Andrew Leigh

Speaker Key:

DW             David Walsh

AL               Andrew Leigh

 

DW             My experience of gambling taught me to understand that the wealth and prestige that I grew to possess had very little to do with my talent. I didn't gain wisdom by standing on the shoulders of giants. I gained wisdom by serving the crowd.

AL               My name is Andrew Leigh, and welcome to The Good Life, a politics free podcast about living a happy, healthy and ethical life. In this podcast, we seek out wise men and women who have lessons to teach us about living life to the full with humour, pleasure, meaning in life. We chat with musicians and athletes, CEOs and carers, about making the most of this one precious life. If you like this podcast, please take a moment to tell your friends or rate us on Apple podcasts. Now, sit back and enjoy the conversation.

                   David Walsh contains multitudes. Australia's most successful professional gambler. He built and stocked Australia's most successful museum, MONA in Hobart. His autobiography, Bone of Fact, jumps from Bayes’ theorem to bullfighting and includes an extensive discussion of his vegetarianism, atheism and past lovers. But for an autodidact, David's also surprisingly self-effacing. In reply to someone who tells him that MONA is the world's greatest museum, it emerged that it's not even the greatest museum starting with him.

                   He places a strong emphasis on luck, which led me to ask him to launch my book, The Luck of Politics in Hobart in 2015. I think he's one of the most interesting people I've ever met. I hope that by the end of this conversation, you'll think so too. David, welcome to The Good Life podcast.

DW             Hopefully, by the end of this conversation, I will have disabused you of the notion that I’m one of the most interesting people you’ve ever met. And not only is my museum not one of the best in the world and it starts with the letter M, also I'm probably not the most successful gambler in Australia, or at least I'm not the most successful Australian gambler. I know that to be so because the enterprise that I set up, I have 28% and there's another guy that has 29. Arguably, he's doing better than me.

AL               So, that's your friend, Zeljko?

DW             That is correct.

AL               You think he’s a better gambler than you?

DW             No, I think he's a more successful gambler than me.

AL               What makes him more successful?

DW             He’s got a bigger share. What makes him more successful? Success is only slightly correlated with talent. And he had some capital when I had the germ of an idea, actually I had the gem of several ideas, one of which turned out to be okay. But capital buys rates of return that are not congruent with the value of labour, as economists are well aware. And he had the capital and I didn't, so he gets the bigger share.

AL               Hence the success.

DW             I’m measuring success in the way that most… You might be circumscribing a different philosophy than simply outcomes. But the most successful gambler in Australia might be someone that loses a few bucks every Friday but enjoys that particular passion immensely.

AL               But many would use the yardstick of dollars.

DW             And I did.

AL               I would normally start an interview by talking about someone's childhood. But I understand from [unclear] I've read from Kirsha that what most relaxes you is talking about numbers, so I thought we would delve…

DW             I’m pretty happy to talk about my childhood.

AL               We do.

DW             And in general, I'm sceptical about analyses that say, okay, this turned out okay for you. What's the wellspring of it? I think the genesis of most of us is randomness, so I guess Kirsha’s on the ball there. I'm happy to talk about my childhood and I'm happy to talk about numbers.

AL               When you ask people to pick a number between one and ten, a surprising number of them choose seven. Why seven?

DW             They won’t choose an even number most likely. Because you phrase it the way you did, they're probably going to eliminate one and ten because you would have to be much more definite, like it is not certain that that was included in your categorisation. They’re trying to be random, so they eliminate three and nine because they seem too stylistically located, like they don't feel like they're random. They won't pick five. Can you tell me why they won't pick five?

                   They won't pick five because it seems non-random because of the symmetry because we use base ten, five. There are another bunch of components, they inadvertently try to choose primes because numbers that have factors seem less random, or numbers that have obvious factors. And small numbers, you typically know whether they're prime or not. Seven results, I don't know what the percentage is, you obviously do. What is the percentage?

AL               My understanding is, in some studies, it's up to half that do choose seven.

DW             I would have thought it would be half. I suspect you would get even more men choosing seven than women because women… Interestingly enough, and I'll throw this in, because there are people out there that are chasing a million dollars on a $2 million to one event because they have a nonlinear utility value function so they buy a lottery ticket. If you buy a lottery ticket and you do what everyone else does, you'll do things like choosing birthdays, which means that there's a preponderance of numbers smaller than 12.

                   And then a preponderance of numbers between one and 31, for obvious reasons. You should choose a few numbers that are greater than 31 because then when you do get a dividend, you'll get a better dividend than you would have gotten had you chose birthdays, because they can't be birthdays. But interestingly enough, if you choose all numbers bigger than 31, that doesn't pay very well either because there are a few people that are aware that numbers bigger than 31 are good.  What you got to do is choose a few birthdays and a few larger numbers.

                   But we're just not very good at being random number engines. In fact, I think it might have been Poincaré, he did some analysis of roulette numbers and Monte Carlo that reporters were making them up. The roulette numbers of Monte Carlo used to be printed in the newspaper and reporters were not doing their job in making them up. When he did the analysis, he found that they’re extremely non-random because humans, like essentially inventing the science of this analysis, and humans are just terrible at generating random numbers.

AL               And this is the kind of imperfection that you're exploiting then.

DW             I'll get back to that. Computers will always beat humans at, what is it called, paper?

AL               Rock, paper, scissors.

DW             Yes, that’s it, rock, paper, scissors. Because computers can be random but humans can't. In fact, they take the one that if they had a loser last time, they’re way less likely to take it than they were before. Sorry, I interrupted your question.

AL               I was just asking if this is the sort of imperfection that you're exploiting in order to get an edge in gambling, isn't it?

DW             Yes. In general, casino games, in particular blackjack, depends on the fact that cards have a memory. Once you take them out of the deck, the remaining cards are not random so that's not particularly sophisticated. More interesting is horse racing because we've done a lot of analysis in horse racing, more than 30 years now. It took me years before I was able to generate a model that was more accurate than the public, which incidentally is a rousing endorsement of a form of democracy.

                   It turns out, it doesn't matter how much analysis you do, doesn't matter how sophisticated how much you know about regression and logistic regression. You end up with a model that's inferior to the public, unless you include the public as one or more factors in the model. What happens is, what you end up with, if you’re actually going to win is a Bayesian [?] model.

                   I'm gesturing which is probably users, that assumes that the public is right, and then finds the biases that they introduce based on the things that in the psychology literature are called cognitive biases, the things… I'll give you an example, hot hand bias. There's a belief, if you watch basketball and these days everyone does, that someone that's shooting, had a great three point percentage through the game is more likely to shoot a three-pointer than he was or she was before. But in fact, the lifetime average is the predictor of the next outcome.

                   Hot hand bias is foraging behaviour. If you go back to where you found food in your ancestral environment, it was useful. In fact, bees do it too. It's built into our evolutionary history. It was once useful, this cognitive bias, this mental shortcut. But now, in certain applications, it leads us to misadventure. We can get the odds mostly right. But when there's been a particular sequence, a jockey’s won a bunch of races, we might think that it's more likely to win the next race, then she's more likely to win the next race.

AL               We overweight the last race and considering they’re just going to do it in the next race.

DW             We overweight trends. In fact, in the stock market, we underweight trends, and I can explain that if anyone's interested. But typically, that's just one of hundreds of them. The most interesting one, and you’d know about this from your economic background, is what's called hyperbolic discounting.

                   If I said to you, and this wouldn't work on you, but if there was an amount of money that was going to make a difference to you, and I said, I'll give you $100 today or $200 tomorrow. You would almost always say, I’ll wait a day. Hang on, I've got this wrong. [Unclear], do you know?

AL               There’s a special feature about today and tomorrow.

DW             $100 today or $101 tomorrow, you'll take the 100 today. But if I defer them both a year, one a year and one a year in a day, you'll take the 101 because you've only got to wait a day longer. And hyperbolic discounting is such that it doesn't seem relevant. But we know that that bias is not cognitively sound if you’re using linear utility because in one year it'll be exactly the same outcome. Hyperbolic discounting is very real and bets in the far future behave in different ways.

                   People who take bets, like having to pick four winners behave differently than if they had the opportunity… For example, if they've taken three winners, they're much more likely to take a longshot than they would have been if they had to pick the outcome of this event. And in general, they assert longshot biases. People prefer to… If you back a favourite in a horse race, you might lose two or 3%. If you back a longshot, you might lose 50% or 40%. That's preferred. And arguably, it's preferred for two reasons.

                   One is boasting rights, if you happen to get a winner, it's very useful. And I believe that plays a big part because men exert this bias more than women. Women certainly seem to need less boasting rights. And the other one is, the thing I talked about before, nonlinear utility. $100 might be worth a lot more than 200 times $1 to you because the dollar has no purchasing or no apparent purchasing value, but $100 does.

AL               Which is why the longshot bias goes up in the last race of the day, right?

DW             It used to. It doesn't anymore because of people like us, presumably, or because people like us have taught the public.

AL               During the stock market is there [overtalking].

DW             What might be happening is we're forcing inefficient players out of the market.

AL               Some of the other things I've seen, you talk about bankroll exploiting the general punters tendency to overweight the amount of weight that the horse is carrying and the weight of probably the jockey.

DW             In Australia, that is a real feature of horse racing, because there was a gambling guru, Don Scott, who based his system around weight. He was published extensively and a lot of people incorporated this ideal into their behaviour. He was himself wrong about it as a feature, it was never as relevant as he thought it was. But once you misdirect the wisdom of the crowd, they're all over playing this to the extent that… I think this feature, as far as I know, doesn't exist in many markets.

                   We don't tend to isolate out features now. But some of them are so obvious, like huge negative coefficients on weight. If you include the public, you get huge positive coefficients on weight, if you don't include the public. Meaning, that weight is relevant to the outcome but it is over discounted by the weighted numbers.

AL               We've dived hard into the way in which you might find an edge in gambling. When you take a step back and think about how you first got interested in gambling, what role did going along to greyhound races with your dad play in this? And indeed, [overtalking]?

DW             I didn't spend a lot of time with my dad, my parents were separated, but extremely working class. Dad was engaged in working class activities and greyhound training is certainly almost an archetype of such behaviours. And I enjoyed Sam's existed, probably at the time I believed he made money out of it. But now in retrospect, I know that dad, he’s very low probability. He didn't really bet much. And if he did bet, it’d be $1 or $2. But he liked to train greyhounds. A lot of his mates trained greyhounds.

                   The greyhound racing, he was on Saturday night, I used to do a paper round and then I'd get on the bus and go meet dad, I think, dogs and it was a bonding session. As I say, I didn't see him much. And I enjoyed it. I can't say that I extracted much principle from the dogs at the time that I thought, simple things like adding up the bookmakers odds, they can tell you how big a disadvantage you're going to have at random if you bet. Those things didn't occur to me when I was very interested in numbers.

                   But that experience started to occur to me when I met people at uni and read books. The university in Hobart where we are and where I live is just across the road from the casino. It does seem to have created quite a lot of serious gamblers, people wander across the road. I guess I was one of those, but I'd already been given some info that it could be beaten by a couple of guys doing physics. And I vaguely knew this guy, [unclear] from playing table tennis. Questions were asked and…

                   I had something of a computer science background and something of a maths background and not particularly deep skilled either, but being the Swiss Army knife made me a useful commodity to these guys that are interested in gambling. They asked me some questions, I got them wrong. I got a feature of one wrong. And I realised shortly after that I’d got it wrong and that made me put in some serious time. And I went back to [unclear] and I said, look, I got this wrong. Here’s how you should do it.

                   And he said, no, I was just testing you, which really pissed me off as you can imagine. Somehow now I got ego in it, and there was that ego in it that got me interested and also his compelling personality. But that was about casino gaming. But it made me think about the possibility of extracting an edge at horse racing. One of our blackjack colleagues had also looked at dogs and greyhound racing. And essentially thought that maybe there's a difference in the embedded wisdom of bookmakers compared to the taught [?], particularly in small pools.

                   It was a valid observation that comprises what stock market traders call a technical model. And there are other stock market traders that have fundamental models that essentially look at the features of the market, like the underlying value of shares. But you don't often see in the public literature people using both fundamental and technical models. Why that is, is beyond me weird. And when I put the technical features with the fundamental features together, is basically…

                   When I didn't just look at the trend of the money, of the public, I also looked at the reality of the underlying factors, I found that they were both influential in forecasting the outcome. You don't need to forecast the outcome. A lot of people aren't aware of this. What you need to do is simply get odds that are better than the public odds. And I use a die analogy quite often. If you roll a die, it doesn't matter what number you bet on. If you get ten to one, it’s a fair die, you'll make money. If you get four to one, you won't make money. Every economist knows that.

                   The one sitting opposite me on the microphone definitely knows that. But it isn't universally known and it probably should be. It's basically turning a business transaction, where you enter a business transaction you're buying a car because you might think that the car plus the status embedded in buying the car is worth more than the cost of the car. All we're doing is reducing a series to a single transaction.

AL               Let's go to that edge that you had in your early gambling. I understand that when you're walking into a race-point in those early days, you're basically playing blackjack. What does a good card counting system like wonging give you? What percentage are you getting and what does that mean you've got a stake in order to…?

DW             Wonging only bets when you got an advantage. Typically, your edge goes up by about a half a percent for each true count of one. And a true count is basically a measure of the numbers and the number of ten and aces above the random number of ten and aces remaining in the pack, and usually it's multiple decks of cards. An edge, 1% might happen one hand in five, an edge of 2%, one hand in 25. Typically, card counting alone, you win about one and a quarter, one and a half percent on turnover.

AL               Per hour?

DW             No, per hand. If you bet $1, you make a cent. Typically, in a reasonably fast game, and you do seek those out, you make about one top bet an hour. If your top is $100, you make about $100 an hour. And if you introduce a few other features like shuffle tracking, or if you find games that make other errors or barred [?] particularly good rules, you can sometimes make up to two top bets an hour, so betting $100. And the top bet is basically, you need about arguably 300 top bets an hour, so you might need a 20 or $30,000 bankroll, a bet of $100.

                   Playing $100 an hour, assuming you can play 30 hours a week, you can double your bankroll in ten weeks. That's pretty attractive. There aren't many, so you can turn $1,000 into 2,000 and then into 4,000 and 8,000. Assuming you're not living off it, but I mainly was which is a big mistake, which is why [unclear] had a bankroll and I didn't. But from growth of small capital, I still know of no more reliable way of turning $1,000 into 10,000. Or these days, it might be 5,000 into 50 because the value of the money has been diluted.

AL               Blackjack’s the early start. And then one of your…

DW             It wasn't meant to be. We thought it was our career. I didn't think it was our career, but we thought it was going to make us rich. And it has made people rich. But what happens is, casinos aren’t great admirers of winners. They tend to have an attitude which goes something like, losers only welcome. What you're going to try to do is convince them that you have the characteristics that losing buyers have. Say, for example, if you're betting hundreds of dollars, or as it became tens of thousands of dollars, you got to look like someone that can afford $10,000.

                   And I categorically didn't. A Rolex might be a good start and a backstory, making them believe, for example, that your drug dealer works very well for young people, because why else would poorly dressed young people have money. The idea is not to be poorly dressed, it's to be suave. And typically, other people than me can carry it off better. We eventually started using strategies. And it is an arms race. I haven't played for years and I know there are strategies out there that are a lot better than any we employed both on the casino side and on our side.

                   We played a thing called gorilla big player and big player, where you might swamp all the tables in a casino with a card counter. It's just flat betting. It's betting, say, $100. And this might happen in a Vegas casino when there's a big boxing match on, so there are lots of people betting a lot of money. On a table with a 10,000 maximum, you'd have someone on every table giving signals to someone that’s just walking around betting the maximum.

                   Presumably, either they've had a few drinks or they're a good actor and can pretend they’ve had a few drinks. And they’re spilling a drink and they've got an attractive lady. And it always has to be a man because women don't behave this way. And this person has to blend in. They're walking around big-noting themselves, showing off, betting 10,000. It appears to be a random process but they're getting signals from people. There's no growth in bets, which is a big indicator of card counting.

                   If you bet bigger towards the end of the shoe, then the consigner [?] can do some quick analysis from, they can do it live or they can do it from the movies that they're making. And quickly, they'll find a correlation in the outcomes or in the distribution of cards that remain, so they'll catch you but they won't if you… I'm sure casinos have become aware of these tactics, but they were not aware of them 30 years ago. And they introduced things like shuffling machines that you feed the cards straight back in after you’ve dealt them.

                   The memory characteristic of blackjack where cards that are out can’t be redealt is eliminated. And I don't know how, but I do know that there's a team of South African gamblers that have figured out a way of leading the shuffling machines. These are interesting things that… And whatever they do to fix the flaw, someone will find a flaw. It is not pure mathematics and pure mathematician, no, I don’t mean pure.

                   I mean mathematicians, statisticians, economists make terrible gamblers because they expect the features in a casino to reflect the real world and they don’t. It's typically, we've successfully played roulette. Because as implemented, it's a random game. No, sorry. As described, the theory of roulette is that it's random. But it isn't often characterised in casinos in ways that are perfectly random.

AL               You have your story of having found the flawed roulette wheel, right?

DW             We found lots of flaws in lots of roulette wheels. But the one I've identified, written about or maybe I’ve tried.

AL               Red 27.

DW             Yes. 27 red is a flaw on as far as I know every American bingo parlour wheel in the world, which means the casinos don't often put them out. But every now and again, for some reason they… Like a few years ago, some appeared in Puerto Rico.

AL               The ball’s about 20% more likely to land on 27 red, is that right? It’s not a big edge [overtalking]?

DW             Between 20 and 30% advantage is a massive edge. Think about it in terms of you’re betting on that number straight out, you can bet, say, $100, $200, you get up to 160 spins an hour. That's 32,000 turnover. At 20%, you're making $6,000 an hour. You can also bet the splits, which halves your edge, and you can bet the corners themselves. Anyway, in that scenario, which was exactly the scenario we found ourselves, we can make about $15,000 an hour. And we did.

                   But people have beaten roulette in a predictive way where they use a combination of technology and strategy to forecast what numbers are going to come up to get much greater than 15, often greater than 100 percentages. This happened for example, at the Ritz in London and they won £1.3 million in two nights. And then the casino refused to pay them. They just understandably protested. It was taken to court. The discriminating factor was found to be whether they influenced the game in any way or not, and they did not, so the money was given to them.

                   But using a technique called edge sorting. It is very rare for a deck of cards to be perfectly printed on the back and you've got a diamond pattern. Where that is cut by the edge of the card can sometimes be red and sometimes asymmetric. I would assume she's very smart. I met her but her strategy was incredible. She was claiming to be superstitious in telling the dealer to rotate cards in certain directions depending on the outcome.

                   But what she was doing was having aces rotated one way so that the next shoe she could read the back of the card until it was an ace. That gives you a 50% edge blackjack. She won £7 million, or was it $7 million US? It was $7 million US. And they didn't pay her. Again, it was contested. And the casino was able to keep their money because, again, the discriminating factor was whether you influenced the game. And the court ruled that she did. I think she was quite hard done by it because a lot of Asian players do a lot of things like that.

                   I think because the casino weaselled out on a loophole. But anyway, these are the sort of things that happen that enable you to beat gambling games, and they're not features that statisticians would typically deduce from an analysis of the game.

AL               Just before we leave gambling, I am interested in what someone who is not betting billions a year can learn from how you think about risk. One of the things you noted in your book, Bone of Fact, is that you should employ a barbell strategy, you should never bet so much as to risk financial ruin. Can you talk a bit about that?

DW             Yes. I think there's only one characteristic of successful investors that they all share. When we do analysis, when we see a magazine interview, they interview some ultra-successful Bill Gates type. And they only look at the output cohort, they don't look at the input cohort. And I always use tennis as an example. There is no point in interviewing Roger Federer because all of the things he's done to make him good, the people that didn't become good also did. It might seem incredible, but most of his edge was a lucky retrospective outcome.

                   He was slightly better than everyone else. And there's a feedback loop and he got better and better. Getting back to your question and looking at barbell strategies, as Taleb calls them, and I call them B-glass strategies.

AL               Because a B-glass is more fun than a barbell, right?

DW             And also, barbells are the same at each end. The first thing that successful investors do that is a discriminating factor based on it being different from inputs, and I didn't do this, is survival. I said before that you divide your 300 top bets when you're playing blackjack, that's because it reduces the risk of ruin. I did also say that when you've lost half your capital due to a negative fluctuation, assuming you've got an edge, and that will happen, you've got to bet half the size. Because the biggest thing that can happen to you is destruction.

                   The first thing you've got to avoid is ruin. And that applies to anything you do. And generically, a lot of people are arguing, most people with some expertise believe that global warming is real. It is absolutely irrelevant in terms of our response to it, whether it is 99.9%, 50% or 10% likely, we should still intervene because the cost is too great to contemplate. The fear of ruin, the four degree global warming will require trillions of dollars a year to be spent at a later date. It's easier to spend billions now than trillions later.

                   Risk of ruin, existential behaviour is fundamental to any activity. I find BASE jumping, for example, untenable in terms of it maybe generating 50% more thrill than parachuting, but tens of thousands of times more risk. And also, there's this other characteristic of some of these activities, they're addictive, so you keep doing them until they kill you, even if they're low risk. If you shoot up here once, you're going to have a good time and it's no risk to you at all.

                   That isn't the nature of the engagement, it's about 8% likely to change your behaviour in the future in a dramatically negative way that has a high risk of ruin. If you bet on the horse races, there's probably about one chance in 5,000 that it will destroy your life because you'll become addicted. But the first time you bet on a poker machine, there’s about a 5% chance that it's going to destroy your life. What's the worst decision I ever made? It was that first bet that I placed.

                   It made me very, very rich. But there was about one chance in 100 million of that happening. But there was one chance in 20 of me becoming an addicted gambler and ending up possibly with other addictive characteristics. In part answer to your question, if you've never placed a bet, the best decision you can make is never to place a bet. If you're ever going to place a bet, don’t do it on a poker machine because the addiction by design, clever bastards that built them, are building feedback mechanisms, and all sorts of other complications that I don't know about.

                   Because they've hired the smartest people in the world and they pay them offensive [?] amount of money. And they've got good outcomes to get the highest addiction rates straight up and to make it very difficult to beat that addiction if it arises. 5%, fear of risk of ruin, very high. Don’t take that initial bet. And if you are engaged in social gambling on an activity like horse racing or playing roulette, don’t cross over into the ones that are a lot more addictive.

                   But in terms of non-gambling outcomes, there are a lot of people, including me, that made really bad business decisions early on and got lucky. People argue, Andrew, talk to them. But their skill was simply… Maybe they've got an edge now because they had the chance to acquire it. But those initial early tosses of the coin, they often took risks that are untenable, I did. I overestimated my edge, and there was about a 20% chance that my horse race gambling resulted in a successful outcome. That's not a good position to be in.

                   Now, my contention would be that there are a lot of other people like me that were just as talented at gambling and figured out the same strategies as me, and made exactly the same mistakes as me but they got wiped out and I didn't. And then, of course, once I'm in most of the markets, people like me that are successful, then make the characteristics of the market more untenable and more deceptive. So that newcomers come in, they think they got an edge and they're less likely to have an edge.

                   And indeed, they're less likely to bet too big with the edge that they have, so they're more likely to blow up. 30 years ago, you can beat me, but now you can't. Because you'll come in and you'll find a way of getting an edge, you'll overestimate your edge, you'll bet too big or you'll have deceived yourself about whether you've got an edge in the first place and you'll blow it up. So, don't gamble. In terms of business decisions, do lots of things, repeated outcomes, open restaurants with part of your capital, don’t blow it up at once and seek asymmetric upside.

                   Uber, at least, for humans on low incomes is much better than a government unemployment policy, because it allows you to do something that will keep you alive while you take risks with part of your capital and it enables you to have an income. One of the best features of MONA that [unclear] from seeing the literature is that we employ about 150 people with fine arts degrees, that would otherwise not be doing anything in fine arts because they'd have to get a job.

                   Because you can't make a living in fine arts unless you pull off a million in one shot and become Damien Hirst. What's happening is these people are working one or two days, three days a week at MONA. And then in their downtime, they're making stuff. And some of that stuff's really great. And every now and again, the forces of talent and chance will in unison propel them to a successful career. But in the meantime, they can do their creative endeavours, their artisanal endeavours, the things that define them as a human being without risk of ruin.

                   And also, it keeps people in Tasmania that otherwise would not be. It keeps artisanship and talent and endeavour and creativity and honour, particularly honour in the community because people that make things necessarily have honour. Or they have honour because it's built into us biologically to endeavour, to strive. But normally we have communities that enable us to take risk. Those features of communities have been extracted.

                   And I believe that those features are the litmus test in the quality of if you can take risk in such a way that if that risk were to succeed would benefit the community, then that is a marker of a well-measured community. The US produces a great deal of the world’s endeavour and a great deal of its innovation. But the downside cost is too great because those that attempt to innovate and fail aren’t protected, there's no safety net and therefore [unclear] act will be a lot less.

AL               In the origin story of MONA, I understand, goes back to the early 1990s when you're in South Africa trying to work out how to get gambling winnings out of the country and go for the strategy of exporting a piece of antiquities, right?

DW             The origin story of everything is multiple path, except when there's no risk. What we're talking about in an origin story is an unlikely outcome that had a lot of risks or ruin. One place that I can pinpoint easily and have done and therefore it's available in the literature is, my mate and I were gambling in South Africa. He was playing with my money. He won what I now think is $18,000, but at times I’ve reported it as $20,000. But I think I did the translation of the rand at the wrong time.

                   But anyway, having reasserted, I believe it was $18,000, I had walked past a gallery in Sandton, a suburb of Johannesburg, and I'd seen this Nigerian palace door made by an artist I’d never heard of, but it turns out he’s quite well-known. His name is Aerogun. It was made, then about 100 years ago. Now, about 130. I liked the look of it. I thought it was pretty cute. And when Patrick tries to leave with the money, he can't. He rings me up and he says, they won’t let me take out more money than I brought in.

                   And I say, I’ll come if he can take that out. And he turns around and he says, sure, they say that’s okay. Rather weird. But anyway, I think I rang up the gallery and I said, how much is that work? They said, US dollars. That's what's counting on here. They said $22,000 US. And that might be the discrepancy between the 18 and the 20. I said, would you take $18,000 US? And they said yes. And I became an art collector.

                   Another outcome of this was it turned out that this palace door was a very important panel of a double door, the other half of which is in the British museum collection. The first work of art I bought turned out to be worth a lot more than I paid for it. And that non-statistical outcome convinces you, you know what you're doing. Now I think I know about art so I'm buying art.

AL               Then Roman coins [overtalking]?

DW             Actually, I had purchased coins beforehand. I bought them but not at any level. But then I got serious. Coins, Egyptian artefacts.

AL               So then in 2001, you open the Moorilla Museum of Antiquities and it's a raging success, right?

DW             No, it was a raging failure. It was part of a winery. I never really intended anyone to visit it. About 10,000 people, therefore, a few 100 people a day visited. A lot of them liked that. We did no publicity at all and they were turning up for a winery. One of the things that makes people have a valid experience is to exceed their expectation. And because they had no expectation at all, it wasn't hard to exceed it.

                   It becomes harder as MONA, this successor to the Moorilla Museum of Antiquities, becomes more successful, people are aware of it and have an expectation. But that wasn't the case. But anyway, having built this thing that no one came to, I declared it a raging success and moved on. But the most interesting characteristic of that in terms of what happened later was that it looked like every museum in the world.

                   I had no interest in the way museums were created. No museological background. The people I hired to do stuff for me also had no museum background. And yet, I build a museum that looked like everything else. And that was interesting to me.

AL               How much did MONA's success depend on having tried something that in your view wasn't a success [overtalking]?

DW             I wouldn't have even contemplated doing it again. As I say, it did for those that came exceed their expectations. And I enjoyed it personally. And I only built it because I realised that it was going to cost not much more to open my collection to the public than it was to appropriately warehouse it, to build a climate controlled endeavour, control the humidity. There wasn't a big reason not to open it to the public.

                   Although it was certainly not a commercial success and it didn't attract a lot of visitors, the most interesting thing about it to me was that it looked like every other museum and I didn't intend it to. I started doing some thinking about why museums look like they do. There's a museum theory called the white cube, and it says that we try to present art in a neutral matrix so that it's most objectively presented to the viewer. After about five minutes, I decided that that was nonsense.

AL               You went for the nightclub aesthetic instead?

DW             No. I decided it wasn't working. That there's a huge amount of subjectivity in the engagement of art. In fact, if you can’t see the same work of art twice, you won't have the same experience, your mood changes. And also, the artist didn't create it in this environment, you've already created an alien environment for the art by putting it in a museum. Very little art was ever intended to be in a museum. But certainly none of mine was because it was antiquities.

                   And I just broke the golden rule, I moved closer to the microphone as I got excited. I won't make that mistake again. The reason museums are white is because they have to put interpretive material on the wall and the easiest material to read is black text on a white background. And to make that work in the environment, that drives a white wall.

AL               How important was…?

DW             And that white wall, I wanted to get rid of. I came up with the idea of not having any interpretive material. And I developed a device called the O that shows you around the museum and knows where you are. And that freed up the design. And that meant I could mess with the lighting, for example. I could only light the art, which creates a very interesting atmosphere. You have to wall wash. You have to light the whole wall if people are going to read the interpretive material.

                   Anyway, I could have written a book called creative aesthetics in 21st century museology, but I didn't. I built the bloody thing instead.

AL               There's many things that strike you when you're in MONA. One is the absolute lack of pretension and that sense that I'm engaging with artworks through the lens of somebody who takes a regular, every person [overtalking].

DW             Can I comment on that because that's very important?

AL               Sure.

DW             State museums are essentially temples to the value of the state. They're not about you making your own call. They're enhancing the status of the state by telling you what good is. And they do it in a very specific way. You walk upstairs, you walk through grand portals. You’re made to feel small in the presence of greatness. It is autocratic rather than democratic. My museum, a lot of it is contemporary art. There is no way I can tell whether anything that I've got will be interesting to anyone in 50 years, except for the antiquities, except for the things that have history.

                   You can pretty well tell if you've got no other data about something that it'll survive into the future as long as it's already survived. When a storm hits the great pyramids, they don't fall down because many storms have hit them in the past. You can expect them having survived a few thousand years to survive a few thousand more. A house that’s just been built next to the pyramids, when a storm hits is much more likely to fall down. The characteristics of my art, the new stuff, I've got no idea if it has value.

                   Because of my statistics background, I can do the analysis and say, this is most likely to be worthless and there might be a few of these objects that become valuable. I can't predict which one they are since I have no idea whether anything I'm showing myself is good, it's very hard for me to assert a state position and try to indoctrinate you into my belief system. And in fact, I believe that [overtalking] emerges [overtalking] consistent application of doubt for multiple individuals. [Overtalking] theory they’re almost always wrong.

                   Scientific method, which is [?] essentially the creative application of doubt. They generate a consensus that is more likely to [overtalking]. In the same way that we’re likely messing up the atmosphere, if we had to carry around a little bit of atmosphere and watch it get polluted by our activities, we would probably behave a bit differently than what we do. I call this compartmentalisation.

                   And I'm railing against compartmentalisation. Typically, the cultural and often the biological expression of taboo is a way to enable us to avoid the knowledge of the consequences of our actions.

AL               [Overtalking].

DW             It looks like I'm railing against taboo, but most taboos make perfect sense. Not eating shit makes sense. Shit smells bad because it has got skatole in it. And it has skatole in it because tryptophan, one of the amino acids, decomposes to skatole. Now, we don't want to eat our own shit and others because it's more likely to cause digestive upset. You put it in a water supply and you eat it and occasionally it happens, you get a whole range of diseases.

                   Rather than having us make a choice, biology has embedded by dint of sieving the failures that died to make us have an aversion to shit. It's picked a compound that is in shit that is in very little else, skatole. But skatole itself is not dangerous. Actually, it can be in very large doses. But we have also have skatole in ice cream for flavouring. But it's just a biological marker that tells us to be disgusted. The taboo makes perfect sense. There are similar compounds in earwax, but not the same ones.

                   In general, there's a group of compounds called indoles that we respond to. Most taboos make perfect sense. I'm having a go at the taboos for which there is differential uptake. But rational behaviour is that behaviour as we talked about before, not that generates correct information, not that generates fact. It's that which keeps us alive, it’s that which keeps us in the gene pool. You and I are probably committed rationalists. I'm an atheist. I believe you're an atheist too. You probably don't want to talk about that.

                   But I might be wrong as well. And anyway, you're not being interviewed. Now I think there are good reasons that atheism is the most likely feature of our universe. What I mean is that the lack of a deity or the lack of hundreds of thousands of deities is a feature of our universe. But belief, for example, in Islam, if either parent is Islamic, then the children aren’t necessarily Islamic, and the parents have to convert. The feature of belief in Islam is propagation.

                   There were other religions like the Shakers that, and I’m onto something here that has to go beyond this, were a feature was that they didn't breed and they died out. Rationalists tend to have less children than believers. As far as biology goes, rationalism is a flawed philosophy because it will result in a decrease in population compared to those that believe. And I don't know exactly where I'm going in terms of the question with this.

AL               Let me take it in a particular direction, which is that I'm struck in your book by the generosity that you show towards those who you intellectually drew on from a strong faith tradition. Gregor Mendel, the geneticist, Thomas Bayes, the father of Bayes’ theorem that you used extensively.

DW             Bayes made me all my money, it'd be hard not to be a fan. The best one is George Lemaître. The Vatican Observatory was mainly established to get the date of Easter right and a bunch of other lunar calendar phase. But George Lemaître is the guy that postulated the Big Bang. A Catholic priest at the Catholic Observatory in the Vatican came up with this big idea. How cool is that? Charles Darwin was going to be a Protestant priest.

                   Now, there are a lot of smarter people than Charles Darwin, but there isn't anyone with more intellectual integrity because he completely undermined his own belief with his own discoveries. Now, I would not have the rigour to do that. I'm in a position where I don't have to. But if a big idea ever thrust itself in my face, I won’t present it in a well-argued way so that it can either be shot down. I’ll avoid it because I won't even know.

                   Mendel 30 years before, three people in the same year discovered his idea, had the rigor to present it even though it took him down a path which wasn't great for his belief system. A feature of a priest is, apart from a few confessions and maybe mass on a couple of days a week, they have a lot of time to meddle, they're messing around, they’ve time. They may back into what I said before. They've got the safety net so they've got the capacity to innovate.

                   Catholicism as broadly, not for the devotees but for the direct beneficiary is those that believe in some such way that causes them to go down to a clerical path. They have a lot of free time and the time to meddle and the time to think. And they're also well trained intellectually, particularly the Franciscans.

AL               They’ve got a built-in B-glass in your analogy.

DW             They've got it. They've got the end of the B-glass that frosts and bubbles.

AL               Let me throw a few final questions at you, David. What advice would you give to your teenage self?

DW             I can give advice because I have a background in a specific way. There are two ways this question can be phrased. One is if I know the outcome is going to be roughly what it is, in other words, whether the following path can be. But if it's my teenage self in the multiverse and the outcome is unspecified, I will definitely say don't gamble. And the other thing I'd say is don't try to be good at table tennis because the things that you do that you think you're good at, you're not good at and they’re going to mess up your life.

                   Do something I would seek now, in other words, I would advise myself to seek. But if I was a ten-year-old with a similar set of talents, I would have done what I should have done, except that I got distracted, which is, be very interested in what I was very interested in. And the two things for me as just slightly spectrum, weirdo were astronomy and computers. Astronomy would have been immensely satisfying and I probably could have made a living but I wouldn’t have been good.

                   The computing, it was already obvious, except that I deviated from the path that I was going to get a good job. Everyone's going to get a good job if you're… The idea is all you need to do is be mediocre. You don't need to be fabulous. Tennis, you have to be fabulous. A lot of endeavours, you look like you have to be fabulous, but all you had to be was fabulously lucky. [Unclear] was luck rather than talent.

AL               Does that inform how you parent as the father of three children?

DW             Yes, and don’t do as I do. I'm very fortunate that my luck was not just luck, but it was meta luck because it gave me a set of analytical tools that can give me the background knowledge of how important luck is and how unrepresented it is in our analyses of what success is. Societies shouldn't need luck. We should design our activities. The larger the population size, the less risk there should be. Global warming is a really bad idea. And to get that, we have built into our biology an in-group/out-group phenomenon.

                   We tend to respond well to our family and then more well to the community that we're embedded in, the religion that we believe, the colour of our skin, and we have to fight to overcome these beliefs. Martin Luther King was completely wrong when he said racism is something that has to be taught. And getting those things wrong messes up our chances of fixing them. And that was a bit of an aside that probably just blew my chance of anyone ever listening to me again.

                   But the things that bubble up from community level tend to be the things that we are most committed to that effect the most viable changes. There should be risk-taking at the individual level, but we can minimise that with good strategies. And then they bubble up to family and to local. And then at each level, the sum of all the statistical risks shouldn't be biased so that the risk is reduced. Because if bias to risk, as it bubbles up to larger populations is still the same, but risk that’s scattered around a distribution isn't the same.

AL               Which is the difference between gambling and investing in the stock market, right, uncorrelated risks versus correlated risks?

DW             Absolutely. The stock market is a crazy way to make a living even if you've got a winning strategy. Because, let's say, you had an edge over the market, from 1929, for example, to 1956, the market declined. Doesn't matter if you had an edge. Now there's a lot of billionaires in the stock market, but the vast majority of them just participated in… Shakespeare says, there is a tide in the affairs of men, which taken at the flood, leads on to fortune. It went up for everyone. They think they know what they're doing, but it could just as easily have gone down, and it will.

                   There will be black swans. There will be grey swans. And when it goes down, like it did in 2008, like it did in 1997, it goes down in a dramatic way and the whole populations get correlated out of the market. We need those that most benefit from the market to have risk in it so that they don't pile it up. And the hedge fund managers that have invested, for example, in their own hedge funds, don't ask what the strategy is, ask what the portfolio is.

                   And if the guy giving you advice isn't in the same portfolio, whether it's social… If he's telling you to go after redheads and he's not married to a redhead, don’t take his advice.

AL               What's something you used to believe but no longer do?

DW             God.

AL               Vegetarian?

DW             I used to believe I knew what I was doing. I'm here talking to you not because I think that everything I'm saying is likely to be right. I'm here because for one thing, I've been lucky enough so that you've asked me to talk to you and you're charming, so I said yes. And also, you do this very clever thing where you use hyperbolic discounting, you ask people to do things. If you ask people to do something tomorrow, they’ll say no. But you know full well that if you ask them to do [unclear], they'd say yes.

                   But anyway, I'm here talking to you. I had the great good fortune to be able to have my voice heard, but I don't think it should be heard as a gospel in any sense. It's just a little bubble in the [unclear] of conversation. I have a few little ideas and they need to be tested by the wisdom of the crowd, they need to be tested by the doubt of the scientific method. But I used to believe that I had insights, I used to believe that left, for example, was better than right.

                   I used to believe that the features of left that aren't necessarily congruent with each other, for example, a typical support of the mother’s right to choose, the right to abortion and the right to choose termination. The right of euthanasia were congruent with each other because they're aligned with the left. Now I believe that they should be explored as entities in and of themselves.

                   I used to subscribe to the idea of find someone that believes what you do but knows a bit more and then you don't have to think as much about the political sphere and you can get on with your life. Now I just don't think that at all. For me, it is a particular necessity if I want to live a whole life with some integrity to analyse the features of each individual belief system and decorrelate them as much as I can. Now, I know you're aligned with the left. I don't know what you believe about…

                   I would say that you're probably a moderate left. But there are some things where I'm radically left and there are some fiscal… I would say, there are huge problems with our tax situation because the only way of destroying inequity is through destroying wealth. And we reposition the tax burden on income in non-progressive ways when it should be… I think estate tax is an obvious way of repositioning, I believe in high tax regimes, I believe in asymmetric upside.

                   Endeavours that work need not to be scaled out of existence, because otherwise the upside is an asymmetric and the reason to engage in it isn't there. I think it's possible to structure a tax system. Like capital gains tax makes no sense at all to me because the wealthy don't need to consume. I make more money than you probably and… The tax from consumption should be repositioned to luxury. There might be administrative problems with that, but I don't care.

AL               When are you most happy?

DW             I'm about to be very happy. After I leave you, I'm going straight to Government House. My wife's about to become an Australian citizen. She's American, she will soon be Australian.

AL               Congratulations.

DW             And that'll make me very happy because it makes her very happy. We share a rationalist view of the world, but she's also a magical thinker because she thinks it creates beauty. I am most happy when someone exposes to me beauty that my worldview could not hitherto encompass.

AL               What's the most important thing you do to stay mentally and physically healthy? Is your vegetarianism partly to do with health or is that entirely an ethical decision?

DW             Entirely an ethical decision. I suspect that if I eat fish, I’ll be more healthy. But compared to the overall domain of outcomes, I'm okay because vegetarian is better than eating a lot of meat. But probably eating a little bit of meat is ideal. In fact, probably eating a little bit of everything. Our diet is very focused on a few things and eating a lot more things might be better, certainly in our ancestral environment, we certainly did. And there's no reason to think that apart from a few modifications for things like wheat and milk that we've changed dramatically.

                   About my mental health, I have a history, in my family, of mental illness. It's something I have to be aware of. I tend to find it very easy to pursue rational materialist disciplines. And I tend to find that when I'm able to make my ingrained beliefs the things that make me feel like I'm a human being consistent with the things that I think are ethical, then I'm at my best. In general, I would say as a rule, we are most likely to behave in a way that's congruent with our ethics when we look inside ourselves.

                   Consciousness is to me an error-checking software. We don't need it to operate. We can just strive on [?] without it. Most of our decisions are made without it. And when we do tests, we find that we might have the capacity to intervene. We might have free want but we don't really have free will. What it seems to be is what Socrates termed as, he put it, a life unexamined is not worth living. I think there's looking inside yourself, and then having that become a mechanism that drives your behaviour is what integrity fundamentally is, and that gives you health.

                   There's a notion that, and it's hard to do. But if you can find it in yourself to realise that most other people, however disparate from their beliefs, probably are engaging in them with as much integrity as you are with yours, even though they're completely in conflict. It is quite likely, although I don't believe that Donald Trump, the outcomes that he's likely to generate a positive or universally positive. It’s quite possible, quite likely, in my opinion, that he is behaving consistently with his belief system and he believes what he's doing is right for the world.

                   And in particular, for America. I think it's very hard, but worthy, to try to realise that those whose beliefs are totally anathema to you, still are behaving in a way that is totally congruent with what they believe. They are not inherently evil. And I'm not sure that evil is inherently a real thing. Hitler was trying to create a utopia for those that shared the characteristics that he was bound to. Individuals trying to create utopias tend to produce very bad outcomes.

                   Communities trying to create utopias tend to produce bad but not as bad outcomes because individuals assert their biases more than groups do. But in general, utopian visions need to be squashed. But they need to be squashed by people that understand that those that are compelled by them believe in them.

AL               Which person or experience has most shaped your view of living an ethical life?

DW             My experience of gambling taught me to understand that the wealth and prestige that I grew to possess had very little to do with my talent. I didn't gain wisdom by standing on the shoulders of giants. I gained wisdom by serving the crowd. And that has been a tool that I've been able to use to analyse. And I'm certainly not suggesting that I'm ethical. Skin in the game is both a hugely positive thing when it drives behaviour that's good for a community, but it's very, very bad when it drives your behaviour, in the Donald Trump instance.

                   It gives you the opportunity to enjoy outcomes that satisfy you or that satisfy a need that aren't compelled by others. That's why we have laws against insider trading, but we also make ethical decisions like that. But gambling has given me the tool to least process when I'm doing it, and occasionally, perhaps redirect my behaviours, occasionally make better decisions.

AL               What a fascinating conclusion wrapping up where we started. David Walsh, professional gambler, art collector, polymath, thanks for speaking on The Good Life podcast today.

DW             Thank you. It's a pleasure.

AL               Thanks for listening to this week's episode of The Good Life. We love getting feedback, so please leave us a rating or a comment on Apple podcasts, formerly known as iTunes. Next week, I'll be back with another inspiring guest to discuss living a happier, healthier and more ethical life.