The Jackpot Mystique: Tool or Trainer Superstition?

A perplexing tool

As I travel around the world lecturing about training, there are certain questions that come up again and again. The end-of-session signal, the keep-going signal, the least-reinforcing stimulus/scenario, and jackpots seem to create the most confusion. Perhaps this is because information about these concepts is not readily available in the scientific literature. These are tools that have come from the practical side of training, and have been promoted by trainers who needed to make certain concepts clearer to their animals. As a result, there are no consistent or accepted definitions of these tools.

This article presents my definition of the term, explores the use of jackpots and the mystique surrounding the tool, and discusses common mistakes with the use of jackpots. Finally, I share some suggestions for sorting through the information about a less–than-clear tool.

The fix-it fish

My first exposure to the use of a jackpot dates back many decades. As a young trainer working with dolphins in a small marine life park in Texas, I used a variety of fish as reinforcers. One of the dolphins’ favorite fish was mackerel, but, for a variety of nutritional reasons, each dolphin was only allowed one of those large fish each day. It was a highly valued reinforcer for the animals, and we reserved it for special breakthroughs in training. When one of us was training a new behavior, we would argue among ourselves to see who would get to use the fish for their sessions. When a behavior had problems meeting criteria, we reserved the fish for when the animal did particularly well meeting all criteria that day. We called this reinforcer the “fix-it” fish; the director of training told us that it was called a jackpot. As a young trainer, I remember marveling at the value of this magic reinforcer.

I remember clearly that the use of that fish both helped correct problem behavior and reinforce specific breakthroughs in new training. But is that memory clouded by the passage of time? Is it a fanciful myth that is perpetuated by our desire as trainers to have that elusive magic reinforcer? I am not certain. I have continued to use jackpots, albeit sparingly, in my training. But as I study and explore the science that underlies training, I keep facing the question: Is the jackpot a unique tool, or is it just the use of good, well-timed reinforcement?

Definition:

Because the jackpot tool does not appear in scientific literature by name, trainers have taken the term jackpot—from its use in casinos, lotteries, and other games—and created their own unique definition. It could be argued that any definition of use of a jackpot as a behavior tool could be considered correct, since there is no resource to turn to for a consistent definition. For the purpose of this article, I would like to propose a definition that is a combination of definitions offered by Karen Pryor (1984; 2006), the zoological community (AZA, 2017), and my own discussions with and observations of professional trainers who seem to use the tool well.

A jackpot is an unexpected, high-value reinforcer used sparingly and contingently to reinforce a significant breakthrough in training.

Application challenges

Even if we can agree on a definition of a jackpot, that does not guarantee consistent application of the tool. In my experience I have seen several ways that the tool is used in detrimental or less–than-helpful ways.

The biggest challenges include:

1. Drawn-out delivery – When they are delivering reinforcement, I have seen trainers give the animal a dozen treats as a jackpot. But, instead of offering the treats all at once, the treats are delivered one at a time. By the time the last treat is offered, several seconds have passed and it is doubtful that the later reinforcers are associated with the excellent behavior that the trainer intended to strengthen. The later reinforcers certainly reinforce the animal for good stationing or paying attention to the trainer. These results are not bad, but they are not the responses the trainer was intending to reinforce.
2. Sloppy delivery – Sometimes trainers toss a large handful of treats on the ground; as the treats scatter across the floor, the animal goes into a panic looking for all of the treats. Unless the animal has been trained to receive and enjoy reinforcement delivered in this manner, that type of delivery may be aversive. This is particularly true if there are multiple animals in the room and the form of delivery creates competition for the food.
3. Used too frequently – One of the supposed benefits of a jackpot is associated with the rarity of its use. Used too often, a jackpot creates an expectation on the part of the learner that large or high-value reinforcers for a particular behavior are common and, thus, expected. This assumption can cause lower-value reinforcers to be perceived as aversive. I am not opposed to the use of consistent high-value reinforcers for a critical behavior; I think that can be helpful. But I would not refer to that use as a jackpot.
4. Too novel – There are some who advocate that one of the advantages of jackpots is the novelty. However, I have seen trainers offer jackpots that were so novel that they frightened or confused the animal, and thus were counter-productive.
5. Use of marker –There is a divide in opinions among those who use jackpots in training about the timing of its use. Some people deliver the jackpot immediately following the marker, which is the way I have used jackpots traditionally. Others claim that the effectiveness of the jackpot is dependent on it being delivered in place of the marker, a technique suggested by Karen Pryor (2006) and a technique that I have used in certain situations. This is clearly an area that needs further exploration.

Studying the jackpot

So far, the efforts to study the jackpot under controlled conditions have not proven very enlightening. One study that was conducted did not demonstrate that the jackpot is an effective tool (Muir & Rosales-Ruiz 2009). However, that study did not use the definition or application of the jackpot that I have seen be most successful; the jackpot in the study referenced above was not used contingently on a significant breakthrough.

An area where there is a great deal of research is the area of conjugate reinforcement (Rapp 2008). Conjugate reinforcement is the idea that exceptional effort or increased intensity of a behavior is reinforced with higher-value reinforcers than behavior at a lower intensity or level of strength. A process that has been identified and studied, conjugate reinforcement is the reason trainers find success in the use of high-value reinforcers.

However, studies have also indicated that variety in reinforcement value can create variability rather than reliability in learner responses. For behaviors that require greater intensity, such as pulling a cart, or increased recall speed, a high-value reinforcer seems to be effective in creating variability that allows the trainer to select and reinforce the desired increase in intensity. But, if the variability impacts the precision of a behavior, the high-value reinforcer has been more of a distraction, and perhaps it has even been detrimental. It is this anomaly that creates doubt that a jackpot truly has the desired impact on learning.

Until trainers begin collecting data and keeping good records on their uses of a jackpot, including a consistent measure of the way the jackpot affects behavior, any claims we make about its effectiveness are anecdotal. While I still use jackpots in my training, I wonder if I am simply applying a high-value reinforcer on some type of intermittent schedule. That question, combined with the fact that I change and adjust my application of reinforcement continually based on the learner’s needs, means that it is difficult to be certain that any major improvement in the animal’s performance is due to the jackpot alone or is due to a combination of factors, including the already-proven impact of conjugate reinforcers.

For example, I have often claimed that the excellent reliability of my animals’ responses to the recall signal are due to my occasional use of jackpots for exceptional speed. This could be true. But there are so many other aspects to the training that are probably contributing to the success of the recall. The primary reason for their success is likely due to my use of high-value reinforcers for all well-executed recall responses and a gradual increase in criteria for speed. I do give surprise and unexpected larger-value reinforcers when my animal shows exceptional response times, but is that just a superstitious behavior on my part? I cannot show data that indicates that an animal trained without those jackpots would be less reliable or less fast, but, because of the importance of the recall and the success I have seen, I am hesitant to change my approach to training that behavior.

To use or not to use?

Based on my personal experiences, my own use of jackpots has never had a detrimental impact on my training. However, its benefits are still only conjecture, ideas based on successful results, without supporting data to back it up. I am not going to throw the jackpot out of the tool box, but I am going to take great care not to overstate its effectiveness—or even its existence as a real tool. For now, I will simply state that I am using high-value reinforcement contingently, and not give that practice a special name.

I hope that we can all begin to find a shared definition of the concept of a jackpot. I plan to look for ways to collect, measure, and interpret jackpot data, and I encourage others to do the same. Only then can we truly alleviate the confusion that surrounds this elusive tool.

Happy Training,

Ken

 

AZA (2017). AZA animal training terms and descriptions. From AZA Professional Development Courses. www.aza.org

Muir, K & Rosales-Ruiz, J. (2009). The effects of jackpots on frequency of response and choice. Presented at Association for Behavior Analysis Conference.

Pryor, K. (1984). Don’t Shoot the Dog.

Pryor, K. (2006). Jackpots: Hitting it Big. In: Letters from Karen. https://www.clickertraining.com/node/825

Rapp, J. (2008). Conjugate reinforcement: a brief overview and suggestions for applications to the assessment of automatically reinforced behavior. In Behavioral Interventions, Wiley InterScience 23: 113-136.