Alex's Adventures in Numberland - Alex Bellos [133]
The flexibility of slot design is that you can vary the symbols, the winning combinations and the payouts to make very different games. Game A is a ‘cherry dribbler’ – a machine that pays out frequently, but in small amounts. Almost half of the total payout money is accounted for in payouts of just $4. By contrast, in Game B only a third of the payout money goes on $4 payouts, leaving much more money to be won in the larger jackpots. Game A is what is called a low-volatility game, while Game B is high-volatility – you hit a winning combination less often, but the chances of a big win are greater. The higher the volatility, the more short-term risk there is for the slot operator.
Some gamblers prefer low-volatility slots, while others prefer high. The game designer’s chief role is to make sure the machine pays out just enough for the gambler to want to continue playing – because the more someone plays, on average, the more he or she will lose. High volatility generates more excitement – especially in a casino, where machines hitting jackpots draw attention by erupting in spine-tingling son et lumière. Designing a good game, however, is not just about sophisticated graphics, colourful sounds and entertaining video narratives – it’s also about getting the underlying probabilities just right. I asked Baerlocher whether by playing around with volatility it was possible to design a low-payback machine that was more attractive to gamblers than a higher-payback machine. ‘My colleague and I spent over a year mapping things out and writing down some formulas and we came up with a method of hiding what the true payback percentage is,’ he said. ‘We’re now hearing from some casinos that they are running lower-payback machines and that the players can’t really pick up on it. It was a big challenge.’
I asked if this wasn’t a touch unethical.
‘It’s something that’s necessary,’ he replied. ‘We want the players to still enjoy it but we need to make sure that our customers make money.’
Baerlocher’s pay tables are helpful not just in order to understand the inner constitutions of one-armed bandits; they are also illustrative in explaining how the insurance industry works. Insurance is very much like playing the slots. Both are systems built on probability theory in which the losses of almost everyone pay for the winnings of a few. And both can be fantastically profitable for those controlling the payback percentages.
An insurance premium is no different from a gamble. You are betting on the chance that, for example, your house will be burgled. If your house is burgled, you receive a payout, which is the reimbursement for what was stolen. If your house isn’t burgled you, of course, receive nothing. The actuary at the insurance company behaves exactly like Anthony Baerlocher at IGT. He knows how much he wants to pay back to customers overall. He knows the probability for each payout event (a burglary, a fire, serious illness, etc.), so he works out how much his payouts should be per event so that the sum of expected contributions equals the total payback amount. Although compiling insurance tables is vastly more complicated than creating slot machines, the principle is the same. Since insurance companies pay out less than they receive in premiums, their payout percentage is less than 100 percent. Buying an insurance policy is a negative-expectation bet and, as such, it is a bad gamble.
So why do people take out insurance if it’s such a bad deal? The difference between insurance and gambling in casinos is that in casinos you are (hopefully) gambling with money you can afford to lose. With insurance, however, you are gambling to protect something you cannot afford to lose. While you will inevitably lose small amounts of money (the