Webinar Transcript: How gambling machines work

Audio transcript (Edited)

Webinar: How electronic gambling machines work

DR LIVINGSTONE: Thank you for that introduction, Alister, and I might as well get stuck into it.  So, I've called this talk today, "How do electronic gambling machines do what they do?" because 
I think it's not just a matter of how they work, it's also the principles that they utilise from psychological understanding that probably make them most effective as they are.  Before 
I do any of that, however, I just put up a slide there quickly on my disclosures, I understand the slides will be made available so you don't have to read it all at once, but I think it's very important for gambling researchers as with other areas of what we call dangerous consumption research to be as transparent as possible about their sources of funding.

So, if you're really interested in the disclosures, then 
I understand the slides will be available later and you can go through them more carefully then.  In this presentation, as this slide suggests, what I'm going to talk about are the principles that support the operation of electronic gambling machines or EGMs as we tend to call them.  These, of course, are also known by other names such as slot machines, in other jurisdictions FOBTs, or fixed odds betting terminals operate on similar but not quite the same principles, sometimes they're called amusement machines with prizes and so on.  But nonetheless, all of these devices are electronic gambling machines.

What I'm talking about today, however, are what are called Australian-style electronic gambling machines, a style which has really taken over the world, particularly in the 
United States and other parts of North America since it became popularised in the 1990s.  The second point on the slides is my discussion about how these devices are attractive and addictive, what are the characteristics of them that make them so.  This is obviously a short-hand version, it's a complex business but what I'm trying to do today is to explain some of the more basic principles and give people a reasonable understanding.

And on top of that, I'm also going to talk about what are called the accoutrements that have increased the harmful effects of EGMs, that is, in a sense their so called addictive qualities, what it is that makes them as addictive as they are and how that has been accentuated in recent years.  And finally, I'll talk a little bit about what we think can be done to reduce the harm associated with EGMs, there's plenty of things that can be done because, after all, EGMs are, at their core, just a computer

So, most of the changes that me and other people have been talking about now for some years can be done simply by software changes, it's not a matter of, you know, vast expense or difficulty to do these things, it's really a matter of political will and regulatory power which, in all Australian jurisdictions, you know, we can find, I think.  The first point, I think, about EGMs is that they operate on a principle of randomness.  Some people mistakenly thing that the machines have this predetermined outcome and that they will periodically deliver prizes.

This is not, in fact, correct, they are actually random, they are what I call highly articulated random number generators.  A way of remembering that is to think about them as a set of dice on speed, so to speak, they are constantly - while the machine is turned on, the random number generators in the machine are constantly generating numbers at the speed of nanoseconds.  So, when the user pushes the button to, you know, place a bet, then the machine grabs one of those numbers from each reel, or four each reel, and converts it into a series of patterns on the screen.

Now, the way most machines operate is that they provide the illusion of spinning, so you have a series of what, in most cases, are five apparent reels spinning across the screen and they usually stop one by one.  And these spins, of course, are associated with a series of clicking noises as though the machine were an old-style mechanical device which operated by a series of mechanical stops.  They're not spinning, the result is known the instant you push the button, there is nothing that you can do to change that result and it is also unknowable in the short-term, that is to say, in the instant in which you push the button, you cannot predict what the number will be.

The best way to think of an EGM in a simplified model, 
I guess, is to think of it like a huge barrel of marbles.  In the case of the machine I'll talk about a bit more today, imagine a device with 35,640,000 marbles in it, a huge barrel, and every time you push the button what you're effectively doing is reaching into that barrel, pulling out one of those 35 and a half million marbles, looking at it, and then putting it back in and then turning the barrel over again and doing that.

So, it's not as though there was a conveyor belt with all the results coming out at you in a orderly manner, it is much more like a giant barrel of marbles with a very large number of marbles in it and with the capacity for any of the results to be repeated time and time again but, of course, at very long odds.  So, what we see is a set of random numbers being converted to a set of symbols on the screen of the machine.  Now, because of the randomness we don't know what the result will be for any single button pusher, indeed, for a series of button pushes.

But because the machine is operating within certain parameters, that is, it has a certain number of available symbols for each so called reel and a certain combination of pay tables and so on, then it is consistent with what we call the law of large numbers, that, over a long period of time, the device will produce the results that have been mathematically determined as its outcome.  So, as I say on the slide, the number and type of symbols are predetermined and consistent.  So, one machine that we've done some work on called, "Dolphin Treasure," an Aristocrat device, has four apparent reels with 30 symbols and the fifth reel with 
44 symbols.

So this means there are 30 to the 4th, that's 30 times 
30 times 30 times 30 plus times 44 which produces 35,640,000 possible outcomes.  So, there's a huge number of outcomes possible with an electronic gambling machine.  I've seen machines which have, you know, more than 50 million possible outcomes.  So, the odds of winning on a poker machine tend to be in the order of one in ten million, when 
I say, "Winning," that is winning the major prize.  I understand in some jurisdictions there are now requirements that that has been reduced to as little as one in seven million, but to put that into perspective, that's roughly the odds of winning a super prize on a gigantic lottery such as, you know, lotto.

In Australia most machines have a top prize of $10,000, so the odds of winning that $10,000 top prize generally are between one in seven million and one in 20 million thereabouts.  Now, you know, that's not particularly good odds particularly since you win much, much more on the lottery.  So, you know, in a way, what the machine is doing is providing you fairly poor returns for what can often be a very considerable outcome – sorry, investment, if I can use those terms.

Now, the table that I've put up shows the real patterns and the pay table for this Dolphin Treasure machine, this is from a paper which was led by Kevin Harrigan, a colleague from Canada, and on which I was a co-author, and it shows some analysis that we did of the Dolphin Treasure.  So, the important thing to look at here is the number of symbols on each reel.  So, if you look at the second set of five columns on the right of that table you will see that, on reel one, if you go down to the king symbol, there is one king on reel one, there are two kings on reel two, four on reel three, five on reel four and three on the last reel.

So what that suggests is that the odds of winning are altered considerably by the placement of the symbols.  If you look at the pay table, which is the left-hand part of that figure, then what you will see is that to get a winning combo of kings you need to get three in a row.  Now, the other thing to remember is that most poker machines, and Dolphin Treasure is an example, pay from left to right.  So that means, in order to get the minimum prize you can get for a series of kings, you would have to get a king on reel one, a king on reel two, and a king on reel three.  So, try to remember those figures, I know it's probably a little bit confusing, we can go back to it.

In Australia, the order in which symbols appear has to stay the same.  In some jurisdictions they allow what's called, "Virtual real mapping" which is the symbols on the screen don't necessarily have to reflect what is on the reel strip in the machine's internal workings.  But in Australia, the order in which symbols appear stays the same and the symbols on that strip have to maintain their number and order.  But because there are different numbers of symbols on each reel, because that possibility is available to designers, the odds of a specific symbol appearing may vary from reel to reel.  And that has a very deliberate effect on the odds of winning a prize, and we'll talk a little bit more about that in a minute.

So, I guess the lesson from this is that the outcome is predictable in the long run but unknowable in the short-term.  So, over the long run, then, we know what the machine will do, it will deliver a certain proportion of prizes back to individuals who are using the machine, but it will collect a highly predictable amount of money from those users which it hands to the operators of the machine.  And it's important to remember that the long run is a minimum of 35.64 million simples, for example, in some cases, as I've said, it can be as many as 50 or more million possible outcomes.

So, the short-term you don't know what's going on and you cannot predict it, in the long-term it's highly predictable, the machine will produce the outcome that has been programmed into it, and the programming consists of a combination of the pay table, that is, the order in which prizes are delivered and how much those prizes will be and the reel strips, that is, the order of symbols on the apparent reels and the way in which they are selected.  Now, this gives rise to consideration in what we call the return to player ratio, so, again, this is something which is a product of the interaction of the pay tables and the reel strips, and it can be calculated with reasonable precision over the long hau

In Australia, the minimum amount of RTP is 85 per cent, that is, the least that the machine has to return to users over the long haul.  Now, this varies between states and venue types, in some states it's higher than 85 per cent, I think in South Australia it's 87 per cent, and it varies if you're in a casino or in a pub or club.  In pubs the RTP generally is higher, in casinos it's generally higher because they have less restrictions on how much you can spend.  Now, what this means is that the average loss of a bet will be 15 per cent.  What it doesn't mean is that you will get 85 per cent of your stake back even over the long haul.

As I say, the RTP is a function of the relationship between pay tables and the reel strip configuration.  So, I'll try to explain that a little bit more.  How do we measure RTP?  In Victoria, the statute, the law, defines RTP as, "A performance of every EGM aggregated in a venue over a period of the calendar year."  So, if you're a venue operator, you don't breach the RTP requirement in Victoria, as long as all the machines in your venue in the aggregate report at least an 85 per cent RTP over the course of a calendar year.

Now, what that suggests is that there is likely to be some volatility in games, that is, some of them might not adhere to their predicted performance over the relatively short-term, and in the life of a poker machine, a year can be a relatively short-term.  In most other Australian jurisdictions, and in most overseas jurisdictions, RTP is measured by the theoretical performance of the game over its game cycle.  Now, game cycle isn't well-defined in most jurisdictions, but what it means is that the cycle of all the possible combinations of outcomes of the game, the time it takes for that to occur.

That can be many years.  If, for example, you played the Dolphin Treasure game 24 hours a day at five second intervals on a single line, then it would take you 5.6 years to have any chance, that's the minimum period of time you would have to play it to have any chance of completing the game cycle, and the probability of all the possible outcomes occurring over that period is quite limited.  So, what we're talking about is a very long period of time indeed.  Indeed, it's quite possible that, over a five year period, the highest prize in a particular machine would not be awarded which would, of course, affect to some degree its RTP performance. And that's why in Victoria the RTP is actually measured by the performance of all the machines in a venue over that given year, and it's also why, in other jurisdictions, what is really being assessed is not the actual RTP, but its theoretical RTP.  Now, in most cases the theoretical and the actual RTP conform very closely to one another, but that is not necessarily the case.  The other problem with all of this is that there is plenty of evidence that users, the people who are using machines, have little conception of the price.  An example of this is, that if you walk into a poker machine venue with $50 and you put that in a machine, you don't know how much time you are going to have on the machine for that amount of money.

You might have quite a long period of time, the probability is that you'll have a relatively short period of time.  On the other hand, if you go to a cinema and you buy two tickets for a film you know how much the tickets cost and how much time you're going to have from the entertainment provided by that film, hopefully it's a good film, but in the end, it is at least predictable.  Unfortunately, the evidence is fairly convincing that people who use EGMs have no real conception of the price, so we tend to use RTP as a proxy for the price. And what that means is that, unfortunately, many users seem to think that, if they use an EGM, then they are guaranteed to get 85 per cent back, that is, if you walk into a venue with $100 and use the machine for a while, you should be able to walk out with $85.  But it's important to remember that the 85 per cent RTP also includes the amount that people are being re-staked.  So, in any – in the course of any EGM session you will lose a bit, then you will win a little bit, then you will lose a bit, then you will win a little bit, but, overtime, the trajectory of your stake is downwards.And that's why most people who use EGMs tell us that when they leave, they leave with nothing because the RTP includes the amount that is re-staked and, as I said earlier, on average, every time you put a bet on the machine will take 
15 per cent of that bet.  So, a way of thinking of that is, even if you do get in front and keep using the device, it will inexorably eat away at your stake until it's gone, and that's almost invariable experience of many people who use EGMs and certainly those who use them regularly and seriously.

So, a way of summarising that is that a user operating an EGM which has a price of 15 per cent, that is, one way of thinking of price is by deducting the RTP from one, so if the RTP is 85 per cent then one minus .85 gives you a price of 
.15 or 15 per cent, obviously if the machine's 87 per cent, then the price is one minus .87, or 13 per cent.  So, in any event, what it means is you're going to lose that much money every time you push the button on average, and it doesn't take a lot of thought to realise that, over a period of time, you are actually going to lose your stake because the machine mostly takes rather than gives.

And the effect is cumulative, so in the example on the slide, if you insert $10 and you bet a dollar every time you spin it, then if the game performs exactly as predicted then you would exhaust your funds in a little more than five minutes, that is, at the rate of one bet every five seconds.  And if you were betting on $5 then it would take you just a little more than one minute.  So, you know, what we're talking about is a machine that has a very, very good capacity to devour considerable sums of money even though the bets seem to be very modest.

All right, so that's some of the background I guess, it's important that we understand why it is that people use EGMs.  One of the things that most – my colleagues ask me is, "Why is it that people get sucked into these things?" and the average sort of person, I guess, would say, "Well, I don't understand it, I don't even like the things, I can't enjoy them" and so on and so forth, and that' s what most people would say to you until they get drawn into the machines.  So there's two well-known psychological principles that apply to the operation of EGMs, they're called operant conditioning and classical conditioning.

Now, these two principles apply to any mechanical gambling device, including electronic gambling machines and old-fashioned one-armed bandits of the type which proliferated in New South Wales from the 1950s onward.  What has changed, however, since the early devices is that all of the facilities available via electronic computers have been incorporated in EGMs, and so they've been very much enhanced over the last 50 to 60 years.  We'll talk about operant conditioning.

Some of you may remember this from your psychology studies but, essentially, B.F. Skinner, an American psychologist and researcher demonstrated in the 1950s using animal studies, rats and pigeons mostly, that animals, including humans, will respond to unpredictable stimuli, a pattern of unpredictability is important to this, rewards in a sense, but developing patterns of behaviour that are difficult to extinguish.  In other words, if you give someone a prize every second time they press a lever, they get bored and they only press the lever when they want the prize, when they want the reward, and this is what he discovered with pigeons and with rats.

But if you randomise that schedule it creates habitual behaviour, the animal will keep pushing away at a lever often under extremely difficult circumstances involving electrified grids on which they're standing and so on, and they will continue at that until the behaviour becomes so established that it's very hard to extinguish.  And this is much, much more likely if you have a random sequence of rewards than if you have a predictable reward sequence.

Now, this has been well-known since at least the 1950s and this illustration with our little ratty friend in there is a Skinner box, this is a little box which was designed for Skinner's experiments and which has been elaborated upon many times in the last half century in which, as you can see, the animal pushes down on a lever, it's taught to push down on the lever because, every now and then whilst operating the lever, a little pallet of food comes in and it gets a buzz and continues to push the lever.

And often animals will do this to exhaustion, they will keep pushing a lever until they're exhausted or they'll keep pushing the lever even though they also are getting an electric shock.  So, it establishes behaviour which is very hard to extinguish and, unfortunately for those of us who think that we're better than animals in many respects, humans operate on exactly the same principle, if you give a human a set of unpredictable rewards, that will establish behaviour associated with that reward which is very difficult to extinguish.

Classical conditioning has been known since the 19th Century and Ivan Pavlov, a Russian physiologist demonstrated that animals, and again, including humans, will respond to stimuli that are associated with a reward.  So, in Pavlov's classic experiments he played a metronome, he set a metronome off when he fed a dog and he had a little device that measured the extent to which the dog was salivating, and eventually, what happens if you feed the dog and played a metronome, the dog comes to associate the sound of the metronome with the reward, the food, and eventually you can play the metronome, not give the dog any food, and the dog will also salivate.  So, that's a conditioned response.

So, if you hear a specific sound when you get a reward, you will rapidly associate that sound with the reward or a smell or a sight like flashing lights or something which indicates that you've received a reward.  So, if you walk into a gambling venue, a poker machine venue or a casino, you will instantly be assailed by sights and sounds that are associated with people getting rewards, machines going off, flashing lights, all the hoopla that goes into poker machine design.  And, if you walk into a casino, you will also encounter quite a distinct smell that you're unlikely to get anywhere else, and that's because most casinos pump odours through the air to develop an association with that activity.  So that's classical conditioning, and here's a little illustration which shows how it works.

So, before conditioning you give the dog some food and it salivates at the prospect of getting the food.  Before you've conditioned the dog, if you play a tuning fork or play a metronome, there's no response from the dog, it's not interested, it's just another sound.  The picture marked three there is the process of conditioning.  If you play the tuning fork and give the dog the food, then the dog will start salivating principally because the food is there.  But eventually, once the dog has been conditioned, and it doesn't take that long, then playing this tuning fork will generate the same anticipation of getting a reward.

So, what this is doing, what these two principles are doing is building a very powerful model for conditioning people to particular responses, so, what's actually going on through this process.  Now, you know, for many years we knew about conditioning principles and we knew that they worked.  It's only really been relatively recently with the advantage of MRI imaging and so on, that we've been able to actually look inside brains to see what's happening while this occurs.  And it now appears that what's happening is, when we get a conditioned stimulus, something to which we have been acclimatised, as it were, then our brain reacts.

And it reacts in various ways, but one of the principle mechanisms is what we know as a dopamine system, this is a brain – a neurochemical which operates the reward system in our brain, it produces a pleasant, euphoric experience and that's because, you know, when we were swinging through the trees or walking across the veldts of Africa a long time ago, the idea of attaining a reward was a rare thing.  So, when we saw something that gave us the idea that there might be some game that we could catch or some food that we could eat, our brains remembered it, they were stimulated, they developed a conditioned response and we came to enjoy that response because we knew it was associated with a pleasurable outcome, whatever that might be.

So, what the poker machine is doing is harnessing these, I guess, developmental principles, and all animals have them, we all have them, it's harnessing these principles to trigger the dopamine system.  And so when a reward is received or when a reward is anticipated, dopamine is released and this causes this pleasurable, euphoric experience.  And it looks like both classical and operant conditioned responses derive from this.  So, if you look at a poker machine, it takes no prizes, unfortunately, to understand that this is exactly what the machine is doing.

The machine provides a random reinforcement schedule, as Skinner would have called it, and it provides sights and sounds and, in some cases, even smells that will condition the user to want to keep doing this.  And EGMs have become extremely good at this over the last few years in particular as they were computerised and transformed from sort of primitive but still reasonably effective devices operating on mechanical stimuli.  If you think about an old-fashioned one-armed bandit, the reels would ring around providing an anticipation, there would be a sound associated with that, they would then go click, click, click and click into place, if there was a prize the coins would drop into the tray below generating another sound with which people came to associate this sort of pleasant experience of having a little win, but of course, you know, they were programmed just as electronic machines are to take a certain percentage.

But what we got with electronic and computerised poker machines was a whole new capacity for multiple tricks, you can call then, which increased the reward rate without necessarily costing the operator any more money.  And I'll talk about some of these now.  These include what we call near misses, losses disguised as wins, jackpots, game features and the lights and sounds.  All of these things maximise reinforcement and, as I said, many of them do it without actually costing the operators anything.

Now, if you want to just understand the concept of losses disguised as wins, it's important for those of you who are not necessarily familiar with how contemporary poker machines operate to have a look at this slide here in which we portrayed the lines which you can use on a 50 line machine.  So, if you look at the top left-hand corner you'll see that line 1 goes through the middle, that is, the middle row of the three available, line 2 operates on the top row and line 3 operates on the bottom row, that's pretty straightforward

Line 4, on the other hand, starts at the top-left corner, goes down to the centre reel at the bottom and then zips up to the top-right corner, row 5 is the inverse and so on.  So, you can see that, using this quite elaborate set of constructions of rewards, it's possible to derive 50 possible outcomes, in some cases you can do even more than this but we'll stick with 50 for the time being, I think it's probably easier to comprehend.  So, it's important to understand the concept of losses disguised as wins, in order to do that, you have to understand how the lines operate.

Now, most serious users of EGMs select as many lines as possible, and the reason for this, of course, is in a curious sort of way they're highly risk-averse, that is, I know it's a funny thing to say, but people are very unhappy if they see a win occurring in a line they're not playing.  So, for example, if you're playing only the middle line and you get a 'win' on the line above or the line below you think you're missing out.  And so it was very common, as has now been known for some years, for user to select what they call the minimax strategy, so you will bet the minimum amount on the maximum number of lines.

So, if you're operating on a one cent machine and you're operating 50 lines, then the one cent machine immediately becomes a 50 cent machine, so every time you push the button you're spending 50 cents.  You can see why the operators would like this device because it accelerates the rate at which people contribute money to their business.  So, the minimax strategy also means that you get a win on a specific line, but if you get a win on one line but lose on all the rest, the net result will be a loss.  So if, for example, if you were to bet $1 on a 50 line machine and you get a win which pays ten cents then the result is a net loss of 90 cents.

But nonetheless, the machine goes off, has a little verbal at you, plays a tune, produces a reinforcement and can, in fact, double the reinforcement rate as has been now shown through some fairly good research from Canada from 
Kevin Harrigan and others.  So, what you've got is, you've got a reinforcement which is produced when people get an actual win, that is, when they actually win more money than they lost, but you'll also get about the same level of reinforcement measured physiologically from a loss disguised as a win, and both of these, of course, contribute to the reinforcement rate but without actually costing the operator anything.

The next thing to talk about is what we call near misses.  Now, you know, in Australia near misses may not be deliberately engineered according to the technical standards, however if you have a different number of symbols on various reels, as with Dolphin Treasure if you can recall that table I put up, then what will happen is that, if you have more kings, for example, on the latter three reels than there are on the first two, then king symbols will appear quite regularly, much more likely to appear than if you're actually winning the prize.

So, if we look at the rather wordy next slide you can see that the odds of obtaining five king symbols on the particular game we described are one in 297,000, which you can – the calculation is there if you want to replicate it.  So, getting five kings has odds of one in 297,000, but because there are multiple king symbols on the later reels, it's not unusual to see a grouping of multiple king symbols on the screen.  So, the odds of seeing a king on the last three reels but not on the first two are one in 660 and, on both the third and fourth reels, they're one in 45.

So, what that means is that these combinations will occur relatively frequently, particularly compared to getting five king symbols which is one in 297,000 and even when compared to the odds of getting the modest payout that you achieve for three kings which is one in 3,375.  So, the effect of this is to allow near misses to appear quite regularly, and we know from research that near misses also produce a response, a physiological response which is not quite as significant as that achieved from either an actual win or a loss disguised as a win, but still, a significant reward. And the next thing I'll talk about is what we call game features, game features are very popular with serious users, now been a couple of pieces of work in Australia which demonstrate that one of the key characteristics of machines that users look forward to are these game features.  In a game feature you get what's called free spins, so the free spin, you know, you get a scattering of symbols across the screen, it triggers the free spins and, often during the course of the free spins, any prizes you win are multiplied depending on how much you've been betting.

But it's not uncommon for them to be three, four, five times as much as a standard prize.  But they're not free, the probability is factored into the RTP, but because they appear to be a really good reward and people really look forward to them, they provide another reinforcement at no net cost to the operator.  So, they are factored in, there is a deduction from the actual prizes made for the cost of having the so called free spins factored into the machine, but they provide, clearly, a very significant reinforcement, again, at no net cost.  And jackpots.

Now, jackpots are different from maximum prizes, the maximum prize on a machine occurs when you get an order of symbols which triggers the highest prize.  Now, as I said, this can be a very, very remote possibility, perhaps one in 20 million, one in ten million, something like that.  But jackpots are funded by a contribution from the amount of money that's put into the machine, and this can be around two per cent or more of the wager.  So, what this means, on a machine with 87 per cent return to player calculated, the actual RTP is reduced to 85 per cent.  And some machines, particularly machines with a linked jackpot, that is, where the jackpot pool is linked either locally or, in some cases, remotely across venues, then there will be multiple deductions made.

So, if you think about it, the raw RTP of the game might be 87 per cent, if that's reduced by two per cent for one of the jackpots, that will take it down to 85 per cent, if it's reduced by another two per cent for another jackpot, a local jackpot, that will take it down to 83 per cent, and then once the deduction is made for the feature and so on, then it could be as low as 80 per cent.  So, you get a very significant reduction in the actual return to play for everyone who doesn't win a jackpot or doesn't have good luck, as it were, with the features, and, sadly, that can be almost everyone.

So, the actual RTP that people experience can often be quite a bit less than the RTP which has been calculated into the machine.  So, you know, these are – this is just a sort of a quick outline of how the machines operate, I hope it's given you some insight, the question is, what can we do about this?  Are there policy implications?  Well, there are and it's important to remember that, despite the bluster that some of the manufacturers in the industry come out with from time to time, EGMs are computers, they can be modified very readily, the software on them can be changed quite rapidly and easily and, in some cases, remotely.

All the EGM parameters are open to change and EGM parameters in Australia are specified by the Australian/New Zealand Gaming Machine Standards which all the jurisdictions in Australia and New Zealand agree to.  Now, my argument would be that these are heavily underutilised from a consumer protection perspective, they tend to be much more concerned with technical issues around how the machine should operate, what should happen when you open a door, obviously that it shouldn't electrocute people, et cetera, et cetera.  But in some jurisdictions, harm protection standards using the parameters have started being implemented, and there are some examples of this.

The maximum bet is a parameter which can be readily altered, and has been in some jurisdictions.  For example, in New South Wales the maximum bet from year dot, really, has been $10, it hasn't been reduced unlike in Victoria and Tasmania, at least, where it's now been reduced to $5 and there are calls by some, including the productivity commission, that this should be reduced to a dollar.  In New South Wales, a load up, that's the amount of money that you can stick into a machine at any one go, used to be $10,000, it's been reduced to a mere $7,500, in Victoria it's $1,000 and in Queensland it's $100.  I should say, the productivity commission suggested that it should be $20.

The number and denomination of bank notes that you can insert is also capable of being regulated, in New South Wales all notes up to and including $100 can be inserted, in Victoria it's $50 notes and in South Australia now bank notes at all are permitted, it has to be operated by coins.  And around issues of losses disguised as wins, in Queensland, although a loss disguised as a win can occur, no sounds are permitted to omit from the machine in consequence of this, so the reinforcement can be modified, the rate of reinforcement can be modified.       

So, these are just examples of what is possible and each of these has been implemented in an Australian jurisdiction, and without huge outcry from the industry.  So, it's interesting to see why it is that, in some jurisdictions, there's massive resistance to modifications and yet, in other jurisdictions, there has been important but perhaps not quite as far – not going quite as far as we might hope.  So, what can be changed?  What are the things that we need to focus on?

I mean, my argument would be that, you know, this is a list, it's not an exhaustive one, but I think these are important priorities for change in the way machines are designed, drawing on our understanding of the structural characteristics.  So, the first of those would be load up limits, allowing people to put vast sums of money into a machine is simply encouraging overconsumption and making it easy.  The maximum bet, obviously the higher the maximum bet, the more money can be lost in the shortest possible time.

Looking carefully at losses disguised as wins, how they operate and what they are – what can be modified around them, as in Queensland where the reward as such has been diminished.  The capacity to utilise starved and uneven reels, starved reels are reels where the number of symbols have been reduced, uneven reels are reels where you have a certain number of symbols on each reel but that changes from reel to reel, and as the Dolphin Treasure example shows.  People do not understand RTP, time and again I have had conversations with people who are very experienced EGM users and still don't understand how the return to play ratio operates and what it actually means, as I've tried to explain today.

The major predictor of harm from poker machines is ready accessibility to them and, in most Australian jurisdictions, they're on every street corner pretty much, every pub, most clubs have poker machines and many of them are very large, so large venues are much more harmful, we know that absolutely, than smaller venues and the accessibility issue is a really, really key predictor of how much harm will be associated. And finally, of course, again as the productivity commission pointed out and others have been arguing for for many years, the capacity of an effective pre-commitment system, one which is universally applied and which would allow people to determine in advance how much they're willing to spend on a machine, will give people some certainty, at least it might allow them to decide, "Well, I want to spend 50 bucks tonight, and if I run out of that in ten minutes, that's as it may be, I won't rush back to the ATM and get more."

So, things could all – they're all amenable to change and none of them are particularly radical, all of them can be easily done and very readily implemented without massive cost to industry, it must be said.  So, I'm getting to the end, I'm glad you'll be – I'm sure you'll be happy to hear that.  There's a slide here with some useful links which helps people who might be interested to discover a little bit more, to find out a little bit more, and some of these come from my colleagues at the University of Waterloo, some are related to other work that's being done, but all of them provide useful and helpful information both for people interested from a policy angle, and people interested at a practical level.

And I'll finish off with my acknowledgements.  Thanks to the AIFS and AGRC staff, including the people listed, to the anonymous reviewer and to many of my colleagues and people with whom I've worked on this over the last couple of decades.  And if you're interested in the references then they're available on the discussion paper which is linked to on the AGRC website.  Thank you very much.

MR LAMONT:      Well, thanks so much, Charles, that was a fascinating presentation.  We now have time for some questions, and they're flowing through which is fantastic.  The first one I'll read out now is, "Can the RTP be adjusted on the fly with linked machines to adjust the time of year or current venue demand?"

DR LIVINGSTONE:  Well, all games have multiple settings, so for example, one game, a Dolphin Treasure game can have maybe three or four different RTP settings and, in some cases, these can be adjusted remotely.  So, it theoretically is possible for the operator to adjust the RTP on a particular machine or a particular group of machines by changing the game set, and they can do that, you know, virtually overnight.  Do they do it?

I haven't seen any evidence that they do and the operators insist that they don't, but it's certainly theoretically possible and I would imagine that there are occasions where the RTP has been adjusted but probably not overnight or, you know, in response to a particular time of year.  There was an experiment which, I understand, when there was a duopoly in Victoria one of the operators set their machines at a relatively high RTP in the expectation that people would discern that they were offering a better deal, a better price, if you like, and the other operator set theirs at roughly the minimum.

And what happened was that the one with the cheaper machines, as it were, just simply got less money.  So, they did an experiment in the eastern suburbs of Melbourne and reduced their RTP, and all it did was make them more money.  So, overtime, what we saw was that both of the operators migrated their machines to the most profitable RTP setting, which is close to the minimum, and essentially they get away with it in Australian jurisdictions because, as I said, the price is not readily discernible.

MR LAMONT:  Okay.  Well, we have another question related to the RTP, this one's a bit long but I'll read it out.  "Would reducing the RTP and therefore the rewards and time in which players are programmed through reinforcement schedules to something like the RTP for lottery tickets be a policy initiative that could improve the safety of machines?"

DR LIVINGSTONE:  So, is that – that question seems to be asking if we make them more expensive, will it make them more consumer friendly.  I think the answer to that is, probably not.  People will go through their money more quickly, but all we know from experience, I guess, of those who use machines is that they become very good at finding new ways of pouring money into the machine.  And really, the harm associated with poker machines has its origins in the amount of money and the deception required to get that money which is associated with the sort of 'addicted user.'  So, you know, if you make them more expensive all it means is that they will find the money elsewhere.  As I said, price elasticity for poker machines is not very high, so people will keep using them, particularly habituated users will keep using them even though the price is very high.

MR LAMONT:  Okay.  And on the same theme again, "How is the RTP monitored in jurisdictions other than Victoria, given it could be over many years?"

DR LIVINGSTONE:  Well, all the jurisdictions now have outsourced the approval process to testing agencies, authorised so called independent agencies that go through the game maths and operate the machines and simulate them to see what the outcomes will be, and if they say that the theoretical RTP conforms, then that's how it's measured.  At least in Victoria there is some practical assessment which the regulator imposes by looking at the aggregate of machines over the year but, in most jurisdictions, it's simply the theoretical RTP as certified by the testing agency.

MR LAMONT:  Okay.  They're flowing through, these questions.  Okay.  "RE jackpots, can these be linked electronically to other machines in different venues or to other machines in the same venue or both?"

DR LIVINGSTONE:  Yes, they certainly can.

MR LAMONT:  All right.

DR LIVINGSTONE:  And they are, I mean, often – it's very common to see a game which will have a local venue, you know, it'll have a series of machines linked together on a bank of machines or what we call a remote-linked jackpot which is, you know, across all the machines operated by a certain operator.  So, you know, they can be – they can offer very significant prizes, obviously, but, you know, essentially the contribution from each machine is paying for that, it's not an add on that you get for free.

MR LAMONT:  Okay.  Another question here, "What do you think about the idea of safe spending limits for EGMs?"

DR LIVINGSTONE:  Well, I mean, you know, there's been debate about whether or not safe spending limits would be effective, and perhaps they could be in the early stages of people developing habituation to EGMs, I think they're likely to be – they may be helpful for people who are just starting to worry about whether they've got a problem, but I think the real problem is that I think, as the VRGF's latest campaign puts it, the harm begins before you realise it.
Many people become habituated to the machine without really understanding quite what's going on, and by the time they realise that they've got a problem, then the only thing that calms them down is using the machine again.  I mean, I think one of the important things to realise about poker machine addiction, for want of a better term, is that it's something which arises from a combination of factors, it's not any particular factor, someone who uses a poker machine once a week or more has a probability of about one in three of developing harm, someone who uses a machine three times a week is very likely indeed to have a problem with machines and they don't necessarily realise is.

And this is part of the problem about the normalisation of venues in places like Sydney where there are clubs everywhere, and many of them are huge, and where it's very common for people to use these places a social outlet, the machine will have snuck up and bit you and addicted you before you've even noticed it, I think that's the problem.  So, yes, if you're talking about people in the phase where they're getting engaged with machines and they finding them fun and amusing, and before they start getting sucked into the machine, then maybe safe spending limits would be helpful.  Personally, I think it would be better to point out to people that it's not necessarily even how much you spend, although that can be important, but how often you go that is going to be a predictor of the likelihood of having a problem.

MR LAMONT:  And one really quick question here, "Does the maximum bet refer to the sum of money bet on all lines played per spin?"

DR LIVINGSTONE:  Yes, it does.  So, in New South wales the maximum bet of $10 is imposed across the top of however many lines you bet.  So, if you've got a – I don't know, if you've got a 50 line game and it's got a dollar maximum bet then you can't – you still can't bet more than a dollar a spin.  However, in casinos it is frequently the case that those limits are lifted.  At Crown Casino in Melbourne, for example, you have unlimited bets on machines but you have to put in a loyalty card to play them, and that's not an uncommon characteristic, but in many jurisdictions – you know, I've seen jurisdictions in Canada, for example, where you can bet hundreds of dollars on a spin in a selected area, you have to, you know, choose to go into that place but it is possible to lift it outside that, but in Australia the maximum bet is the maximum bet in a pub or club venue.

MR LAMONT:  Thanks, Charles.  One final question here, "High alcohol products are taxed much higher than low alcohol products with the view to reducing harm through reduced consumption.  Could a similar regime be applied to EGMs?"

DR LIVINGSTONE:  Yeah, I think that's a good idea and, in fact, what we know is that, whilst price is not an issue that most ordinary people understand with relation to EGMs, the price of operating EGMs is a factor which the operators would very much understand.  So, you know, some years ago I wrote a paper which suggested that the way to help ease the operators away from the machines, to help them get over their addiction, if you like, was to impose a highly progressive tax so that the more money you made per machine per annum, the higher your tax rate.

And our modelling suggests that government could impose that, it would reduce harm and it would actually help them offset the losses to government revenue from reducing the amount of money spent by poker machines.  So, what it's about is reducing the incentive to operate machines that are ripping people off to extraordinary degrees, so for example in Victoria very high performing venues frequently have more than $200,000 a year revenue per machine, so taxing those sorts of machines at very high rates might reduce the reliance of the venue on that sort of almost free money, the licensed print money which the government has granted them, unfortunately.

MR LAMONT:  Okay, I'm just going to have one final last question here, and that's, "How do the multi-terminal machines work that allow something like $100 maximum bet?"

DR LIVINGSTONE:  Okay.  Well, multi-terminal machines are allowed in clubs in New South Wales and are increasingly happening in casinos around the world including in Melbourne and Sydney.  So, what these are is essentially simulated or electronic casino-type games such as roulette, blackjack and so on.  And so what you could do is you walk up to this thing, it's got a multitude of terminals sitting around a particular location, occasionally in casinos there's a real person spinning a real roulette wheel and you can put your bets on through a screen, a multi-touch terminal screen.

More commonly, particularly in clubs, it's an entirely automated process, there's a big screen, a TV screen up in the middle of this display and a invariably scantily clad young woman simulated on the machine who's spinning the wheel and they allow you to place a bet using a screen, often they will have multiple games going on at the same time, you can select from those games, in some operations you can select from multiple games and bet on all of them at once so it's just like playing roulette except it's in a club or a pub and it allows you to spend up to – in New South Wales you can bet up to $100 a spin on these machines.

Very similar to the type of machines that are operating in the UK, FOBTs, fixed odd betting terminals, and which are associated with huge levels of harm in the UK and subject of recent protests throughout the policy.  So, they are, you know, essentially automated casino games and in New South Wales they're available in clubs and, you know, I've played them in clubs in New South Wales, and I have to say, they certainly seem to stimulate great interest and very high expenditures from the people using them.

MR LAMONT:   Well, thanks, Charles and I think we're going to have to wrap up the questions there.  If you do have any other questions, please send them through to us at comms@aifs.gov.au, and yeah, once again, thanks very much, Charles, for your presentation today and thank you everyone for attending.  Please remember you can download the paper upon which this webinar is based at our website and you can also signup to our mailing list at this site to stay informed about future webinars and reports published by the Australian Gambling Research Centre on our website.  So, thanks very much everyone.

Gambling support

Are you experiencing issues with gambling?

Find the help which is right for you by visiting our Gambling help page.