The agility course, that is.
It’s common practice in my area and many others to not allow an agility dog to continue running the course after certain mistakes are made, the most common of these mistakes being a missed contact. Because it is so common, and because I feel it is typically being done poorly/incorrectly (and because I actually did it for the first time EVER recently!) I decided some clarity was in order. These are my thoughts on this technique, and if you have other thoughts or opinions, please feel free to leave them in the comment section.
First, let’s be clear about what removal from the course is; it’s punishment. Negative punishment, to be precise, because you are removing something the dog really loves; agility. (Heads up, if your dog isn’t wild about agility, you shouldn’t even be considering this technique). There’s likely some positive punishment mixed in there as well, because the “event” of being taken from the course is likely aversive (no matter how nice you are about it). So, given that this classifies as punishment, that means we know a few things to be true:
- We must be timely with punishment for it to work. That means you must remove the dog from the course immediately after the offense occurs–not three jumps later when the initial offense catches up to you and you flunk. I also recommend having a punishment marker signal, like a clicker signifying a reward is on its way, the punishment marker signal (which is truly different from an NRM, a no-reward-marker) signifies that the punishment is on its way. The dog must know the split second she went wrong, or you won’t get very far with the punishment.
- We must be completely consistent. Once we decide we will pull a dog from the course for a certain behavior (or lack thereof), we must do so every time, even if we are still “clean” at that point.
- The dog must have a strong history or reinforcement for the correct behavior in a variety of environments. I call this not asking the dog to fix something they don’t have the tools to fix. On this same point, we should have massively proofed this behavior, making it much more difficult in practice than it will ever be in the ring (like placing a raw turkey neck just out of reach at the bottom of the dogwalk, and only allowing your dog to eat the turkey neck if she sticks her two on two off).
- Preferably, our dogs would have an immediate opportunity for success after the failure that just occurred. In a perfect world, you could pull your dog from the course, get back in line, and try again in one or two dogs (albeit forfeiting your Q). We don’t live in a perfect world, so a better way of handling this is to set it up in training so that your dog is likely to understand the punishment in the competition setting better. Meaning, if my dog blows her dogwalk contact in class, I will pull her from the course, let someone else have their turn, and then try her again, following up the newly improved DW contact with massive positive reinforcement. Now if I pull my dog from the course in competition for the same behavior she should have better carry over.
- Know that if a punishment-based route is going to work, you will only have to dole out the punishment ONCE or TWICE maximum. If you are pulling a dog from the course (or giving out any other kind of punishment for that matter–like collar pops in training or time-outs for bad behavior at home) on a semi-regular or frequent basis, your dog is not learning (and, in my humble opinion, you are being abusive). If your dog has the tools to fix the problem, he will do so after one punishment.
When trying to change behavior, it is important to focus on one thing at a time. This holds true in all areas of learning, be it shaping a dog to go through two poles to begin weave pole training, helping an aggressive dog to make non-violent choices, or indicating to a dog what is not acceptable in the ring. If you routinely pull your dog from the course for a bullet list of naughty behaviors, it’s time to prioritize. Let’s say I’ve decided to pull my dog from course if she blows her startline, doesn’t hold her contacts, or misses a weave pole entry. I’ve now set my dog up for five potential punishments on any given standard course. How fair is that? Instead, go into the weekend knowing which problem you want to focus on; if the punishment of leaving the ring is appropriate you won’t need to do it next weekend, so just pick one and move on.
A lot of you might be thinking that you feel a slight improvement in your dog when you pull him from the course, but that he always backslides. That indicates the punishment is effective but the reinforcement in training is not. Punishment alone will not uphold behavior–it suppresses behavior. Reinforcement is what is needed to uphold what you want to actually see in the ring.
When I pulled Idgie from the ring a few weeks ago it was because she ran her dogwalk when I asked her to stop. I have been proofing this behavior, I believe her to know what I am asking, and she made zero attempt at stopping. I gave her my punishment marker, thanked the judge and left the ring. She appeared disappointed and slightly confused, but we still went for our post-run walk, she still got her toy, and all was well. Our next trial was a USDAA trial where she did five dogwalks total in three days, all were excellent and met my criteria perfectly. I am not naive enough to think that this isolated incident of pulling her from the ring made that happen on its own. Did it contribute? Probably. But what is more important is that every single time she does her DW correctly in practice I reward it, and I proof this performance frequently. THAT is what will uphold the behavior when I need it.
I certainly hope this clears a few things up, and I really hope I save a few dogs from the headache of confusing punishments.
Idgie with her highly-valued flying squirrel toy, that I use to both proof and reward her dogwalk on a regular basis.