It would be useful to know what the distribution of session losses looks like between £1,000 and £5,000 (probably policy relevant); but given that there are some betting shop session losses above £5,000 and no bingo or arcade session losses above that level, it seems plausible that the distribution of losses in this band for the bingo and arcade sectors may be lower than for betting shops.
What we can say is that a small minority of machine sessions in all three licensing categories involve heavy losses – but to describe the risk as “similar” may stretch simile (assuming these prima facieextrapolations are valid).
The Commission’s advice also compares losses on B1 machines in casinos to B2 machines in betting shops, pointing out that losses above £200 are more than twice as likely to occur on the former than on the latter. Yet here again not everything is quite as it seems. The comparison here is between proxy session data for betting shops and visits to casinos by loyalty card holders. This is not a like-for-like comparison as studies of loyalty card data tend to involve skew towards more highly engaged players. This is an important caveat and ought to have been mentioned in the analysis (the likelihood of skew from loyalty cards, which is germane to data interpretation is included elsewhere in the advice but not here).
There is another reason to exercise caution in making comparisons. The study of machine play in casinos (carried out by Forrest & McHale in 2015) states that: “In particular, spending above £1,000 [on a machines visit to a casino] is very rare and there are no cases of a player losing as much as £1,500.”
Indeed, even with skew, expenditure of £1,000 on machines play is higher on sessions in betting shops than on machines visits to a casino. By comparison, we note that a study of loyalty card holders in betting shops (Wardle et al, 2015) revealed at least one session loss in excess of £13,700.
In other words, one can make a plausible case that the incidence of very high losses appears higher on B2 machines in betting shops than on B1 machines in casinos and B3 machines in arcades and bingo clubs. There is all manner of reasons to be cautious when making comparisons between different studies and different data-sets; and the Commission’s interpretation could be questioned by casino, arcade and bingo club operators.
Part of the ABB’s ‘risk similarity’ hypothesis rests on the notion that while B2 roulette has very high stakes, its Return to Player (c98%) is also high. This, the trade association has pointed out, reduces theoretical loss and therefore actual harm – a contention the Commission appears to support. As with so much in gambling, the truth is more complex. Return to Player (‘RTP’) is calculated over long cycles of play and needs to be understood through its interaction with game volatility (for example, the prospect of many small wins compared with few large wins).
A number of researchers have argued that high Return to Player (‘RTP’) ratios can actually encourage persistence in gambling. For example, Parke et al (2016) observed (in a study commissioned by GambleAware): “While a higher RTP generally leads to a lower cost of play (all things being equal) this may also provide a more reinforcing and exciting experience which may encourage excessive play.”
Similarly, Williams et al (2012) commented that “from a theoretical perspective, higher payback percentages will produce higher reward frequency, and heightened prospects for an ‘early big win’, both of which will normally strengthen the behaviour. However, this is offset by the fact that higher payback percentages will also be associated with less financial loss and harm. Thus, there is certainly a payback range that should theoretically have a closer relationship with problem gambling: i.e., one that offers a high enough reward frequency but also ensures significant financial losses and ‘chasing’ (these conditions will also maximise profit for the operator).”
To state that high RTP in and of itself mitigates risk is an over-simplification. If low game margin is so harmful, why are we not more concerned about lottery draws, the football pools and bingo? Are those Las Vegas exhortations to “play the loosest slots in town” really promoting ‘responsible gambling’? A simplistic view of high RTP as risk mitigation could lead to some alarming (and false) risks for ‘softer’ products that do not encourage immediate recycling (including over-the-counter betting). The role of RTP is more complex than the Commission’s advice implies; and alternative theories warrant consideration.
Interpretation matters when it is being used to drive policy. In this case, claims of risk similarity between B2 machines in betting shops and other machines in casinos, arcades and bingo clubs appears to be the justification for mandating player tracking and limit-setting. There may be broader and evidentially strong reasons for pursuing widespread adoption of these measures, but the publicly available analysis does not make this clear (in our view).
The Commission advice on these matters appears to echo the ABB’s 2016 submission to the DCMS review: “The ABB contends there is sufficient evidence around the benefits of responsible gambling measures introduced in LBOs that DCMS and the Gambling Commission should consider mandating similar measures for all establishments offering Category B or Category C gaming machines.”
The highly subjective concept of “similarity” creeps into the policy discussion again – and seemingly at the cost of nuance. Yet, if the DCMS review demonstrates one thing, it is that one size emphatically does not fit all.
Risk mitigation should be proportionate to risk; and broad programmes tend to be more helpful than a focus on a few totemic initiatives. As Williams et al. (2012) observed, “there is almost nothing that is not helpful to some extent and, conversely, there is almost nothing that by itself has huge potential to prevent harm. There is no ‘magic bullet’ to prevent problem gambling”. The authors conclude that operators, regulators and governments should “employ a wide range of educational and policy initiatives” and ensure that they are coordinated.
The ABB’s submission in 2016 contained a table setting out risk factors (e.g. availability of alcohol, presence of ATMs) alongside risk mitigation (e.g. staff training, limit-setting) across the betting shop, arcade and casino sectors. It was factually correct but highly selective. For example, it did not include door control (most casinos and bingo clubs) within mitigation; nor did it mention the use of debit cards for slots play in betting shops within risks. Such selectivity may be fair game for a trade association bent on point-scoring – but should be treated with more than a pinch of salt by policy-makers.
The Commission’s advice to the DCMS seems to endorse (at least in part) the ABB’s perspective, by describing betting shops as “more tightly regulated” than bingo clubs and arcades. The reason for this is not made clear in the document. The other land-based sectors will point out that they are subject to more frequent regulatory inspection than betting shops; casinos and bingo clubs will reflect that their managers are required to hold personal management licences; while casinos are subject to strict anti-money laundering regulations. On the other hand, betting shops may be subject to a range of restrictions that don’t apply in some other sectors – such as stricter opening times or bans on the provision of alcoholic drinks. The point here is that if there is a hierarchy of regulation (such as the ‘pyramid’ often invoked by casino lobbyists), it requires clear articulation.
It would be a shame if disagreements over the data analysis impeded progress on social responsibility measures. Tracking play and offering customers the ability to set limits are positive initiatives – but their use involves trade-offs (both in terms of customer enjoyment and resource deployment).
In 2016, the authors of the Reno Model of ‘responsible gambling’ – Robert Ladouceur, Howard Shaffer and Alex Blaszczcynski – wrote that: “we cannot justify mandatory pre-commitment and technical adjustments to player information displays because these costly strategies evidence minimal effectiveness. More importantly, these strategies hold potential to create iatrogenic effects that can counterproductively increase gambling expenditure.”
The professors may or may not be right – but their views are at least germane to the matter in question and probably ought to have been mentioned in any balanced assessment.
Of particular concern to operators is the idea that raising standards on responsible gambling will be rewarded not with greater trust (as in Matthew Hill’s vision) but increased liabilities. Indeed, the advice from the Commission that “making tracked play mandatory across Category B1, B2 and B3 machines… would increase the availability of information about play, giving…operators no excuse if they fail to identify players that are starting to show signs of problematic gambling” seems unlikely to generate enthusiasm for the project within the industry.
Moreover, the implied threat seems both draconian and simplistic. A clinical diagnosis of problem gambling is based upon a number of aspects of the individual’s relationship with the activity. Only two of these – loss-chasing and tolerance – might be (imperfectly) observed in player data; but tracking is unlikely to yield many insights regarding preoccupation, lying about gambling or risk to significant relationship.
This shortcoming would be an issue if operators were simply being asked to spot problem gambling – but the idea that tracking will enable companies to develop foolproof capabilities for detecting when customers are “starting to show signs” of problems seems ill-defined and out of touch with reality. There has been some promising work on ‘markers of harm’ but the realm of diagnostic data analytics is still in its infancy (and requires encouragement and understanding rather than outcome-driven positioning, however tough-sounding or well-meaning).
On the positive side, the Commission’s detailed advice does seem to accept that full tracking and mandatory limit-setting may involve excessive penalties to non-problem gamblers (perhaps tellingly, while the Commission’s advice recognises the rather antiseptic notion of “consumers’ legitimate preferences” it does not countenance the idea that some gamblers may actually experience enjoyment from gambling). It also acknowledges that trials should precede full deployment. These are fair and important points.
The Commission’s advice appears to have come as something of a shock to some in the gambling industry – but if so, they have only themselves to blame. There have been repeated warnings from the regulatory establishment that this was where we were headed. The regulator has rightly shot down the canard of “evidence-based decision making” as a barrier to intervention, stating that “lack of conclusive evidence need not be a reason for inaction”; but at the same time operators have a right to challenge the evidence base if they feel that selectivity is distorting policy.
So where do we go from here? As the industry considers how to respond to the Gambling Commission, there are five points that warrant attention: