Training Results are Not Always Linear

Training Results are Not Always Linear

When we talk about progress in training, it is very rarely, if ever, strictly linear.  This is especially true when we talk about behavior modification.  There will always be forward and backwards steps in the process.  Some days will look incredibly promising while others will be less encouraging.  It’s okay and normal and should be expected to some extent during our training.  I do think it’s unrealistic to expect perfect progression as we train, I think this ideal of linear progression causes quite a bit of distress to handlers.

Most people, when they start out training, expect the motion to always be advancing and always be forward moving but learning (in general, not just with canine friends) is a process of forward and backwards movement.  There are successes and failures and that’s part of learning–there are days where, for whatever reason, new skills are easier to do than on other days.  I think of it like a child learning to do a cartwheel.  There are lots of failures at the beginning but the more they practice, the more their movements start to resemble a cartwheel.  Then, one magical day they do the perfect cartwheel–their legs are straight and the land it perfectly on their feet.  Just because they did it perfectly once does not mean that from that point on they will ALWAYS do perfect cartwheels.  It’s more likely that after that one perfect cartwheel comes a bunch of less than perfect cartwheels.  As they keep practicing, they will have more good cartwheels and fewer bad cartwheels. Along with the increase of good cartwheels, the “bad” cartwheels will be closer and closer to good cartwheels–so a “bad” cartwheel at the beginning may looks like a glorified tuck and roll, while a “bad” cartwheel after lots of practice may simply have bent knees (which was better than the “good” cartwheel at the beginning).

scatter plotWhat people need to hold on to is the overal “average” or trend of life moving in a positive direction.  I think of it like a mathematical “scatter plot.” We are looking for fewer “bad” days as the training moves along and the “bad” days should be less “bad”as training moves forward (so a “bad” day after lots of training is probably significantly better than a “good” day at the start).  So the general trend is in a positive direction even though each day is variable and may be a good day or a bad day.

If im working with a reactive dog, for example, I’m not looking for zero reactions right off the bat.  What I am looking for is fewer reactions and for the reactions to reduce in intensity OR for the trigger to become less like a hair trigger.  Perhaps at the beginning of training the dog would have 3 or 4 reactions a day and each resulted in a total and complete melt down–if he saw any dog, walking (in any direction) within 150ft, it took him 15 minutes of being out of sight of the other dog to settle.  With some training, perhaps the dog is generally only reacting 2 times a day, when he has a reaction it takes him about 5 minutes to come back down, and now only reacts if the dog is within 10ft–since the threshold is cut down significantly, there are moments now when he can see a dog in the distance and not react at all!  The dog will have some days where he’s more reactive than others but the averages should be moving in a positive direction.

There are far too many people who expect training to be linear–always moving forward, never having “bad days” and any “good day” that was above the normal linear path would be the new expectation of behavior.  Training doesn’t work like that and nor does behavior modification.  There will always be ups and downs, forward steps and backwards steps–what’s important is that the trend is in the positive direction.

It’s easy to become discouraged if you are expecting a linear path in training, so it’s important to try to remember that training isn’t a ever going to be a straight line of improvement.  Even the best trainers in the world have good days and bad days when they are working–it’s normal!  Look at the bigger picture when evaluating your work… is the typical behavior you have now better than what you started with? That’s the important piece to hold on to!

1 Comment
  1. I have a pair of clients whose border collie puppy is one of the best “jumping-on-people” executioners I’ve ever worked with. She could have aptly been named Tigger. She is so exuberant and loves people and has been inadvertently reinforced (I’m guilty of the same thing, it’s hard to never interact with a dog flying for your nose or trying to leap up your back!) At her first session, we began working on teaching her an alternate behavior- we started with sit because she has a nice auto-sit already. At her second session, I arrived to a dog whose impulse control was just as poor as ever and she couldn’t help but spend the first few minutes in the air, but who CLEARLY understood what behavior she was supposed to offer instead. She was pressing her butt down as hard as she could when she came to her senses, but when the excitement got to be too much, she launched herself again. Her family was disappointed, afraid that I would be upset with their “lack of progress.” I told them that their dog has LEARNED what to do, she simply needs more practice, and increasing levels of distraction and excitement, so she can really get the behavior down. Though her jumping rate was still nearly as bad as before, maybe 25% lessened, the fact that she now understands that she has an alternative behavior to offer in order to obtain reinforcement IS progress!

Latest Posts

Contact Us


"Like" us on Facebook


Professional Organizations

Professional Organizations



Pet Professional Guild


Kennel Pro Insurance


Certified Professional Dog Trainer

Certified Prof. Dog Trainer

CLASS Evaluator

Canine Life And Social Skills


ABC Mentor Trainer

Canine Good Citizen