Q&A: How can gamification help identify problem gamblers?

Mindway AI founder Professor Kim Mouridsen chats to EGR Compliance about the development of the industry’s first self-assessment game for problem gambling and how this might yield a more insightful approach than current thinking

The identification of at-risk and problem gamblers is something of a stone in the shoe for online gambling firms, as however many are identified there is always one which slips through the net who doesn’t receive the help required in time.

Conventional industry wisdom is that the best way of finding at-risk gamblers is to make them question their own play, be it through the implied restriction of deposit, time or withdrawal limits, which encourage gamblers to question their own behavior or via the medium of the self-assessment.

But is it time for a change in strategy? Should gambling treatment be less about the gambling and more about the psychological science of addiction and what sort of test is going to be able to cope with a wide range of gambling addiction types while also giving specific help. Mindway AI have recently begun implementation of a self-assessment game, which uses the values of gamification and the psychological science behind addiction to deliver a very individual solution to a very troubling problem for the industry.

Kim Mouridsen, professor at Denmark’s Aarhus University and founder of Mindway AI, explains to EGR Compliance why the answer to identifying problem gamblers might have a little bit of gamification to it.

EGR Compliance: Why have you chosen to launch a self-assessment game for players?

Kim Mouridsen (KM): We’ve been looking at what the market is currently offering in terms of self-assessment of gambling-related harm. We believe that a self-assessment test is a useful tool for both online and physical casino operators and tests like the current PGSI, which is widely used, is good but it comes with a lot of limitations. Firstly, we cannot be sure that people are responding truthfully to these questions, as quite often people will answer inaccurately to avoid being classified as a problem gambler. Leading on from this there is an implied bias on gambling by the person who is taking the test. There is also lots of research which shows that gamblers can be very uncomfortable with the sorts of questions being asked, leading them to miss certain questions or answer them incorrectly when compared to their own perception. All these issues can skew the results of self-assessment testing at both an operator level and national level.

Some studies have shown that on average about 26% of the questions asked in the current PGSI are misunderstood. Moreover, concepts such as ‘problem’ and ‘lying’ is interpreted differently by different individuals in different social settings and different countries. Finally, the PGSI is known to express only a single dimension of problem gambling, namely severity, but does not suffice as a means of characterisng the many facets of problem gambling behaviour. This is important in recognising individual differences and serve as the basis for better and more efficient intervention.

A test which understands and recognises these nuances is more useful than one which just measures your gambling habits against a set scale, especially when it comes to problem gambling treatment because it allows players to embrace their habits more fully than granular answers.

In any case, self-assessment testing often targets the risk rating of the player rather than addressing the problem. The underlying problem, something we’ve studied in neuroscience, is certain disfunctions in the brain which impair decision making. The PGSI and other common self-assessment tests measure the result of that dysfunction, but they don’t measure the mechanism which drives the dysfunction in the first place.

That sort of reasoning drove us to think about gamification in the first place, thinking that if instead of asking people about the consequences of gambling we could get a better understanding of their gambling habits by measuring their decision making. We would do this in a controlled environment where we could provide good choices and bad choices, manipulating the test in ways to get us hopefully closer to the actual mechanism behind the problem gambling.

Mindway's test comes in the form of a game

Mindway’s test comes in the form of a game

EGR Compliance: What does gamification bring to the table?

KM: In a gamified environment we are basically free to define the testing using neuroscientific principles rather than standard self-assessment guidelines enabling us to get a richer and more useful information out of the testing than just having a few ambiguous questions.

There is a rather rich vein of scientific literature as evidence of this, it’s not something which has only come up recently. Over 400 papers on the subject came out in 2012 alone and by 2017 that number had doubled to 800. Some of them are concerning gambling disorder because certain designed games target a part of the brain which serves as a kind of control mechanism for decision-making under risk. Without getting too technical, there’s a part of the brain that’s called the ventromedial prefrontal cortex and one of the purposes of this is to control impulses, which typically arise from our limbic system, another deep structure in the brain.

The ventromedial prefrontal cortex sits between the limbic system that generates impulses and the memory system. In a healthy individual this part of the brain functions normally, dampening and filtering these impulses that we receive because the memory system compares them to memory and expected outcome. In a problem gambler, this ventromedial prefrontal cortex often doesn’t work properly, hence why a lot of gamblers cannot control their impulse to gamble and make poor decisions.

Players are invited to select 100 cards randomly, with their scores increasing and decreasing based on selections made

Players are invited to select 100 cards randomly, with their scores increasing and decreasing based on selections made

EGR Compliance: How do you risk classify players without using standardised models?

KM: Risk stratification is now something that we can do in detail and something we could not do previously. In the early years of these kinds of tests you simply counted how many times a subject would pick the different parts. This has since been shown not to be the best way because a learning rate for instance is a parameter and will play a role and, in that way, if you simply take one big average that’s not going to give you good information.

What we have today is a range of cognitive models, and from these models we can then obtain a wide range of parameters. These can be related to things like how sensitive to winnings you are, how sensitive you are to losses, what your learning rate is and how good you are at forming a strategy (do you randomly pick or do you select your bets/options specifically to target the best results).

As I said earlier, instead of the one dimensional PGSI where you are one of three risk groups, here we can put a lot more colour on the result, not only to compare you to normative values within a certain range but we can also very accurately highlight the areas where players need to be careful about the choices they are making. In this way, there is an educational aspect to the game where we can educate players in a much better way than just a one-size-fits-all approach, where we give information to gamblers about better understanding gambling to giving specific personalised education to gamblers about their own habits.

EGR Compliance: What do players receive at the end of the game?

KM: We do two things. One is that we provide immediate and very intuitive feedback, which says where are you compared to others like you and overall advice, which is the simple approach. Some players will be interested in simple answers and feedback, while others would like to know from a neuroscientific point of view how they are making decisions and what is good and bad.

We only need a split second to make those complicated computations, so we can give that feedback in the form of a report which can be emailed to the player almost immediately.

The game delivers a report which classifies players into three types

The game delivers a report which classifies players into three types

EGR Compliance: When do you plan to launch the self-assessment game?

KM: We are planning to launch the game within the next month or so. We think that this is something which will appeal to the entire industry as it’s a new approach to an existing issue, that of testing of at-risk gamblers. We’ve had some interest from regulatory authorities I wouldn’t mention any names but a range of government institutions operators and regulators in Europe have already expressed an interest in taking this forward.

Analysis | Mindway AI | Problem gambling | Professor Kim Mouridsen | Self-exclusion | self-regulation | Technology