Agility is changing and now nearly every agility organization offers or allows some type of training in the ring (or is at least going to test it out). How can you use these allowances to your advantage?
One thing to keep in mind is that you were always training in the ring. Your dog was learning something, even before we could bring in toys, or reattempt the entire contact obstacle.
Now that we have the ability to maintain better criteria, we will, right? Well. Maybe. We always have had the option to maintain criteria, just not one that allowed us to keep running the course afterwards.
Before you use fix and continue (without a primary reinforcer on you), take a quick study of how you respond to mistakes in training, and how you have been responding to mistakes in competition. Then, study how your dog responds to that feedback from you.
Is your dog happy to reattempt obstacles or sequences? Is your dog frustrated by having to reattempt something? Is your dog demotivated when asked to do an obstacle again?
How can you help prepare your dog for these situations?
- Split your training into two types of sessions: a) training and b) working
- Train your re-do procedures as behaviors that are on cue and act as conditioned reinforcers
Training vs Working
Most of my time spent with my dogs is what I call training. My reinforcers are present, the rate of reinforcement is high, and I have a specific objective for the time spent. When I plan a training session, I plan out the details. What the conditions are for the training (antecedent arrangement) what criteria I am looking for (behavior), and the reinforcement strategy I plan to use or, what I will do if I fail to see the criteria I set (consequence).
When I want to sequence multiple things together, I transition into working. This is my competition prep work. I set up a course of an appropriate length and complexity for their skill level and my reinforcers are not on me. They are on a table near by, similar to a competition. I still plan out the details, just like a training session, focusing on the ABCs and making sure I know the path to reinforcement is clear to my dog.
Train Your Resets
When I'm training, resetting my dog is usually pretty easy – cookies and toys are present! If I'm training an obstacle, I usually set up a two cookie contingency: cookie for the correct behavior, cookie for the reset. So, if my dog doesn't offer the correct behavior, I can skip the cookie for the correct behavior and just cue the reset & reward that, and continue training. This keeps your feedback look clean and avoids frustration or demotivation in dogs.
When I'm training a handling skill, I use a similar approach, if the dog is truly responding incorrectly to my handling cues, I use a two cookie/toy contingency like above and will eliminate the first reward for an incorrect response. However, most of the time, the dog has responded correctly to an incorrect handling cue, and therefore I deliver reinforcement as planned.
When I'm working, resetting requires trained behaviors because I do not have cookies or toys present. So, if my dog misses a contact or pops out of the weaves, or drops a bar during a working session my options are:
- ignore it and keep going
- ask for another stationary behavior before continuing
- end the run, removing opportunity for reinforcement
- fix and continue – go back to the beginning of the obstacle where a mistake has happened and start again
Whichever way you go, understand that the dog should see these consequences before he gets into the trial setting. The dog should have some expectation of what your behavior will be in each context. I do not have a problem using any of the above consequences, so long as I have trained them with food/toys, then trained them without food/toys, and applied them in a working session at home/in class.
Observe how your dog responds to each of these and decide what is best for you and your team. A lot of different factors go into my decisions about responding to mistakes for each of my dogs.
#1 – is usually a no-go for me as it is reinforcing the incorrect behaviour, so I am trying to correct the incorrect by circling and re-doing. The run is done, so if I want I can give my reward cue and leave the ring and head to the reward – dog understands that the last thing they did correctly earned a big reward.
#3 – zoomies, direct disobedience definitely.
#4 – answered in #1.
What is the run isn’t done? What if the dog hits the contact but self releases? Run technically “not done”. Are you acting in the same way?
Technically, the circle also reinforces the incorrect behavior. Because it is a trained cue that has become (hopefully) a conditioned reinforcer. However, it is likely *less* reinforcing than continuing to the next obstacle at full speed.
The goal is that we are getting more of what we want, and less of what we don’t, so as long as you are getting correct criteria MORE often, then you don’t have a problem. If you started to see you were having to circle & re-do more often than not, then we have to go back to the drawing board 🙂
That’s where the mental conflict comes into play for me. If the dog flies the contact – do you ignore it or circle and fix? If you ignore – what is the dog learning. The run is an NQ so why continue and have the dog think that behaviour is OK. If “I” circle and re-do then I have the opportunity to either leave the ring for a “cookie” reward or continue the course. Either way the dog knows that the second time they did the contact correctly.
It’s really up to you to decide, and it’s based on my dog, my dog’s learning history, my goals, and the agility organizations that I compete in. Whatever I choose, to be the most fair, I need to be able to be consistent with and have clear criteria for “if this happens, I will do X” and “if that happens, I will do Y”. For example, if I cue a stopped DW to Shrek, and say he stops about a foot off the board. I am very unlikely to bother “fixing” this – I’m going to continue as if he stopped in the correct place. But, if he didn’t attempt to stop at all, that would be worth cueing a circle and redoing. However, I wouldn’t even have these different contingencies if I didn’t truly believed that Shrek knows the difference in these two behaviors (stopped dw and running dw).
Make it easier on yourself. Your dog isn’t learning anything in one repetition. So if he flies off the contact and you ignore it once, who cares? He’s got 1000 repetitions of good behavior being rewarded that won’t be undone in one repetition. It’s what you do in the NEXT run and in training that is going to have a bigger effect. If my dog makes an error once in a weekend, I’m not worried about it at all. My dog makes the same error twice in a weekend? I’m not asking for the behavior ANYMORE until I get home and train it and build up some reinforcement history.
Does that help? I don’t want there to be conflict in your head! I want you to have a clear plan for your dogs so that their path to reinforcement is clear 🙂
Hi, my dog is new to agility, we have learned the obstacles and now have started to put some together. If we are doing a sequence of obstacles and he misses one or does not stick his 2on 2 off or pops out of the weave poles am I right to just keep going? Or should I repeat the obstacle he missed.
It is absolutely ok to keep going, Yvonne. Take note of anything you or your dog struggled with during the sequence and then work on that separately during another training session. If you imagine your training as a set of scales, you want the plate with reinforcing the correct behavior to be much heavier than reinforcing the wrong behavior!