Eye Gaze used as the only cue

The video below is of maybe the 10th session of Lil playing a new “eye gaze” game. The only cue I am giving her re: which toy to target is shifting my eye gaze towards one of two toys.  She picked up on this game so quickly (like in the first session) it is likely she has already been using the gaze of my eyes as a “cue” beyond this game. The inspiration to set up this game was to determine how much dogs pay attention to where their “people” are looking since this has been a big topic amongst agility folks lately.

 

Yet another reason to love Silvia Trkman….like anyone needs one!

My recent post for Silvia’s Agility Foundations class:  Here is a short video of Lil’s current serpentines plus a few reps jumping over angled jumps. We have not practiced serps for a long time due to focusing on straight jumping and cip&cap.

Lil will normally seek out and take serpentine jumps with very little support from me. I wonder if the reason she ran past some jumps was because I was not in my usual position (which is typically further ahead).  It felt good to feel rushed by Lil’s increased speed and drive over the jumps.  I definitely felt more comfortable when I took more of a head start and was a bit further ahead of Lil.  I’m not sure any of this is visible due to the camera angle though.

Lil’s speed has picked up considerably since starting this course and as a result I am having so much fun running with her….not that it wasn’t fun before! It’s just more fun now!

Now I just need to be disciplined enough to do very little jump training between now and next weekend since we have another trial. Jumping has become so much fun its like eating ice cream… hard to stop until the bowl is empty! But I want Lil’s “jumping bowl” to be over-flowing for the trial!

LoLaBu’s avatarLoLaBu

That sure was fast! Great job! She is really flying, so… – who cares about a couple of missed jumps :) You do want to be ahead for a serpentine yes and you can help some with the arm too, you just don’t want to do any extreme turning as you want to keep running.

A theoretical model for training running contacts

I developed this theoretical model as an exercise in how I might approach training a running dog walk using back-chaining.  There are already proven methods for training fast and consistent running dog walks.  My intention in posting this is to generate ideas and share thoughts about training RCs with other agility enthusiasts and see where it leads.  I see it as an experiment in collaborative, creative thinking, and a way to use what we know about dog training and the behavioral sciences to see how good a plan we can come up with as a group.

Participants read the rough draft below and posted comments (plus I received some private emails) about what they would do differently and why, or posted their personal experiences with their dogs, or asked clarifying questions.   I have made changes based on the great feedback I received. This is still a rough draft so please excuse typos, potentially unclear descriptions, and inconsistency in formatting.

The following concept is based on back-chaining.  I would train this away from any “real” agility equipment to test the theory so no harm would be done to my dog’s contact performance.

STEP 1) Since this is based on back-chaining, training would start with the last behavior:  Dog knows how to GO ON, turn RIGHT or LEFT over jumps or cones…
*Foundation Skills:  an independent and fast GO ON, and RIGHT and LEFT directionals.
*Criteria: Handler should be able to send the dog, run with dog, run ahead of dog, peel away, hang back, do FCs, RCs, and… gasp… even blind crosses :).  Handler should be able to move to every imaginable position at a distance, up-close, and mid-range.  Performance should be  fast and consistent in at least 2 habitats.

STEP 2)  Use shaping to train the dog to step onto a small mat or board with all four feet and then jackpot when the dog steps off the mat/board with front feet into a 2o2o.
*Foundation Skills: dog has a ton of value for seeking out, walking over and standing in 2o2o on a small mat/board.
The reason using a short length mat or board is to encourage the dog to really think about the mat or board.  I’d try to come up with an unusual surface that the dog can really feel when they step on it.  Examples: a  yoga mat wrapped around a thin, foam “pillow,” or a mat on a board that wobbles slightly, or a yoga mat crinkled up and glued onto a board so it forms ridges, or any of these items glued to the top of a “hit it” board (with volume turned off to start).   If I used a foam “pillow” I would fade it over time by making the foam thinner and thinner until it was using just a plain yoga mat.  If using a wobbling board, I’d fade the wobble over time and end up with just a thin board or mat.
*Criteria: The dog is clicked and rewarded first for putting front feet on the board or mat, then for standing with all four feet on board, and finally jackpotted for stepping off the board into a 2o2o.  Dog is “reset” after each rep by releasing forward to a thrown toy, MM or food reward.

STEP 3) Adding motion: Dog runs to the mat/board and as soon as the dog hits the 2o2o, release forward to a toy, MM or bait bag.  I might mix crate games by putting a crate 5  front of the board, then gradually increasing the distance to 30′ or more (why not add speed and distance early on?).
*Foundation Skill:  Same as STEP 2.
*Criteria:  Click and reward 2o2o and then release forward to a manners minder or throw a toy or bait bag about 5 ‘ straight ahead of the board.  After a bunch of successful reps in a row, I’d sometimes send to the dog forward as soon as the dog hits the 2o2o position vs. rewarding the dog on the board every time.

STEP 3a)  Add a jump after the mat/board on the reward line so the dog takes the jump on its way to getting the reward.   NOTE: The reason adding a jump is so important is because it turns the mat/board into a directional “obstacle” so that you can send your dog ahead to the mat/board and it knows to continue focusing forward.  If there is nothing in front of the mat/board, the dog would be correct if you sent her to the mat/board and she did a 2o2o facing back at you.  Progress to using three jumps: one straight ahead, one ahead to the right, and one ahead to the left to practice GO ONs and directionals.  Cue the dog to GO ON, RIGHT or LEFT as the dog is approaches the mat/board.
*Criteria: Dog still hits the mat/board with all four feet, briefly pausing in 2o2o.

STEP 4) Get rid of the slight pause so the dog is striding over the mat/board and continues over a  jump to a toy or MM without pausing in 2o2o.
*Foundation Skills: a dog that will GO ON and go RIGHT and LEFT independently and has enough value built up for standing in 2o2o on the mat/board that the dog strides across the board with all four feet  touching the board but does not stop.
*Criteria: All four feet hit the board.  Due to the short length of the board, this will be a very compressed stride and should be easy to see.

STEP 5) Gradually add more extreme turns to  jumps including 90 and 180 degree turns. I’d repeat all the variations from step #1.

STEP 6)  Increase the distances between the starting point and the mat/board so the dog is running a full speed with long strides when approaching the board.  A “hit it” board would come in very handy here (especially with my small dogs since their short, fast legs are hard to see in real time.
*Criteria: Dog approaches the mat/board running at full speed and then compresses its stride so that all four feet come in contact with the board.  I’d continue to increase the distance until the approach is at least as long as the length of a dog walk.

STEP 7)  Add a 8-12′ long x 12″ wide plank about 5’ in front of the mat/board and gradually move the new plank closer to the original mat/board until they are touching and finally place the mat/board on top of the plank.  The reason I would not start with the mat on the end of the plank is because my dogs already run over a plank and I wouldn’t want to suddenly change the rules.  But I think by gradually reducing the space between the plank and the mat, they will not be confused when they are faded together.  I would paint the new plank a different color than the mat/board so I could easily see where the mat/board begins.  I might even add more planks until the dog is racing across the entire length of a DW on flat planks at top speed. Handler moves to every imaginable position listed in Step 1 mixing in GO ONs and turns.

STEP 8)  This step could be done before step 7 too on just a plank (no mat/board yet).  Train angled approaches to the plank by setting up every possible scenario for dog and handler, using a small object placed at the front edge of the plank to encourage the dog to approach the plank straight on (like Silvia Trkman does).  I would keep switching between using different small objects placed at the corner of the plank and I would also sometimes not use any object to mark the corner so the dog did not become dependent on the object being there to have a straight approach.

STEP 9)  Not sure this would be necessary, but I might also run the dog over a plank on a gradually increasing angle (like Silvia Trkman does), with the mat/board at the end.

STEP 10)  Transfer the behavior to a real DW by placing the mat/board at the end of a DW.  I’d start every session (early on) doing a few reps with the dog starting close to the mat (on the down ramp) like Step #2 to reinforce the behavior.  I might start with a lowered DW (if I had access to one) but I’d quickly increase to full height since a DW is not very steep and doesn’t affect a dog’s striding like the extreme angle of an A-Frame does.

STEP 10a)  Fade the mat/board.  If the dog is running fast across the dog walk and hitting the mat consistently,  I’d quickly start alternating between using the mat and not using the mat so that the dog’s performance on the DW did not become overly dependent on the mat.

ENDING NOTE:  I don’t think this theoretical RC training method would create as fast a running DW as Silvia Trkman’s method, since it creates a very compressed stride at the end, but perhaps training slight collection to maintain the criteria of 4 feet running through the contact zone (vs. 2) would result in a more reliable performance.

Please post ideas as comments on this blog vs. posting on a closed Facebook Group page so that everyone can read your thoughts, insights, comments.  Thanks!

Weave pole entries, sends to tunnels, and a lot of fast running

Below is a video of Lil’s 4th session working on Silvia’s weave pole challenges and I am amazed at how much she has improved!  Using channel weaves makes so much sense.  Lil did more reps in this one session than she normally does over a 6 month period with no strain on her body whatsoever.

We also did SENDs and GO ONs with a tunnel and mixed in some weave poles.  My matting is too slippery to work on tight turns to FCs or post turns but that is OK because GO ONs are just what both of my dogs need right now to find a better balance between handler and obstacle focus.

Throwing a ball as a reward has increased Lil’s speed and drive considerably.   Lil has always loved to chase balls and fetch.  I can’t believe I had this “speed tool” in my tool bag all along and wasn’t using it!

LoLaBu on March 8, 2012 at 13:08

Silvia’s response to my video: That’s sure really great progress already! Looks like she figured out what this game is about! And yes, I also think go is more important as come at this point. And she sure seems to like it! Really great distance skills with tunnel sends and some really cool entries!

ps–I am only uploading videos of Lil right now because I am posting them as part of Silvia’s Foundations class.   But I am also video taping and reviewing Jake’s performances.  He is not being neglected and is doing very well too!

An another note, I am I am taking a few days off from training “Loop and Wrap” since my yard is now thawing out and very wet.  Plus I think taking a few days off from working on any skill is good especially if a snag occurs.

A couple of days ago, I had a brilliant idea….or so I thought at the time.  I would do a short “Loop and Wrap” session in the bedroom to provide a new environment and I would rig up an impromptu jump with a tiny base so Lil wouldn’t have to hop over long support bars.  The new jump consisted of a 2′ tall PVC pipe that Lil would wrap around, a partially deconstructed wing on the far side, and a 2′ long PVC bar set at 4 inches.  I think the combination of a new environment and the strange mini-jump opened up a can of worms.  Lil started hitting the bar with her back legs so often that I wondered if she thought we were playing “touch that object with your back feet” game which she likes to play.  She even tried stopping while straddling the bar once and looked at me with an expression of “Is this what you want?”

OMG.  How did this happen?  Well, I have been saying YES when Lil approaches the wrap, thinking I was marking the commitment point and the set-up step.  But I think my timing contributed to Lil’s confusion because I was saying YES before she cleared the bar, so I was marking both clearing the bar and hitting the bar.  I play a lot of shaping games with my dogs and so when Lil happened to hit the bar a couple of times in a row and I had already said YES, I think that she thought she was supposed to touch this strange apparatus with her back feet and proceeded to do so consistently for a few more reps.

Bob Bailey is so right when he said: “What you click (or mark) is what you get!”  Of course, at the time I didn’t fully recognize what was going on so I ended the session and brought Jake in for his turn.  Jake had no trouble generalizing that this was a jump and that he was supposed to jump it.  Whew!  Anyway, I finished by taking Lil out to a dry spot in the yard and did a short “Loop and Wrap” session around a pole to end on a positive note.  I think  taking a few days off from “Loop and Wrap” will erase Lil’s memory of that strange little “shaping” session!

Finding the balance

Over the past couple of days,  I have started working on improving my mechanics in order to reward my dogs by flinging a toy on a rope forward to initiate a game of tug.  Silvia Trkman makes is look so easy but there is a lot going on and it all happens while she and her dog are running super fast.  Based on my observations, here is a description of what I think she is doing:

Silvia drops the toy out of her hand at a precise moment to mark a behavior (such as her dog turning tightly while wrapping a jump at a distance).  She continues running with the toy flying behind her while her dog chases her and the toy.  The split second before her dog can grab the toy, she flicks the toy forward so her dog accelerates past her and grabs the toy as it flies forward.  Then they play a game of tug while walking back for the next rep.  Silvia’s timing is always perfect.  Her dog gets the reward while it is running super fast vs. the dog having to slow down to grab a toy dangling at her side.  It makes perfect sense to reward in this manner since dogs can run faster than people.

I have practiced the mechanics without my dogs for a few minutes here and there.  I run while glancing over my shoulder to see my imaginary dog, then I drop the toy out of my hand and let it fly behind me while I continue to run and my imaginary dog chases me and the toy.  Then I fling the toy forward so my imaginary dog accelerates to grab the toy.  After doing it a handful of times, it seemed to work well… but that was without a real dog.

When I tried it for the first time with my dogs, it didn’t work nearly as well. I thought the problem was solely with my timing and mechanics, but after a short session of just playing with my dogs and a toy on a rope yesterday, I realized that I had inadvertently trained my dogs NOT to run past me when I am also running.   Jake and Lil do a lot of freestyle tricks in heel position on both sides so they have been heavily reinforced for being at my side.  We also play “recall to side” games with distractions ahead (like the Manners Minder) and flat work where they know the game is to stay at my side no matter what I do, whether I am running fast and stopping abruptly, doing front crosses, post turns, or circling with my dogs on the inside and outside of the circles.  They both really enjoy these games, but I am fairly certain that when they are chasing me while I am running with a toy on a rope, they think they are doing the right thing by staying at my side vs. driving ahead to grab the toy when I fling it forward.  They are just being good dogs!

There are a lot of awesome dogs in Silvia’s Foundations class, including some amazing high-drive, herding dogs.  As I watch those dogs drive out of turns or tunnels in pursuit of a thrown ball, I am amazed at their speed and toy drive.  I suspect if I had a super high-drive dog, I’d make sure I played a lot of games that reinforced the dog for coming to side to keep the balance (like the games I have been playing with my dogs).

But with Jake and even more so with Lil, their natural tendency (combined with past training) is to respond to my movement (acceleration/ deceleration/ shoulder turns) so instead of running ahead to grab a toy that I fling forward on a rope, they will pace themselves to stay at my side.  This tendency has caused us to almost trip over each other and get tangled up with the rope a few times when I’m forced to decelerate to avoid running into a wall or something.  Not a pretty site and potentially dangerous.

Now that I am aware of this, I am going to shift the balance by playing more “Race Me” games and GO GO GO games (which are already part of Silvia’s Foundation class) so I can reinforce my dogs for running past me until each dog has a good balance.  I’m not saying I want my dogs to think it is OK to run past me when I decelerate or stop when I am calling their names, but when I say GO ON, GET IT, or GO GO GO, I want them to know it is OK and GREAT to race past me to whatever is in front of them, whether it is an obstacle, a toy on a rope, or a thrown ball.  I can envision this being a lot fun for all of us!

In the mean time,  I will continue to reward mostly by throwing a ball, while I play with toys on a rope as a separate activity until my dogs figure out how to play this new and fun chasing/ tugging game.

Update: I just used a toy on a rope to reward Lil for a few quick Loops and Wraps around a pole in the backyard and she flew ahead of me and grabbed the toy every single time.  WOW!  That didn’t take long for her to GET!  Now.. onto Jake…

I am enjoying this class so much.  It is challenging and I love learning how to become a better trainer and handler. Plus Jake and Lil’s enthusiasm for Silvia’s games keeps going and going…. they are like Energizer Bunnies.

Below are Silvia’s responses to comments I posted relating to my dogs and toys on ropes:

My Comment:  Gosh! Sorry to be posting again about toys on a rope but I just successfully rewarded Lil with a toy on a rope doing Loops and Wraps around a pole in my backyard. While I was running, she raced by me and grabbed the toy every single time. This new game certainly didn’t take long for her to get! Then she tugged like there was no tomorrow! Her best tugging ever. She wouldn’t even drop the toy like she always does when I say drop it. I think this is a very good thing for my normally very obedient little girlie. Prey drive won out! YEY!  She did eventually drop it and I immediately flung it out again for her to chase.  Sorry if I’m being overly gushy here but I can’t believe Lil got so crazy (in a good way)! I promise this is my last post about Toys on Ropes!!! :)

Reply

LoLaBu’s avatar

LoLaBu on March 6, 2012 at 19:13

Yay for Lil! Not dropping a toy is always a good sign, I agree!!! :)

Very good observation, it’s probably in fact the reason why they won’t drive after a toy! But if they will drive after a ball, that’s a good start already. I never understood why running past me and throwing balls would be bad – only heard it’s a bad thing a couple of years ago anyway :) – I think it’s great!!! I want my dogs to drive ahead as hard as they can – and when I want them to stay close, I just tell them so :) I want them to really understand both, handler and obstacle focus. Works great for me!

My Response:  Silvia, One of the things I love about the way you coach your students is that you suggest different approaches for different dogs vs. sayings stuff like: “All dogs in this class must do X, Y, Z and if you cannot get your dog to do X, Y, Z, then there is something wrong with your training or (worse yet) you have a bad relationship with your dog…” which is so ludicrous!

I heard chasing a ball is a bad thing to do with agility dogs so I stopped playing fetch with Lil for about 9 months (and Lil loves to play fetch). Silly me for not questioning that statement!

Reply

LoLaBu’s avatar

LoLaBu on March 6, 2012 at 21:46

Very often, finding ways to work around NOT-having X, Y, Z builds the best relationship :) I noticed that with my dogs – letting them be who they are and working around their weaknesses, focusing on their strengths instead, makes us a real team. I think they really appreciate I appreciate just how they are.

How Lil learned to throw a Frisbee like a discus

Teaching Lil to toss a Frisbee in a discus-like manner involved stringing together a few rather complex behaviors including: catching a Frisbee, spinning in a 360 degree circle while holding the Frisbee, and whipping her head to the side and releasing the Frisbee at a precise moment so it flies back towards me.

This trick started by capturing a behavior that a lot of puppies do naturally: throwing toys into the air. Over time, Lil learned to flip toys higher and higher in the air, and it was not uncommon to see a toy fly across the living room while we were watching TV (which looks so funny).  The “Flick It” trick is now mostly under stimulus control so she rarely  flips toys across the Living Room anymore.  This is a good thing since Lil considers anything on the floor to be a potential toy and a marrow bone came close to hitting the TV screen once.

Here is how Lil learned this trick:

As I mentioned, she already knew how to flip a toy over her shoulder so I started by throwing the Frisbee just beyond her reach so she had to turn away from me to pick it up, which put her in a perfect position to flip the Frisbee back to me over her shoulder.

The first few times, I clicked and rewarded Lil for just picking up the Frisbee.  When she heard the Click, she’d drop the Frisbee and run to me for a treat.  Little by little, I started delaying the click, so Lil started turning back towards me with the Frisbee still in her mouth, before I clicked.  Her enthusiasm for the game, along with her past experience of flipping toys over her shoulder, made it easy for her to start tossing the Frisbee by whipping her head to the side while she turned.

At that point, my placement of the Frisbee dictated the direction of the toss.  So I refined that position to make it easier and easier for her to throw it back to me vs. in a random direction.  And since Lil was then able to throw the Frisbee 90% of the time,  I started rewarding only “better than average” performances (direction OR velocity).   It looked good, and was fairly consistent, so I left the trick alone for a long time, thinking it was complete.

Somewhere along the line though, the idea occurred to me that it would be super cute if Lil learned to catch the Frisbee before flipping it back to me.  But this would mean that Lil needed to learn how to flip the Frisbee back towards me from any position.

I started this new phase of training by rewarding Lil for just holding the Frisbee in her mouth. I quickly added that behavior to another trick she already knew:  how to do a 360 degree spin.  Once she was able to hold the Frisbee and turn 360 degrees without dropping it, I needed to figure out how to add the “head whip.”   The only way I could think of was to alternate training two separate behaviors:  “360 spins while holding the Frisbee,” and the “head whip” from Lil’s original Frisbee trick.

I figured eventually the two behaviors would start to blend.  Plus I couldn’t come up with a better plan! 🙂   Not surprisingly, the direction of Lil’s tosses went out the window for a while as she often released the Frisbee too early.  But I rewarded any tosses that had good “head whip”/ velocity regardless of the direction, and I also rewarded any full 360 degrees turns while hold the Frisbee, even it she just dropped it after the spin.

I didn’t train this regularly at all, often taking weeks or even months off between sessions. When I did bring the Frisbee out, I just clicked and rewarded “better than average” performances (velocity OR direction).   And as you can see in the video it is slowly coming together.  ps–Some sessions have better Frisbee velocity than what was shown on the video, but I was very happy with the results of that particular session because I thought Lil was really trying to spin AND toss the Frisbee towards me in one motion.

Over time, I think Lil will be able to blend “the catch, the spin and the release” into one fluid motion.   I also anticipate the Frisbee will gain more and more velocity as Lil learns to whip her head to the side at the precise moment necessary to use the momentum of the 360 turn the way a discus thrower does.

Another reason to love Silvia Trkman

I recently sent the following FB message to Silvia Trkman:

Hi Silvia, I took a workshop with you in NH a couple of years ago with my 2 Australian Terriers and loved it. I took detailed notes and continue to follow your suggestions. I love your fun attitude and approach to training and fun is always the #1 thing I focus on with my dogs! As a result, both of my dogs are fast and enthusiastic! I hope it is OK to ask a question about Lil’s increased drive and speed (Lil turned 3 years old today). She is very enthusiastic (and talented) in freestyle and agility. She moved up to Masters in USDAA very quickly. Her drive and speed continue to increase and it seems to me that she doesn’t know how to manage jumping well when she runs REALLY fast, especially on long straight-aways. I feel like she has lost confidence about jumping and so I was thinking about taking a break from jumping and just running Lil through wings without bars when we practice agility for a while. And then starting up again with low bars and gradually bringing them back up to her jump height. Is that a good or bad idea? Any suggestions you have would be greatly appreciated. All the best, Devorah Sperber (Jake and Lil)

Silvia’s Response (a few hours later):

Happy to hear Lil is doing so well!!! Sounds like a good plan to run her without bars some and then slowly adding height to let her learn how to jump at her new speed, it’s definitely harder when running real fast! But I’m sure she will master that too!

Jake and Lil at Silvia Trkman Workshop, December 2009