressed not with anxiety, but with an ever-growing confidence.

Once Thanzi was on a platform, she was super at staying on. I would click and treat for a short round, then I walked to the other platform holding the target out for her to follow. Thanzi raced after me, but she didn’t automatically jump up onto the platform.  She’d by-pass it on her way to get to me.  I had to maneuver myself so the most direct route to me was via the platform.  Click as she stepped up, and then another round of clicking for all four feet on the platform and her head in good “grown-ups” position.

Trixie’s sessions were similar.  They were just done in slow motion compared to Thanzi.  I noticed with her that when I switched sides, she completely lost track of the target. She had not yet generalized so I could present the target from different orientations.  I had to be on her left side, and the target needed to appear as expected.

Certainly with horses if I don’t vary my presentation, I can end up with a very one-sided learner. Trixie’s reaction to the target was more than this.  It made me wonder if her nervous personality made it harder for her to notice patterns. And then does that contribute to her feeling even more nervous because it is harder for her to predict what is going to happen? She starts out nervous and becomes more nervous because all that worry gets in the way of making connections.

Or do nervous individuals start out having more trouble making connections, and that’s what gets the ball rolling.  They start to feel anxious because they don’t understand why things are happening the way they are.  Which comes first the chicken or the egg?  Either way, it’s a snowball scenario.

As very young horses, Robin and Panda were both superb at making connections and they are both supremely confident. Thanzi is so like them.  She’s very quick and very confident. If we collected data, would this correlation between pattern recognition and confidence hold?

Is this one of the reasons clicker training helps nervous learners? By slowing things down for Trixie I am helping her make connections.  She was understanding that A leads to B in a predictable way.  That was making her bolder and more confident in her responses.

Does predictability help reduce generalized anxiety by giving you more control over what is going to happen?  If I don’t understand how A and B are connected, it’s much harder to adjust my behavior to avoid the things I don’t like or to get to the things I want.  When you’re frazzled and worried about every little thing, all that noise makes it harder to understand what is really happening.  You see connections that don’t exist, and you miss the reliable predictors.  When the world seems to be built on shifting sands, of course you will be more nervous.  It remains to be seen if the stability that platforms offer will help Trixie to become an increasingly confident learner.

With both goats I had the beginning of platform training.  They would both step up onto a platform and stay there while I clicked and treated them.  But I wasn’t convinced that they were really aware that there was something special about being on a platform.  Once on, they tended to stay on, but when I took them off with a target, I had to direct them back to the platform.  Neither one of them went to a platform on her own.

So I switched tactics.  I set a food bowl down just far enough in front of the platform that they would have to take a step off to get to the bowl.  The question was what would they do after they got the treat.  Thanzi’s answer was she consistently backed herself onto the platform, looked regal, got clicked, dashed forward to get her treat and then backed herself onto the platform.

All the backing to deliver the treat had primed her well.  She easily solved this puzzle.  Trixie was similar, just slower.  So now I had two goats who could get themselves back onto a platform after getting their treats.  Progress!  For Thanzi in particular it was time to create a real puzzle moment for her.

Kay Laurence talks about puzzle moments.  This is when you set a test for your learner to see if what they have learned matches what you thought you were teaching.  When you play the table games, you often have times when the person gives you the correct answer.  You think they have the concept you were trying to teach.  Maybe you want them to touch only the yellow object.  You’re setting two objects out on the table, one yellow and one purple, and they are consistently touching the one you want.  Click and treat.  So now you add to the choices by setting out three objects, and they touch the purple one!  Surprise, surprise.  What is going on?

They’ve been operating under a different rule.  Maybe the yellow object was always the first one you put out, or the one you put closest, or the smallest, or the biggest.  There are lots of different variables to choose from in any system.  Your learner found a rule that worked to produce the correct answer – until it didn’t.  You can go on for quite a long time thinking your learner understands what you want, but unless you test it, you really don’t know for sure.

Maybe Thanzi had no idea that the platform was significant.  Maybe she was backing up after she got her treat because she’d discovered that backing up got me to click.  The platform just happened to be in the way.  That was certainly part of how I got her onto the platform in the first place.  The question now was had the platform itself become significant?  Did she understand that being on the platform mattered, not just backing up?

I had been gradually moving the food bowl further and further out from the platform.  That meant a couple of times after she got her treat, Thanzi was no longer lined up directly with the platform.  When she backed, she ended up broadside to it.   What would she do?

I love watching puzzle moments.  This is when you can really see your learner processing what is going on.  When you train in tight loops, a rhythm emerges.  It’s get the right answer, repeat.  Get the right answer, repeat.  You slide the criterion along so smoothly that your learner really doesn’t notice that you’ve been gradually making the lesson harder.  There’s no break in the rhythm until you stumble across a puzzle moment.  Now there’s a pause.  This is when you want to be very still.  No prompting.  No helping the learner.  No giving away the answer and depriving them of ownership of the solution.  You wait and watch.  If it’s clear your learner needs help, you offer it.  You give the clue that will make solving the puzzle possible, but first you wait to see if that extra clue is needed.

The reinforcer for waiting was seeing Thanzi step sideways up onto the platform.  Click and treat.  She was understanding.  Several more times in that session she got off line from the platform and each time she very deliberately deviated from where she was to step up onto it.  Her actions told me that she understood the platform was indeed significant.

So now she was ready for the next puzzle moment.  She was solidly on the platform.  I could click, and then walk to her to give her a treat.  She would wait for me to bring the treat to her.

And I could also drop the treat into the food bucket.  She would leave the platform, get the treat, and then back up onto the platform.  I was impressed by how well she understood the two forms of treat delivery.  There was no confusion between them.  She was reading my body language well, and she was waiting on the platform when that was indicated.

I was also impressed by how quickly I had been able to open up space between us and how solid she was about staying on the platform.  What a smart goat!

So now I presented her with a puzzle moment that made her head hurt.

I tossed the treat into her food bowl.  The platform was behind her.  As she lifted her head, I began to take a further step back in the opposite direction from the platform.  The food was going away!

In the world in which she had lived up until now you followed food, and you made sure you were the first with your head in the bucket.  In this alternate universe that she now found herself in, backing up away from the food bucket got you treats.  She had worked her mind around that concept.  Going back to the platform got me to reach into the metal bowl I was holding and hand her a treat.  But I had never moved the bowl away.  Now it was leaving!

The question was could she back up as I backed up?  Could she allow the gap between us to keep expanding.  Would she be able to cope with this truly inside out world?  I backed just a little and waited.  It was so clear her head was hurting with all the computing she was doing.  As smart as she is, she wasn’t used to having to think so hard.  No, she just couldn’t do it.  Not yet.  Not this time.  I helped by stepping slightly forward.  She went onto the platform.  Click and treat.

I did another couple of rounds where things were kept as she expected them to be. I clicked and reinforced her a couple of times while she stayed on the platform.  Then I clicked and dropped a treat in the bucket.  She came forward to get her treat.  As she lifted her head out of the bucket, I took a step back.

And this time she could find the answer.  She let me take the food bowl further away as she backed up to the platform.  What a truly fast learner she is!  What a smart goat.  No wonder she’s such a powerful leader.  When she stood up on her platform waiting for me to click, she looked so regal.  I hope she was as pleased as I was by how well she had solved that puzzle!

So that is pretty much where I am with the four I am working with.  Elyan and Pellias are learning about working together and sharing.  It sounds like kindergarten.  Always I am reminded of Robert Fulghum’s charming book, “Everything I Ever Needed to Know I Learned in Kindergarten”.  The ladies are learning about platforms.   Marla has continued to work with Galahad on targeting.  She was away over the Thanksgiving Holiday.  When she got back, she also introduced him to platforms.  So all the goats have taken a major step forward in their basic education.

Coming next we’ll see where all of these good puzzle pieces are taking us.

 

 

 

 

by

The Goat Diaries – Day 3: Arrange the Environment for Success

The Goat Palace – Journal Report for 11/19/17: You Never Know What You Have Taught

Galahad had the first session of the day.  He’s an eager, happy learner, and he very much chose to go into the far end to play.  I stayed for a few minutes down in the front section visiting with the other goats. Elyan and Pellias were up on the top platform of the jungle gym.  They were eager for head scratches. Surprisingly, so was Thanzi.

By the time I extracted myself from their appeal for more, Marla had already begun Galahad’s session.  She commented that what she thought she had taught him was not what he had learned.  Ah yes, that’s the clinic mantra: You never know what you have taught.  You only know what you have presented.  Yesterday he had been going to his target, click, followed by Marla dropping a treat in one of the food buckets.  He went promptly to the bucket, got his treat, and then touched the target again. Marla would then drop his treat in a second bucket, so he was going back and forth between buckets with a quick stop in between to touch the target.

His takeaway from that was just to go from bucket to bucket – never mind touching the target.  It reminded me of the table games that we play to learn about training and to work out procedures for teaching concepts.

Training game

Playing the table game during the Five Go To Sea Caribbean conference cruise.  I’m hiding from the sun under the funny hat.  Kay Laurence is sitting behind us.

Kay Laurence is the originator of these games. Several years ago we were together at an airport, both with long waits for our flights home.  So we found a quiet corner and pulled out a table game kit.  I was the learner, something when I’m teaching I rarely get to be, so that was a treat.  Kay had a plan in mind for teaching me to use the pieces from the game to draw a pentagon.  Of course, I had no idea what she had in mind.  But I was a contented learner because I was making lots of correct choices and getting clicked and reinforced  at a high rate.  The only problem was the rules I was using to produce the actions she was reinforcing were not the same rules Kay was trying to teach.  So I was coming up with the right answer but for the wrong (from Kay’s perspective) reason.

Every time Kay presented me with a puzzle moment I got stuck.  Puzzle moments are small tests to check to see if what you think you are teaching is what your learner is learning.  It was a fascinating and fun experience, though it could easily have been a frustrating experience if either of us had brought a different mind set to the game.

My flight was coming up, so we had to end the game.  Kay explained what she wanted me to do.  My reaction to being told the “answer” was interesting.  I felt deflated.  I wanted to go on and work through the puzzle.  Being told the answer was far less satisfying than discovering the answer on my own.  I missed the puzzle solving, and I missed seeing what strategies Kay would have used to get things sorted out. But my plane wasn’t going to wait for us to finish the game, so we had to jump straight to the final answer.

Galahad had come up with a solution to the puzzle that made total sense to him.  Go from bucket to bucket and expect your person to drop a treat in when you get there.  He had completely by-passed the target.

Watching him, I also didn’t think he was noticing Marla’s tongue click. With horses I suggest that people begin with an actual clicker.  The sharp sound that a box clicker makes is very noticeable, and the horses seem to catch on fast to the significance of the sound.  After a couple of targeting sessions with the clicker, you can switch to a tongue click, and the horses are very aware of the new marker signal.

I suggested to Marla that she get an actual clicker.  At the stage where you’re using target sticks, clickers are easy to use. You can duct tape a box clicker onto the end of the target stick so you have easy access to the clicker.

Marla got a box clicker and continued on with the lesson.  Galahad quickly remembered that he was supposed to touch the target. Yesterday’s fluid pattern was back. Now it was: orient to the target, click, go to the indicated food bucket for a treat, look for the target. A clean loop was reappearing.

This experience highlights another part of the start-up process.  I like to begin with very short sessions.  With horses I have people count out twenty treats.  That means handlers who are new to this process have to stop frequently to reload their pockets. This also gives them time to think about what has just occurred and to consider what, if any, changes need to be made.

With five goats to juggle I was certainly finding I needed to do a lot of adjusting.  It wasn’t just what was happening with the individual I was focusing on.  What was going on with the other goats?  When I had Pellias out by himself, he was having a grand time, but how stressed was Elyan?  Was he being chased by Thanzi?  Yes.  When I took Thanzi out, was Trixie able to cope?  There was a lot to think about, a lot to keep shifting around to find the right training combinations.

Keeping your initial training sessions short lets you check in with your animals more frequently to see what they are actually learning. Each time you go back in and start up the session, you get to see what’s been processed from the previous session. If your learner has come up with a different answer, these short sessions mean it hasn’t become so entrenched that it is now hard to shift the pattern.

It is ironic that I am writing about short sessions, because I am known for using long training sessions. With an established learner I’ll fill my pockets with treats and keep going. That seems to suit the learning style of horses, but these long sessions are broken up into smaller units. I give breaks through the behaviors I’ve taught. For example, I might be working on lateral flexions. We’ll have a bit of success, then it’s off to find a mat. The mat acts both as a conditioned reinforcer and a way to give a break. The change in the rhythm of the training provides a break without having to stop the play.

At the heart of this is the training principle: for every exercise you teach there is an opposite exercise you must teach to keep things in balance.

The balance that I thought was needed now for the other goats was a morning session of quiet visiting.  I was very pleased that Thanzi wanted to participate in some head scratching.  I had the two ladies in the back section so the three youngsters could relax and not worry about dodging out of Thanzi’s way.  She stayed by the gate while I scratched her head.  Normally, she’s been drawing away when I try to touch her, so I consider this real progress.  Trixie came up to me repeatedly through the morning, but she’s not yet ready for a proper scratch.  The boys, on the other hand, had a blissful time enjoying a prolonged cuddle session.

Afterwards, Marla and I worked some more on the Goat Palace.  We’re getting close to the finish line, but there always seem to be a few more things to do.  Years ago my family did some remodeling to the house.  The process dragged on and on.  Every day my father would make a list of things that the builders still needed to get done before he could sign off on the job.  He remarked that they always seemed to get done only half the remaining jobs.  You would think on a finite project like that, you would be able to check everything off the list, but it never seemed to happen.

At the moment we seem to be caught in that twilight zone of always completing just half the remaining tasks.  One of yesterday’s tasks was tidying up the section we’ve designated for storage.  I was very pleased to see how little we have left to store.  We have managed to use up an amazing amount of miscellaneous clutter.  So perhaps when we run out of stuff to find a use for, we will also run out of tasks that still need to be done. That will finish off phase one of the goat palace.  (I say phase one because phase two is obviously going to be expanding the goat jungle gym. That will be as much for our entertainment as it will be for theirs.)

One of the things that contributed to the tidying up of the storage area was the snow blower went out to be serviced for the winter.  That left a clear area that could be used for training.  So in the early evening I took advantage of this space to work with Elyan and Pellias.  It was a good time for training.  The goats were beginning to settle down for the night.  It was easy to close the middle gate so only Pellias and Elyan were in the front section.

I had everything set up for them out in the storage area.  I had my chair, a food bucket and a couple of platforms, including the very distinctive foam platform I had introduced them to in July.

Elyan came out first.  I brought him out on a lead, and then turned him loose.  He stayed nearby.  He was clearly interested in playing, but he wasn’t sure what to do.  I let him explore for a couple of minutes, then I brought out the baton and directed him towards the foam platform.  He hopped up onto it, click, I dropped the treat into the bucket.  He had to step down from the platform to get to the bucket.  So now the question was what would he do?  The answer was he backed up to get back on the platform. Click! Drop treats in the food bucket.

Elyan seemed to catch on fast.  The “rule” was get back to the platform, and you’ll get clicked.  At least that’s what was happening.  His “rule” might just as easily have been: back up, and you’ll get clicked. The platform was just in the path of the backing. I’ll need to have a puzzle moment to check whether he is going to the platform or simply backing up.

In any case, while he was getting his treat, I nudged the platform a little further away.  He continued to back himself onto the the platform.  We could have kept going all night, but this was a session that should be kept short.  I got up from my chair, and he followed me back in to the front section.

Pellias was eating hay.  He hadn’t been at all fussed having his brother outside the pen.  But now I wanted to do a swap, and they were both at the gate.  I got Pellias out and sat down in my chair.  He went straight to the platform.  Click.  I dropped treats in the bucket.  He stepped off the platform, got his treat and went straight back to the platform.  I repeated this a couple of times, and then I exclaimed; “Wait a minute.  You’re not Pellias!” In the fading light I hadn’t noticed that little Elyan had pushed past his brother for a second turn.  With his jacket on to keep his coat clean, it was harder to tell them apart. No wonder he was so good!

I got them switched around so now it truly was Pellias’ turn.  He’s always been a platform superstar.  He went straight to the foam platform.  Click.  But now the food delivery was different.  He’s used to getting the treat from my hand, not a food bucket.  I moved the bucket close to the platform and helped him find the hay stretcher pellet.  He got his treat and then stepped off the platform. He wandered away from the platform. I waited.  He began to eat the leaves that we hadn’t swept out of this area.  I got out my baton target and gave it a little shake.  That got his attention.  He followed it to the platform, click, drop the treat.

The hay stretchers make a very sharp noise as they fall into the bucket.  That helped draw Pellias’ attention, and he began to look in the bucket for his treat.  He only had to take his front feet off the platform to get to the bucket, so it was easy for him to step back onto it and get clicked.  My concern was the sound of the treat dropping into the bucket might become the functional marker signal, so I clicked, and began to wait to see him react to the click before I made any move to drop the treat into his bucket.  I got lucky several times with that.  He had turned on the platform so he could look down the driveway.  The sound of my tongue click turned him around, so it was clear, at least in this situation, that he was responding to the sound of the click.

Again, I kept the session short.  When I opened the gate to let him back in, I dropped treats on the floor to distract Elyan.  Pellias came in to get the treats, as well.  I’m not sure I want the others out in this area yet, but for these two their July visit prepared them well for going outside of their pen.

I filled their hay feeders, opened the middle gate and left the goats tucked in for the night.

Today’s July Goat Diary appropriately enough continues with the initial training of platforms.

The July Goat Diaries: Clicker Training Day 3: Arrange The Environment for Success

I described earlier the morning sessions of day three in which I introduced both goats to platforms.  This was an errand day so I wasn’t able to fit in as many sessions as usual. When I got back to the barn around 5, E and P were clearly hungry. They were standing on a bed of hay, but none of it was to their liking. I gave them fresh hay and left them to eat while I did barn chores.

7 pm session with P

P was very rambunctious – literally. He reared up several times. I managed to dodge him and get him on the platform, but the session didn’t feel very productive.

I wasn’t satisfied with the way he was orienting to the target. I thought a second platform might help. If a platform was the end destination, it might make more sense to him why he was following a target. I decided to consider this a data collecting session.  I knew where I needed to head, but I would wait until tomorrow to add the second platform.  Training success depends very much upon having a good set-up.  I suspected adding the second platform would help smooth things out.  Instead of continuing on with a session that wasn’t going well, I would wait until I had a better set up.

In contrast to P, E’s session was great. He was so very soft and sweet. I had him target the baton, click, treat. Then I scratched him around his ears. His eyes got soft, and he leaned into my hand, clearly enjoying the feel. I asked him to follow the target again, click, treat, scratch.  Who knows what E was learning.  I certainly found it very reinforcing!  I began his day with bliss, and that’s how I ended it.

The password to open this video is: GoatDiariesDay 3 E Learns

Note: When I was in town, I stopped at the new bird store that’s just opened.  I bought some black sunflower seeds which the goats really like. So now they are getting a mix of sunflower seeds, peanuts and hay stretcher pellets.

8 pm final session of the day.

We ended the evening with “cuddle time”.  While Ann groomed Fengur, I took my chair into the stall and enjoyed a few minutes of goat bliss.

Coming Next: Clicker Training Day 4

Please Note: if you are new to the Goat Diaries, these are a series of articles that are best read in order.  The first installment was posted on Oct. 2nd.  I suggest you begin there: https://theclickercenterblog.com/2017/10/02/