This post was originally published on January 27, 2022. It has been updated by the author, Megan Foster, on January 28, 2023.
A common theme that comes up when discussing obstacle skills and problem-solving is ring sustainability. For example, this question below was asked of me by an online student, particularly about weave pole performance.
Can you elaborate a little more about rate of reinforcement in training and lowering the rate of reinforcement to get to an appropriate point of delayed reinforcement for a trial situation?
While this applies to any obstacle or behavior, I believe that weave poles break down easily because they are both physically AND mentally taxing for many dogs. They are an expensive behavior that doesn't hold enough value for a lot of dogs.
When I'm training, I can be doing a number of types of sessions:
- skill training
In skill training, my focus is on a high rate of reinforcement for *just* the weaves or a portion of the weaves. Rewards are happening frequently and immediately following the behavior (weaves). The behavior-to-reward ratio is 1:1 and my focus is on the precision of behaviors.
In sequencing, my focus is to build chains around the weaves and to provide a high rate of reinforcement for a randomized amount of work. Rewards are happening frequently and happening immediately after ANY obstacle/behavior in the chain/sequence (not just the weaves). The behavior-to-reward ratio is based on effort. So, if I ask the dog to complete a chain of 4 behaviors, I'm going to reward the effort of that chain regardless of the precision.
In testing, whether this is at home or a competition, I build sequences that the effort is reinforced with delayed reinforcement. Rewards are stashed off my person and are not accessible until after I've leashed the dog and we've exited the training space to the reward. This requires the dog to move away from reinforcement, perform a chain of behaviors, and return to the reinforcement before collecting reinforcement.
When is each type of training appropriate?
You need skill training to build fluency for the behavior(s) that you're training and to build value for the obstacle/behavior. When you do not have reinforcement on your body, you are relying on the HISTORY of reinforcement to motivate the dog to perform that obstacle.
You need sequencing so that you can build more than one behavior before the dog expects a reward. When sequencing behaviors together, you need to be using behaviors that already hold value, because the second behavior reinforces the first and the third behavior reinforcers the second, and so on.
You need testing when you're ready to put a bit more pressure on the behaviors to see how they hold up. The major thing here is that you need a trained reinforcement strategy of food (or a toy) stashed off of your body. The dog needs to understand before you ask them to work that this is the contingency and that it's attached to leashing up and leaving the ring. I typically back-chain this process and I start it as soon as I have a leash-on behavior and build on it over time.
When it comes to the behavior of weaves falling apart in competition, take a look at your areas of training: skills, sequencing, and testing. Are you lacking in one or more of those areas?
If you feel like you've built up robust reinforcement history in all of these areas and still see a lack of motivation to weave in a trial setting, you likely don't have a weave pole issue, but a ring confidence issue that's presenting as a weave issue, and you're likely seeing other behaviors degrade in trials, but not to the same degree as weaves.
I talk about analyzing your trial data correctly in Episode 27 of Fostering Excellence in Agility, The Podcast.