Predictability vs Unpredictability

I have worked in various zoos and aquariums around the world over the last 11 years what gave me many new experiences along this road. Im always amazed about how trainers/keepers find solutions for problems and how they have been training their animals. One of the topics that pops up in my mind on a regular base is when should we be predictable and when should we be unpredictable towards our animals for higher motivation?


When I started working in Marineland Canada in June 2010 they had a mixed specie show with walruses, sea lions and dolphins what was in my eyes a very cool way to do a show. Very interesting and challenging but there was always one big problem and that was the turn over rate of new people coming in what made introducing the animals a very hard and almost non succesfull job to do. I was still pretty young in this business, I just came from another zoo where I worked for 5 years with sea lions. We were working with a high variety of ways of reinforcing with primary reinforcement and with variable positions. In Canada we had pouches, “boxes connected to a belt” where the reinforcement 9 out of 10 times comes from. As you can imagine we did have some pouch divers there.

Our show was pretty set and even to a point where when we forgot the coming behaviours the animal would exactly know what it had to do. To be honest this has his pros and cons. I mean if you have a turn over rate of trainers coming in and going with not to much training experience it is great that the animals just know what to do but I do have to say im a big fan of variation. Finding motivational objects and scenarios what gets the animals through the steps we set out for them to grow in their behavioural repertoire, that’s what I enjoy the most especially when I can challenge my training skills and see if I can find other ways to train various behaviours. To come back to the sea lions in Canada, the department gave me the goal to get the young animals ready for the show so the oldies could retire. Challenge accepted I thought but I didn’t know how big this challenge would be when they told me this. We had aggressive animals, we had young animals, we had animals who liked playing more then actually focussing training and we had animals who never been together preforming before. Lots of things we had to do to get them into the current show. The show had a very set line for the sea lions and there wasn’t to much variation for the animals. Making them become robots was an easy task but as mentioned before I didn’t really want this. Meanwhile the shows had to keep going with the oldies. The only change I made was reinforce the correct animals and don’t pay attention to the others what means that I had to practise my bridging with working 3 animals at the same time. Quickly enough the animals started to be much faster on my signals. I think this was because they observed the correct animals get reinforced all the time. Now the behaviours stayed the same what was very predictable to the animals but I discovered that between every asked behaviour I could snoop away 2 seconds of adding some sort of variation. What means that this could give me time in the set show what was for me essential for the young animals to stay focussed in the show. So I used these 2 seconds in a way that for example, once I would reinforce and send the next behaviour right away another time I would only bridge and send the next behaviour right away or I would reinforce wait reinforce again and then send and so on. The pace in the show was the same but now I still had animals who were very focussed to me even though they would knew what was going to happen next. That’s where unpredictability came in.

conditioned reinforcer

When training animals we use a bridge. For you guys who do not know what this is, A bridge is a conditioned reinforcer also known as a secondary reinforcer. The bridge pinpoints the moment when you want the reinforcement to happen. This is how we can communicate to the animal when the animal did a great job. A bridge is basically a very predictable strategy because the animals know they did something good. The unpredictable part is what comes after, the reinforcement. This makes the information what comes after the bridge very unpredictable but because its always positive the motivation is very high for the animal and there for the animals will repeat the approximation you asked the animal to do next time again.


At Kolmårdens Djurpark we work with many recalls. We use recalls for the safety of the animals itself but also for the visitors and the keepers.  When we start training a recall we make a very predictable message to the animals. Look if you respond in some sort of way what we want you to do (in our case mostly for coming inside) you will get a big amount of meat (when we talk about carnivores), this is the first step. I mean it doesn’t matter what the time criteria they give us or what ever, they just have to come in. Believe it or not after 3 times they know very well what’s going on if we talk bout the carnivores I mean this is very specie specific especially thinking about their natural behaviour. This is where we step away from being predictable. We start changing the reinforcement with the amount, the size, the kind of meat and the position. The predictable part is that they know they get something but the unpredictable part is they do not have a clue what it will be this time. guest-love_aspen_051By training their mindset they are giving you a higher criteria then asked for just because you are unpredictable in the reaction towards the animals.

I do believe if we know well enough how to play with predictability and unpredictability your motivation will rise and not only that you will get animals who start to think a lot more. They start to value out scenarios, they start to think about their choices and even how or what they can do to receive the bridge in some cases but you know what I do not want to be anthropomorphistic. At the end of the day we never know 100% and therefor we should observe all the time, before within and after our sessions because believe it or not a lot of times we as trainers are figured out even before we figure out the animal due to predictability.

I always explain to the teams at our zoo, if I reinforce you with a carrot every time you do something correct, after 10 times getting a carrot you probably throw the carrot back at me. I can ask you why did you do that? I mean isn’t a carrot part of primary reinforcement? Even primary reinforcement can become boring although we think they want it. If we become predictable with our primary reinforcement you will not see the motivation you are looking for but if we stay unpredictable with our reinforcement the motivation rises.

Plan your reinforcement, your positioning and your sessions.


Peter Giljam

“Thinking Outside the Zoo”



Peter is a passionate Animal Consultant that beside teaching you about Operant Conditioning makes sure you will go home motivated and inspired. Make sure you read his Bio!

1 Comment

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: