We talk a lot about setting our dogs up to succeed, but what does that actually mean? When we set criteria (what exactly must the dog do to be paid?) and parameters (level of distraction, distance from the dog, and duration: how long the dog must do the behavior) and the dog meets it, we have a few options. We can just pay, we can praise and pay or we can mark and pay. If we are only paying, we need to make sure we “feed for position” (we should do this in all cases, actually)- that the dog has not moved into, say, a down from a sit, if sit was the behavior we asked for. Praise is generally less formal than clicking or marking and of the “good boy!” variety. If we are marking a behavior with either a clicker or a verbal marker, like “YES!”, it’s important that we keep our end of the deal and pay each and every time we click or mark. The precision of clicker/marker training is part of what makes it so powerful. You can read more about distractions, distance and duration here.
In addition to marking the precise moment a dog has done the correct behavior, marker training has some pretty awesome side effects: it buys us time to get to the primary reinforcer (usually food, though play, toys and affection can also be primary reinforcers) and we get some classical conditioning in there as well, as the dog develops a nice positive conditioned emotional response to the clicker, or as my partner-in-crime, Kelly calls it, “Pavlovian Perkiness”- the click signifies something good. See this excellent post from Life As A Human.
This is important stuff if we want our marks or clicks to continue to have meaning to the dog. We don’t want to weaken the power of the marker. So, if we want to build a chain of behaviors and we have used a clicker/marker to build each one, how do we get rid of it? We adjust our criteria and make sure that the dog can do what we are asking from step to step.
These two short videos are examples of what I am talking about. In the first video, I am marking (YES!) each behavior Hazel performs. I am doing this because I have changed a parameter: distractions. We are outside, working on a chain and had never done that before. It’s a shorter chain of 3 behaviors because I could have easily lost her attention to the kid next door who was playing paintball (very noisily). I wanted her to be successful, so my criteria was: “Hazel will perform 3 behaviors in a row and be marked and paid between each.”
The second video is inside- less distractions in an area we often train. In this video, she does a chain of 5 behaviors, with no marking or paying in between. Don’t ask me why I am using hand signals, she doesn’t need them, so I chalk it up to trainer error. In this video, the criteria is: “Hazel will perform a chain of 5 behaviors and be marked and paid upon successful completion of the chain.” I could have praised her along the way, but we were moving fast, so I chose not to. It would have been not so great to mark with my “YES!” and not pay her. I don’t want to take away from the value of the marker, which means that the primary reinforcer is on it’s way. It would be breaking the contract. The contract that says “You just did something good and have earned something for it.” That contract needs to be protected. We don’t want to weaken the predictive value of the marker.
Marking behaviors is a powerful means of communication. We must always make sure that we are communicating effectively and consistently. We do this in our contract: we mark and we pay. If we aren’t paying, we don’t mark- we just keep going or we use informal praise. Dogs are always making associations. The association they make early on in training with a clicker or marker word is extremely valuable and protecting it along the way actually helps strengthen learning because of the very thing it communicates: You have done something correctly and will be paid for it. Mark the Dog. And then pay him. Every. Single. Time.