We may never truly comprehend how dogs see agility. It is fascinating when we see dogs react the same way to something, despite being trained in completely different ways.
Every now and then, either in trials or in training, we see a situation where several dogs act the same way, often completely against their handlers´ intentions.
In this article we share some experiences which have broadened our horizons, challenged what we thought we knew about how dogs see the world, and taught us little by little to understand our K9 companions better.
Right or wrong
Our first lesson in dogs’ perspective on agility was several years ago in our own training session. We were training with Janita’s male Border Collie Cosmo, and Jaakko’s female Border Collie Zen. In the middle of the training course we had the following sequence (picture 1). In jump number 2 we executed a Tandem Turn, then Cosmo surprised us by going around the first weave pole and entering the weaves from the wrong side. It would have made sense if Cosmo had entered from the other side of the weaves and gone straight to the second gap, but the type of error he made did not make much sense, and was an uncharacteristic error for him. Janita ran the same sequence for the second time, with the same result. Janita started to wonder what on earth had confused Cosmo so much that he made such an uncharacteristic error, not once, but twice in a row.
Sequence where the dogs entered weaves from the wrong side.
There was no problems in entering the weaves after the staircase pillar was covered with a tunnel.
Trust your dog
Janita and Cosmo took a break and Jaakko started training with Zen. Coming in to the weaves, Zen made the exact same error and missed the weave entry. Now it was clear that there was more to it than we could see from our perspective. The dogs were, for some reason, seeing the weave entry differently to normal. We tried to figure out what was going on by going down on the ground to the dogs’ level on the landing zone of jump number 2, and trying to see exactly from the dogs’ point of view how they saw the sequence. About six feet behind the weaves, there was an eight-inch thick staircase pillar, more or less in perfect alignment with the weaves. We were unsure if this staircase could be the reason for our dogs to miss the entry. So to find out if our dogs had interpreted the pillar as the first weave pole, we blocked its view with a straight tunnel between the weave poles and the staircase pillar (picture 2). After this neither dog had any problems entering the weaves correctly several times in a row.
You could only wonder how much confusion we would have caused our dogs, and how they would have reacted, if we had punished and corrected them for doing - from their point of view - exactly what they have been taught to do.
Pictures 3,4,5,6 & 7.
The dog seeks the gap between two first poles.
Sometimes you can clearly see differences in how dogs visualize the entry. Some dogs seek the gap between the first two poles (pictures 3-7), some seek to get the first pole close to their left cheek (pictures 8-12) on entrance and some try to get the second pole close to their right cheek.
Pictures 8,9,10,11 & 12.
The dog seeks to get the first pole close to her left cheek.
One time Janita was giving a seminar to students from across Finland. Many dogs, way too many dogs, missed the weave entry by going around the first pole on the wrong side. She wondered why they had this unusually high percentage of errors. Janita noticed there was an extra support in the weave pole base before the first pole and wondered if that could be the reason (picture 13). The other end didn’t have this extra support, so she turned the weave poles around and this solved the problem. It seemed that many dogs were not making the entry to the weaves by looking at the actual weave poles. Either the different base partially affected their decision-making, or they were entering the weaves looking solely at the base and not the poles. We often don’t know or realize exactly what the dog has associated his behavior on.
The extra base in the beginning of the weaves caused problems in entering.
We have also seen an increase in the percentage making mistakes on entry, when:
- The dog can see the long jump corner marker poles behind the weave entry (picture 14)
- The color of the first weave pole matches the background
- The first poles are in the shade and the rest of the weave poles are in the sun
The corner marker pole of the long jump can fool the dog behind the weaves.
Let me in
One of the most common cases where you can see dogs are conditioned on something else the handler intended, and seeing the situation differently from us, is the blind side tunnel entrance. Instead of seeking the actual entrance of the tunnel, like we would do from our point of view, many dogs evaluate where the entrance is based on sandbags or something similar which holds the tunnel in place. Where the sandbags are not placed right beside the tunnel entrance you often can see the dogs hitting their head on the side of the tunnel which is visible between the sandbag and the actual tunnel entrance.
Inexperienced dogs often struggle to perceive and separate the teeter and the dogwalk from each other. From the dogs’ perspective they can’t see the level part of the dogwalk on a straight approach (pictures 15 & 16), so they often are quite unsure going in to the contact obstacle. We can only guess how the dogs eventually learn to distinguish between the teeter and the dogwalk. Are they paying attention to the base of the teeter; do they learn the dogwalk has crossbars and the teeter doesn’t; do they learn to pay attention to the colour difference in the end part of the teeter; or do they react to the movement of the teeter? Sometimes in trials you can see the dogs getting really cautious going in to the dogwalk if the rising part of the obstacle is wobbly, like they would think it is a teeter, because it moves under their feet.
Pictures 15 & 16.
The dogwalk and teeter looks a lot alike from the dog’s point of view.
Bars or wings
Just as we are unable to always determine how dogs perceive to go correctly in to the weaves, similarly we are sometimes taken by surprise with what the dogs look for when they go in to a jump. Every now and then you see dogs taking off for the double jump from way too far away. Some dogs don’t perceive the double jump as three dimensional, but instead see it as two bars on top of each other. The dogs who perceive the double jump like this correct their jumping arc on a second try.
Some dogs don’t look at the bars at all. Their decision on how to approach and take off the jump is made by looking at the wings. How the dog perceives a double jump can be tested by doing the double jump test with several sets of extra wings. This test was introduced in a previous Clean Run article on the OMD basic training. Dogs which jump based on looking at the wings also often jump over the obstacle with the same jumping arc even if the bar has already fallen on the ground before they enter the jump. We have also observed that often this type of dog takes off from further away the larger or wider the wings are, and assume this is one of the reasons these dogs often have problems clearing the wall consistently.
There is a reason
A few years ago, Janita’s female Border Collie Hitti had always cleared the long jump without any problems. Once, when competing abroad, she knocked the last piece over in three courses. Janita was puzzled as to why Hitti made the same, extremely uncharacteristic mistake, not once, but three times in a short space of time. Back home she practiced difficult angled approaches, turns, approaches with full speed, different handling techniques and different obstacles after the long jump, with a consistent outcome. No matter what she did on the long jump, Hitti cleared it perfectly every time.
Janita forgot the whole episode, until few months later she was giving a seminar in another city in Finland. While setting up the course, she was about to place the corner marker poles as usual, when she realized that the base was constructed so that when it slotted in, the corner marker pole was in the middle of the last part of the spread (picture 17) instead of in the outside corner (picture 18).
The corner marker poles are set in the middle of the first and last part of the long jump.
The corner marker poles are set in the outside corners.
Janita remembered the problems they had in the trials abroad and wondered if the placement of the poles could have been the reason for Hitti’s problems clearing the long jump. Back home, to test if this really could have affected her performance, she placed the corner marker poles in the middle of the first and last piece of the spread to see what would happen. On the first try, the last piece was knocked over. Hitti had learnt to clear the spread by looking at the corner marker poles instead of the pieces of the long jump. Janita was pleased, because now she knew what they needed to work on. So they started to change the placement of the corner marker poles randomly all along the sides of the long jump. At first Hitti became really uncertain, because now she couldn’t evaluate how to clear the long jump in the way that she was used to. She started to jump about one and a half feet further than before, because she had a hard time perceiving the length of the obstacle from the outside edge of the last piece. Nowadays we have often noticed similarly placed corner marker poles in trials abroad. We can help dogs that have difficulties with this issue because now we know what to look for and how to address the problem.
It’s a blur
Inexperienced dogs often have problems clearing the tire. In trials, dogs often face different types of tire to those they are used to training with, and they have difficulties finding the right gap to go through. If the spacing between the edge of the actual tire and the frame is wide (picture 19), the dog can easily make the mistake of going between the tire and the frame from the side, or they might even try to clear the obstacle by trying to jump through the gap above the tire.
Wide space between the tire and the frame.
At the European Open 2013, the opening sequence in the standard agility large category consisted of a jump - tire - tunnel combination where several dogs had problems clearing the first jump. When it was Janita’s turn and she was setting her dog up for the start, because of all the mistakes by previous dogs, she kneeled at her dog’s level to check her perspective of the opening sequence. From the dog’s perspective, the first bar and the lowest part of the tire were perfectly aligned with each other and therefore the bar kind of blended in to the tire in the background, making it challenging to clear the jump (picture 20). Janita moved her dog six feet back from that spot and checked from her point of view that she could now clearly separate the first two obstacles from each other. Her dog cleared the first jump without problems. When several dogs are making the same mistake in the same place, it is more than likely there is a clear reason, from the dogs’ point of view. If we see this happening we should always trust the dogs, pay attention to what they are trying to tell us, use our knowledge to figure out what is going on and why, and try to solve the problem to help them clear the sequence safely.
The first bar and the lowest part of the tire are aligned from the dog’s point of view.
Don’t fool me
Judges have a great responsibility to design safe courses and to understand the dog´s perspective to agility. In one trial, the teeter and the A-frame were aligned with each other about 15 feet apart (picture 21). Over half of the dogs in the large class flew off the teeter without it moving down even an inch. One of the fastest dogs practically landed on the A-frame. Luckily all the dogs escaped the teeter without injuries despite several dangerous looking situations.
The teeter and A-frame aligned from the dog’s point of view.
Sometimes judges may accidently misguide dogs in tunnels. The judge passes or stands further away from the exit of the tunnel. From inside the tunnel, the dog is unable to distinguish who it can see outside, and starts to turn the “wrong way”, towards the judges legs instead of to its handler.
Know what you teach
Handlers are often oblivious to what they have actually taught their dogs. The most typical example is in 2on-2off contacts. Handlers think they have taught their dogs to independently offer the behavior of going to the end of the contact to the 2on-2off position, and releasing with a verbal command. Often when you test what the dog actually has learned to do, you learn that when the handler is asked to run ahead of the dog when the dog enters the dogwalk, and stop half way along the level part before the dog catches up, the dogs most often also stop before the contact zone instead of going down to 2on-2off (picture 22). It is rarer to see the dog continuing and doing a perfect 2on-2off despite what the handler is doing. To begin with it is natural for dogs to react to motion. Most dogs have unintentionally in every training session and trial had reinforcement to base their contact behavior on the handler’s motion instead of ignoring it. Handler stops, dog stops. If the handler decelerates, the dog also decelerates, and if the handler runs with full speed past the contact zone without watching the dog, it follows the handler motion and runs through the contact zone without stopping to perform a 2on-2off. When giving a verbal release the handlers have unintentionally taught their dogs to follow the body cue instead of a verbal. If a handler gives a “wrong” verbal and continues handling by starting to move or executing a handling move the dog releases, despite the hander not giving the dog the correct verbal release the dog is supposed to wait for. These are partly the reason behind the problems at contacts in trials. Dogs have conditioned their contact performances to different cues to those intended by the handlers.
If the dog is conditioned to the handler’s motion, it will stop in the middle of the contact obstacle, if the handler stops.
Live to learn
Over the 20 years I have been involved with agility, I have seen many similar examples and learnt a lot from them. The purpose of these examples is to help handlers start thinking outside the box when they encounter training problems. When you are trying to teach your dog a certain behavior, and it feels like the dog is not learning as expected with the method you are using, try to condition the dog to something else than you originally intended. In the right conditions, with suitable, properly used teaching methods, the dogs learn extremely fast, after five to seven repetitions, for even quite challenging tasks. If you have to use multiple repetitions to teach something and there is no progress, your dog’s perception to the behavior you are trying to teach differs from yours. Different individuals learn differently, and sometimes even with the same method, different dogs condition themselves to completely different things. Your dog is the best judge to tell you if everything is OK with you method and how you are using it.