Skip to main content

By chance

Truck outside Wrest Point Casino with a message from David to Greg

David Walsh

Posted on Thursday 1 March 2018

Share on

What’s the worst decision I’ve ever made?

I’ve made a few bad ones. In my brief bad bridge days, I put down a cold four clubs contract, or so my mates told me. One former friend was particularly precise—‘the only bridge you seem to know is an open space spanning construct’. Those words taunt me thirty-five years later. And romance—I’ve messed up great relationships, and persevered with some shockers. After a bikie taunted me with, ‘Don’t bash us up’, I heard my weaselly fifteen-year-old voice, ‘I won’t, but only because you asked nicely’. When asked by the Israeli airline (El-Al) security guy if I was traveling with anyone, I gestured towards my Palestinian mate and replied, ‘Only that terrorist over there’.

None of those are anywhere near the worst decision I’ve ever made. The worst decision I’ve ever made was to give up my studies and start gambling with a bunch of itinerants I had just met. Ok, so that turned out well for me. Very well for me. Almost everything I’ve become is a result of that decision. Nevertheless, it was an appalling choice. Staying at uni, and studying, would have lead to a decent outcome 99% of the time. Gambling—not so much. I very much doubt that 1 in 100000 gamblers is able to make a living. I made a bad choice, but fell in with a good crowd, and got lucky.

If you ask anyone about the good and bad decisions they made in their life, they will offer those that resulted in positive outcomes as good decisions, and those that resulted in negative outcomes as bad decisions. They shouldn’t. The outcomes are only slightly coupled to the choices. But human beings are not biologically primed to measure randomness. They are designed to seek cause and effect. Cause and effect has positive symmetry. That’s because, years ago in the savannahs in Africa, where we were mostly designed, it didn’t hurt to see a lion that wasn’t there. But those who didn’t see the lion that was there—their genes departed the gene pool. They won the Darwin Award.

Possibly more importantly, seeing the world as cause-and-effect meant they could make quick decisions. Probabilistic reasoning doesn’t help, because it is too slow. Let’s see, there's a 400 to 1 chance that that’s a lion, but I can’t be sure without a closer look... The shortcuts that help us to make a useful decision faster than the correct decision are called cognitive biases, or heuristics. They kept us alive, but we don’t get a choice about where we apply them. In cultural situations heuristics are often useless.

Consequently, and I’ll get back to this, we are very bad at measuring low probability events. And we are very bad at decoupling a stimulus from an outcome, when the outcome is random, or linked by low probability. So we learn poorly in situations that are mostly random. Chess is deterministic—it’s relatively easy to learn chess, or to teach a computer. It’s really hard to get good at poker, and so far computers are poker novices (at least in real world situations). Learning poker requires building appropriate reinforcement on non-deterministic outcomes, and not being ‘fooled by randomness’. Since bridge is like chess—easy—and I couldn’t play bridge, I made a good decision to not play poker. In general our brains are repositories for believing we are right when we make lucky stupid decisions (so are they suppositories?).

Our savannah generated pattern matching engines make games fun, and they make randomness (stochastic processes) look ordered. Poker machines tap into this—they are fun because the biological engines that once kept us alive are pleasurable. Sex is pleasurable for the same reason. We don’t have a drive to have sex, or to play pokies. Our drive is to seek pleasure because, well, it’s pleasurable. Sex is good for us. Pokies, not so much. But we, savannah programmed meat machines, can’t discriminate. And we don’t have identical programming—there’s also an addiction lottery, that we all buy tickets in.

As I said, years ago I made a bad choice to gamble, but it turned out okay, so it’s easy to conclude that it was okay. I had already placed bets then, so addiction wasn’t likely, but a misspent life was. But a few years before, I placed my first bet. That’s a really bad decision. It was on the doggies (greyhound racing)—my dad was a dog trainer. That bet probably had about a 1 in 500 chance of causing a gambling addiction. (Later, I say that for pokies, that might be 1 in 20. And a pokie addiction is much more ruinous.) So, if I understood unlikely events, I would never have placed that first bet. That first bet is the only choice you get. And choices have consequences beyond the individual. Some choices have clear (but, perhaps unlikely) consequences for other individuals, and for society as a whole. That’s why they’ll put you away if you keep driving drunk, even though a negative outcome is unlikely. And that’s why smoking in hotels isn’t permitted anymore. You only get to make bad decisions if they don’t affect others, or because the causative links haven’t been demonstrated (as was the case with passive smoking for years).

Twenty-five years ago Greg Farrell said, ‘Direct access to gaming machines in pubs and clubs would have a disastrous effect on the social and special culture of Tasmania’. He was most likely correct, because each link in the causative chain was highly probable. However, in terms of outcomes he was apparently wrong. He believed that he might not get the poker machine licence, and if he didn’t, and others did, it would impact his casino business. In terms of probabilities he was right. In terms of ethics, he was also right. But I bet he wishes he hadn’t said it.

Right now, I’m actively opposing pokies. I’m not sure what I thought in ‘93 when Greg was so vocal in opposition, but the social evidence has ossified my opinion since. That could have negative consequences for me. The Liberal Party has been a significant ally, particularly with the festivals, but it is the favourite to win the election, and, even though I stated my opinion before they did, some Liberals might be grumpy at me thereafter. I might regret my decision to spend $250 grand opposing pokies, but that won’t make it the wrong decision.

Because individuals make bad decisions, but there is (with the right distribution) wisdom in the crowd, democracy works. Betting markets, which essentially comprise aggregate opinions, are very efficient. A Russian newspaper had its readers play Boris Spassky, then the second-best chess player in the world. The readers’ moves were chosen by tabulating options and choosing the one that had the most votes. They achieved a draw (so, ironically, one of the bests tests of democracy was in the Soviet Union). But that wouldn’t work for poker. Biases that everyone possesses push aggregate opinions in one direction. I said earlier that betting markets are efficient. In fact, we can only beat them by assuming they’re right and looking for biases. The pro-pokie lobby is trying to bias our population. That’s why I’m involved. I can live with pokies, provided they are what the people want, so I’m spending money (but much less money that the pro-pokie people) to attempt to de-bias the election. At its best, democracy works even for those that bet on the wrong side, because it needs the weight of opinion to become an efficient market. Ideally, unlike betting markets, those opinions should be weighted evenly. And when they aren’t (because some, like me, can get their opinions heard) the right to express an opinion should be voided by skin in the game. You shouldn’t be able to buy an election for the purpose of enriching yourself.

Getting back to our imperfect natures, we find meaning in stochastic (random) links because in our biological background (i.e. the savannahs and other ancestral environments) seeing a connection when it was there does not inhibit survival, but seeing when it is probabilistically linked abets survival. Since it is difficult to discriminate between the two, and there is no survival impediment to linking them, evolution has not decoupled them. So here we are, trained to respond to the stochastic as if it is causal.

Addiction-by-design machine designers know this. They have degrees in all the right disciplines. And they are biologically primed to seek status (status gets you laid, status gets genes into the next generation), and money is a proxy for status—so they take the cash and self-justify. It works even when the stimulus is decoupled from the effect, such as when the guy that owns the machines has no benefit from extra cash since he already has enormous privilege and prestige and is beyond breeding (but not the almost equally efficacious grandparenting), and it works even when the guy writing about the effects possesses similar privilege and prestige, and needs no more.

Let’s play a game called, ‘Bet your life’. The game goes like this. A weird rich guy offers you a proposition. ‘Put a bullet in a chamber of a six gun. Spin it. Point it at your head. Pull the trigger. If you survive I’ll give you a million bucks’. Would you do it? After all, the odds are pretty good. Five chances in six of getting rich. Of course, it’s a stupid bet. It seems logical that one should never make a bet that entails a downside of ruin. Here ruin means death, but it could mean financial ruin, or a loss so great that it is intolerable.

As it turns out, I’m a weird rich guy. Let me offer you another proposition. Here, you’ve got one chance in 45 million of getting seriously rich, and one chance in 45 million of losing your life. This version is worse than the version with a gun, because in that scenario the chance of getting rich was five times the chance of getting dead. Here, the chances of dying and getting rich are equal, but the signal is hidden in the noise. Would you play that version of Russian roulette? If you’re like most Australians, the answer is yes. When you play Oz Lotto you’ve got about one chance in 45 million of winning. But for each five kilometres you drive to buy the ticket, you have the same chance of crashing your car and dying as you have of winning. So if you’re buying a lottery ticket, do it online. And don’t think about what getting the first few numbers up does to your risk of having a heart attack. Why would anyone drive to buy lottery tickets? The noise disguises the asymmetry. More particularly, we see the positive consequences—rich lottery winners on TV, but we don’t see the lottery losers in the morgue. Our genes do not benefit from being able to discriminate between the chance and potential consequences of very unlikely events that are ever-so-slightly predictable. Stochastic events are, in terms of our ability to discriminate between them, inseparable, like death or a lottery win. Our biology sees, and seeks, the bright side. We seek serendipity.

When I’m in a taxi, and the taxi driver is tailgating, I tell him he is taking an unnecessary risk. He says, ‘I’ve never had an accident’. I don’t tell him, but I’m telling you, that when he isn’t tailgating his chance of dying (and mine) is about 1 in 300 million for every kilometre driven (taxis are slightly safer than random). By tailgating, he is at least doubling that risk. If he is a lifetime taxi driver, and he drives that way all the time, he is lopping two years off his life. Noise disguises his risk. He is playing Russian roulette.

There are two things worth noting here. We are not designed to understand risk. Even when we understand it mathematically, we don’t understand it biologically. I know smart, mathematical gamblers—card counters who win at blackjack—who got addicted to poker machines. And, although it looks like freedom to be able to have a punt on the pokies, it isn’t. If you stick a spike in your arm, you are about 23% likely to become a heroin addict. If you have a drink, you are about 8% likely to become an alcoholic. And if you play a pokie, you are about 5% likely to become an addict (a bit of guesswork here—20% of people occasionally play pokies, but I’m guessing around 50% have played pokies). The heroin addicts don’t choose to be heroin addicts (although, unlike pokie players, they are aware there is risk). They choose to try heroin, and then the biological lottery takes over. Willpower, honour and intelligence don’t come into it once you’ve had a taste. All that matters is biology (provided there is access). That’s why heroin is illegal. Alcohol isn’t illegal (everywhere), but alcohol has a significant social upside (just ask Bacchus). It nevertheless does great harm. If you advocate for poker machines you are advocating for an unlikely (but not very unlikely) addiction for every individual that tries them. You are not advocating for choice, because addiction doesn’t discriminate, except by chance. And, apart from the spurious freedom argument—where’s the upside with pokies? It does seem to be the only form of gambling that is addictive by design. I note that it’s the addiction by design element that has cigarette companies on the back foot (a key discriminating factor in U.S. v. Philip Morris was, ‘Defendants have designed their cigarettes to precisely control nicotine delivery levels and provide doses sufficient to create and sustain addiction’).

One approach to protecting freedom is harm mitigation. But the harm is indiscriminate, and blameless. A supply of bandaids does not make inadvertent harikiri a better option. This sort of harm mitigation is like getting rid of traffic lights, but buying more ambulances (these ambulances run others over—each pokie addiction affects several other lives negatively).

Poker machines use noise to disguise the fact that they are financial Russian roulette. Addiction-by-design machines reinforce the positive asymmetries. Every outcome feels like a lottery win. But it’s a bullet in disguise. So here we are, a community that has had a stochastic gun pointed at its head. On Saturday you get to pull the trigger. Are you feeling lucky, punk?