Category Archives: handling analysis

Changing how I train

As a result of quite a bit of reading on what constitutes effective practice, I have changed how I train dog agility.

First, what is effective practice? Back in November, I read this book and found it left me with a lot to think about. I’d heard a Freakonomics podcast on learning new skills (this one) and also discussing the “10,000 hour rule” (from this book by Malcolm Gladwell). These sources collectively introduced me to the idea of effective practice and got me thinking about how I learn new skills.

Effective practice, as I see it, introduces new skills carefully and deliberately–not necessarily slowly, but definitely thoughtfully–so that they’re learned without error. In agility, it would be the idea of teaching a tunnel-contact discrimination as follows:

  • Tunnel and contact are ten feet apart and you’re standing between them, closer to the obstacle you want, with the dog on the side closer to the desired obstacle. You release the dog, say the name of the obstacle, and take a step toward the desired obstacle; in short, you make it fairly easy for the dog.
  • Next, you move a little closer to the middle, then repeat.
  • You move closer to the “wrong” obstacle, then repeat.
  • Gradually you make it harder for the dog, but you still keep the dog on the same side as the obstacle and take a step toward the desired obstacle.
  • Next, you put the dog on the other side and repeat the steps.
  • Now, you start moving the two obstacles closer to each other and repeat the steps.

For the dog, you’ve built the skill of differentiating the two obstacles by name very slowly and in a logical step-wise fashion; the dog’s practice has been gradually getting more challenging, but in small increments.

Now, let’s say that, you–the person–have identified skills you need to improve. You want to work on your timing of blind crosses, or the footwork of your front crosses, or giving cues to your dog earlier. Effective practice dictates that you break it down into steps. By working a sequence without the dog, in slow motion, you can learn how the steps feel without confusing your dog. Once you’re comfortable walking through a sequence–or doing that front cross–slowly, you can move a little faster, then faster the third try–and then add your dog. You can look at course maps and visualize how you will move through a course before you run the course; visualization has been proven to improve performance.

When I started thinking about effective practice and how I learn, I started breaking down the long sequences presented in my lesson and only doing seven or eight obstacles at a time. I tried different handling methods and different dog paths at critical points.

Today Daisy had this sequence in the middle of the course:

course mapThe question, of course, was which path to use for Rush when going from 13 to 14 to 15. I walked three different methods of setting Rush’s path* and Daisy timed the three methods. We worked on what to do with that segment, in detail, for about fifteen minutes. Then, we moved to the next segment. The course as a whole had 21 obstacles; it took an hour of analysis and testing to determine the optimal path for Rush. We worked entirely on the single course, without trying to do more.


We analysed it one segment at a time and I refined my handling one segment at a time, deliberately working my way through the course rather than flinging myself at it and biting off more than I could chew (to use a dog analogy). Rush repeated each sequence without getting frustrated–short sequences mean a higher reward rate for him–and paid attention to the differences in handling as well.

As a result of this change in my practice sessions, I have improved my course analysis skills, become more confident with using blind crosses, gained confidence in Rush’s abilities and his understanding of agility, improved Rush’s focus (which was considerable)–and improved my Q rate.

*For the curious, the fastest path, by more than half a second, was a front cross on the landing side of 13 followed by a push (“go round!”) to the backside of 14 on the side closest to the teeter (the top of the diagram above) followed by a push to the a-frame with the verbal contact cue (“climb!”), with dog-on-right over the a-frame. (The next obstacle was a weave-tunnel discrimination, believe it or not, with the entry to the weaves (the desired obstacle) about five feet from the tunnel entrance. Challenging course!)

Tricky bits

The judge for the Top Dog trial was Cheryl Huffman, who has kindly given me permission to post some segments from her courses this weekend. There were two course sections where I struggled with how to handle Rush through the course because of his speed.

The first section is this one, from Saturday’s Standard run:

standard tricky bit

12-13-14-15 is a sprinter’s line–it was all I could do to not get too far behind Rush–and a push to 15. From 15 to 16, not so bad, but then… which way to bring the dog to get to 17? I walked it with a post turn, dog on right, over 16, to a front cross (the red path); I walked it with a push to the far side of 16, rear crossing 16, 180-degree-turn to 17 (the blue path).

standard tricky bit paths

I decided that the push to the outside of 16 (blue) made more sense for Rush. Watching other people handle the turn, most of them chose the red path.

You can see video here. Notice that I had to stop Rush just before 15 so I could catch up to him to handle that bit… That would be why I’m trying to get faster.

I ran into a different problem in the Jumpers course on Sunday. Here the path from 7 to 11 requires that you push the dog along the line of jumps past the off-course tunnel (the exit end of tunnel 11). When I walked the course, I was worried I’d be so far behind that Rush would turn and see the tunnel. I needn’t have worried. He was going so fast that he bounced 7 to 8 and couldn’t have turned to the tunnel if I’d wanted him to (note to Cheryl: I assume we’ll be seeing that course again, this time with the turn to the tunnel the correct path?).

jumpers tricky bit

You can see video of the run here. Note the handling error at jump 2, where I stopped, which called him to me instead of over jump 2. Fortunately, he handles backside jumps well, and using the backside meant he had a nice line to tunnel 3. And no refusals are called in CPE, thankfully.

Handling Excellent FAST–Some Observations

Yesterday at the Evergreen State Shetland Sheepdog Club trial at Argus Ranch, I watched the Excellent FAST competitors from my in-ring seat as a bar setter. I love bar setting in the Excellent ring. I learn so much from watching the best handlers on the course. No one is talking to you, you see every run with no distractions, and you get to feel good about having such a good seat; the Clubs need their volunteers to run the trial.

The judge was Lisa Potts. (She also has a poodle.) I enjoyed running her Open and Novice courses, finding them challenging but not impossible. I liked that she made a point of complimenting really nice runs. Thanks very much to Lisa Potts for her permission to post the course map here. This is the Excellent FAST course:

Excellent FAST 3-1-2008 Lisa Potts

Okay, let me say upfront that several people decided during the walkthrough that they would do the course only for training, because they didn’t want to be forced to layer a jump. I’m going to ignore any runs that were clearly for training when I discuss what people chose to do.

For those of you unfamiliar with FAST, it is a strategy game that requires that the handler accumulate 60 points (in excellent), which includes a required distance challenge (the send) worth twenty points plus the value of the obstacles. The obstacle value is set by the judge as part of the course design. In this course, the send is worth 33 points, so getting a Q would require 27 additional points and a successful send. There is a limited amount of time to complete the course. Every second over is deducted from the points earned before time ends.

If you’re trying to understand how people handled the course, I suggest printing out the course map and tracing the routes described with a color pencil. It really helps to use one color for the handler and one for the dog.

The send was 5-1-7 (jump, jump, weave). Not many competitors completed the send successfully–fewer than five of a class of 30.

The most common opening was a leadout to the entry end of tunnel 4, followed by the dogwalk and a swing to tunnel 3 (9 points to this point, 18 to go). From there most handlers did a cross to put the dog on the left over the A-frame (ten more points, 8 to go), over the two jumps (1, 9, points finished) and around to the send and out over the finish jump.

One handler, who did Q, took the dog over the broad jump with the dog on the right, then front crossed to put the dog on the left for the 1-9 jump sequence and the send. Still short on points, she swung the dog back and took the 8-point jump in the corner, then ran hard for the finish jump, finishing as the horn went off.

Another option, which I saw two people use unsuccessfully, was to lead out from the dog between the dog walk and the finish jump, take the 1-point jump, then the A-frame, then the 1-9 sequence (all with dog on left, total to here 21 points (6 points to go if the send is successful), then the send. From there, they pulled the dog back to the 8-point jump, then over the broad jump and around the one point jumps and the 3-point tunnel (in one case) or out (in the other case). Both of these handlers crossed the line to complete the send, so they didn’t Q, but they did complete the course within time.

Finally, a competitor who was clearly avoiding all contacts chose to do the 4-point tunnel, the two jumps to the 3-point tunnel, the 1-point jump after the tunnel, the broad jump (total to here 15 points, 12 to go), the 1-9 jump sequence (2 to go), the send, the 8-point jump (by stepping in after the weaves were complete, putting the dog on right, and sending the dog from left to right over the jump), and ran hard for the finish jump. I believe this worked for her, although my memory of exactly who Q’d and who didn’t is hazy.