Elliot Ludvig, Philip Newall and Lukasz Walasek – Written evidence (GAM0089)

 

Gambling regulation as a strategic game

 

Who we are

  1. We are three psychologists at the University of Warwick, whose work is united by a shared interest in decision making in gambling. We are expressing our personal views about taking a strategic, long-term approach to gambling regulation. These comments are most relevant to the Committee’s third question, “Is gambling well regulated, including the licensing regime for both on- and off-shore operations?”.

 

Summary of recommendations

 

  1. Based on the research detailed below, we recommend the following policies:

 

Background on strategic games

  1. Effective gambling regulation, we contend, will come about from viewing gambling policy as a strategic game. This is best illustrated by example. In poker, a novice player might make a big bet when holding an unusually strong hand. A more experienced poker player understands that any individual action must fit into an overall strategy that makes sense. Just betting big with strong hands telegraphs to observant opponent players both 1) when the player has a strong hand, and 2) when the player has a weak hand (since they did not bet big). The result is catastrophic, and occurs because the novice player did not anticipate how their strategy could be taken advantage of. This is why experienced poker players perform every action with a range of hands, and bluffing is perhaps the best example of this concept (Chen & Ankenman, 2006). We believe that gambling regulation must similarly avoid getting distracted by details around the current status quo, and focus on how to ensure the best deal for consumers in the future given the anticipated response to any regulation from gambling industry actors. 
  2. Often the gambling regulation discussion gets focused on individual products, such as fixed-odds betting terminals (FOBTs). The £2 staking limit on FOBTs is a welcome policy change for us. The policy change is not strategic, however, since the industry has an incentive to create new products which leverage similar psychological mechanisms as FOBTs, but which are sufficiently different enough to not be defined as a FOBT.
  3. Expert poker players prefer to make things as hard as possible for their opponents. Any poker situation has an “optimal strategy,” the details of which expert players and computer scientists frequently agree on (Newall, 2018). Expert poker players never want recreational players to learn the correct strategy in any given situation, which is expressed colloquially as, “don’t teach the fish.” Similarly, expert poker players will never reveal information unnecessarily; they will never reveal their private cards unless forced to by the rules of the game.
  4. Gambling regulators must understand this and change the rules of the game to force the gambling industry to finally start providing consumers with better information and to share their private information with other stakeholders. The gambling industry rarely shares data with researchers (Cassidy, Loussouarn, & Pisac, 2013), and this is just one other demonstration of the industry’s gambling expertise. This last point has been made by a number of other gambling researchers, going back many years. So, in the rest of this submission, we will detail research underlying our three unique recommendations for how to approach gambling regulation as a strategic game.

 

Area #1: Warning labels for online casinos

 

  1. Online casino games contributed £3.0 billion of the remote gambling sector’s £5.6 billion gross gambling yield in 2017/2018. Currently, the Gambling Commission requires online operators to include “information that may reasonably be expected to enable the customer to make an informed decision about his or her chances of winning” (Gambling Commission 2017, p.12).
  2. This information is almost always displayed as what is known as the “return-to-player” percentage, e.g., “This game has an average percentage payout of 90%.” A return-to-player of 90% means that for every £100 bet on average £90 will be paid back out as prizes. If this statement sounds confusing, then don’t worry -- you’re not alone. One study showed that a group of regular gamblers were no better than random chance at determining the correct meaning of this statement (Collins, Green, d'Ardenne, Wardle, & Williams, 2014).
  3. However, Gambling Commission regulations state that this information could also be displayed as what is known as the “house-edge” percentage, e.g., “This game keeps 10% of all money bet on average.” Gamblers will on average lose £10 for every £100 bet on the game. A return-to-player of 90% and a house-edge of 10% are therefore factually equivalent, but psychologically they are quite different. An identical glass might be described as either “half-full” or “half-empty”, with each equivalent description yielding quite different feelings. This common-sense observation would be called a “framing effect” by psychologists (Tversky & Kahneman, 1981).
  4. Research that we have conducted shows that gamblers think they will have a better chance of winning when given return-to-player information than when given equivalent house-edge information (Newall, Walasek, & Ludvig, 2019). This framing effect occurred across gamblers and across a range of different payout rates. Furthermore, these perceptions were backed up by differences in factual understanding. Fully 66.5% of gamblers correctly understood the house-edge statement correctly, compared to only 45.6% of gamblers given the return-to-player statement.
  5. We surveyed 363 online roulette games across 26 major operators: we found 357 games with return-to-player statements and none with house-edge statements. Return-to-player statements were further hidden in dense blocks of text on obscure help screens, with 95.5% of statements using the smallest text size on the screen, and 99.7% using the lowest level of text boldness.
  6. Current Gambling Commission regulations state that, “information that may reasonably be expected to enable the customer to make an informed decision about his or her chances of winning” (Gambling Commission 2017, p.12). However, this regulation has allowed the industry strategic flexibility, in terms of how to display this information. This strategic flexibility has been fully exploited, by an industry-wide strategy of only displaying hidden and confusing return-to-player statements.
  7. Our recommendation here is to remove the industry’s strategic flexibility, by requiring prominent house-edge information to be displayed for all online casino games.

 

Area #2: Warning labels for online sports betting

 

  1. Sports betting is another major contributor to online gambling. The prominent marketing of sports betting on TV might be one of the major contributors to rising public concern about gambling (Newall, Moodie et al., 2019). The gambling industry has a strategic interest in deflecting this public concern, and in 2015 the Senet Group started a prominent awareness campaign, called “when the fun stops, stop.” However, an experimental study that we conducted (the first testing this warning label’s effect on behavior) showed that including this warning label did not help online gamblers to gamble less (Newall, Walasek, Singmann, & Ludvig, 2019). Perhaps this is unsurprising to many readers. Like with online casino warning labels, the industry has a strategic interest in minimizing the effectiveness of any warning label.
  2. By comparison, warning labels for food and alcohol provide consumers with factual information that can be used to compare rival products. One research stream we have been involved in attempts to understand the variation in the house-edge in football betting -- where the house-edge is gambling’s equivalent to the calorie or alcohol unit.
  3. One potentially attention-grabbing result is that some football bets have a house-edge more than fifty times higher than other football bets (Hassanniakalager & Newall, 2019). Furthermore, this research shows that, on average, football bets with high odds (e.g., 10-to-1) have higher house-edges than football bets with lower odds (e.g., 2-to-1). The large occasional wins on bets with high odds are not sufficient to compensate for the on-average long string of losses required to find a winner.
  4. This finding is especially relevant to football gambling advertising, as our research shows that it is predominantly bets with high odds and high house-margins that dominate advertising on TV and in bookmakers’ shop windows (Newall, 2015; Newall, 2017). Furthermore, we have evidence suggesting that the bets advertised on TV have become riskier over time (Newall, Thobhani, Walasek, & Meyer, 2019).
  5. Online sports betting again demonstrates two areas where the gambling industry has exploited their strategic flexibility. Industry players have introduced an ineffective warning label (Newall, Walasek, Singmann, & Ludvig, 2019), and have all advertised bets which a more informative warning label would reveal to be unattractive.
  6. Our recommendation is to again remove strategic flexibility from the gambling industry, and to require independent testing for warning label effectiveness. Our first suggestions for potential labels to consider are prominent house-edge warning labels for football betting, and qualitative reminders of the risks of bets with high odds.

 

Area #3: Product innovation in online gambling

 

  1. Lastly, innovation of new gambling products is a constant feature of online gambling. At present, gambling operators have no incentive to innovate on better value products. This is because it is not clear to recreational gamblers when a better value gambling product reduces the “price” of gambling (the house-edge). At present, better value products would only attract skilled gamblers, who can calculate statistical edges for themselves, and whom the gambling industry does not make consistent profits from. Instead, new gambling products can make money for operators by appealing to the cognitive biases that are especially prevalent amongst problem gamblers (Walker, 1992). An appeal to cognitive biases is a common feature of dangerous gambling products, as is clear for example from Schüll’s (2014) work on electronic gambling machines in Las Vegas.
  2. Request-a-bet products are one such new online sports betting product, which allow gamblers to request odds on custom bets via the social network Twitter. This appeals to problem gamblers’ “illusion of control”, the belief that personal control can help gamblers beat the odds (Goodie & Fortune, 2013). But quoted odds on request-a-bets are extremely unattractive, and offer gamblers essentially no hope of winning in the long term (Newall, Walasek, Vázquez Kiesel, Ludvig, & Meyer, 2019). The first proper request-a-bet product was started in August 2017; by summer 2018 at least five other operators had mimicked this product and were advertising their own request-a-bet product on TV. Furthermore, our research shows that sports betting products that allow gamblers to customize their own bets are especially attractive to problem gamblers. In one recent survey, 16.0% of participants who had placed at least one custom bet were problem gamblers, compared to 6.7% who had never placed a custom bet (Newall, Cassidy, Walasek, Ludvig & Meyer, 2019).
  3. Our final recommendation is to reduce the industry’ strategic flexibility to innovate new online gambling products. At present, if one operator designs a new gambling product which successfully exploits problem gamblers’ biases, then this product can be mimicked by rival operators. Mimicry means that no operator has any incentive to expose this exploitation to either regulatory or public attention (Heidhues, Kőszegi, & Murooka, 2016). By contrast, during the debate leading up to the stake limit on FOBTs, only Betfair/Paddy Power, which does not have a large UK retail gambling operation and therefore few FOBTs, publicly backed the stake limit (Ahmed, 2017). If gambling innovation is to be permitted, any new product should not be allowed to spread across operators before other stakeholders have been able to check this product’s effects on gambling harm.

References

Ahmed, M. (2017). Paddy Power Betfair backs curbs on fixed-odds betting machines. Retrieved from: https://www.ft.com/content/e785b6a4-a350-11e7-9e4f-7f5e6a7c98a2

Cassidy, R., Loussouarn, C., & Pisac, A. (2013). Fair game: Producing gambling research - the goldsmiths report. London: Goldsmiths, University of London.

Chen, B., & Ankenman, J. (2006). The mathematics of poker. Pittsburgh, Pennsylvania: ConJelCo LLC.

Collins, D., Green, S., d'Ardenne, J., Wardle, H., & Williams, S. (2014). Understanding of return to player messages: Findings from user testing. London: NatCen Social Research.

Gambling Commission. (2017). Remote gambling and software technical standards. Retrieved from https://www.gamblingcommission.gov.uk/pdf/Remote-gambling-and-software-technical-standards.pdf

Goodie, A. S., & Fortune, E. E. (2013). Measuring cognitive distortions in pathological gambling: Review and meta-analyses. Psychology of Addictive Behaviors, 27(3), 730-743.

Hassanniakalager, A., & Newall, P. W. S. (2019). A machine learning perspective on responsible gambling. Behavioural Public Policy. doi: 10.1017/bpp.2019.9

Heidhues, P., Kőszegi, B., & Murooka, T. (2016). Inferior products and profitable deception. The Review of Economic Studies, 84(1), 323-356.

Newall, P. W. S. (2015). How bookies make your money. Judgment and Decision Making, 10(3), 225-231.

Newall, P. W. S. (2017). Behavioral complexity of British gambling advertising. Addiction Research & Theory, 25(6), 505-511. doi:10.1080/16066359.2017.1287901

Newall, P. W. S. (2018). Commentary: Heads-up limit hold’em poker is solved. Frontiers in Psychology, 210(9) doi:10.3389/fpsyg.2018.00210

Newall, P. W. S., Cassidy, R., Walasek, L., Ludvig, E. A., & Meyer, C. (2019, In preparation). Who uses custom sports betting products?

Newall, P. W. S., Moodie, C., Reith, G., Stead, M., Critchlow, N., Morgan, A., & Dobbie, F. (2019). Gambling marketing from 2014 to 2018: A literature review. Current Addiction Reports, 6(2), 49-56. doi:10.1007/s40429-019-00239-1

Newall, P. W. S., Thobhani, A., Walasek, L., & Meyer, C. (2019). Live-odds gambling advertising and consumer protection. PLOS One, 14(6), e0216876. doi: 10.1371/journal.pone.0216876

Newall, P. W. S., Walasek, L, & Ludvig, E. A. (2019, Under review). Equivalent gambling warning labels are perceived differently.

Newall, P. W. S., Walasek, L, Singmann, H., & Ludvig, E. A. (2019, Under review). Testing a gambling warning label’s effect on behavior.

Newall, P. W. S., Walasek, L., Vázquez Kiesel, R., Ludvig, E. A., & Meyer, C. (2019, In preperation). Betting on intuitive longshots.

Schüll, N. D. (2014). Addiction by design: Machine gambling in Las Vegas. Princeton University Press.

Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453-458.

Walker, M. B. (1992). The psychology of gambling. New York: Pergamon Press.

 

20 September 2019