6 Games people play

Aside, Card, CardGrid, LinkCard, Steps, Tabs, TabItem, Badge, } from '@astrojs/starlight/components';

Core idea

Every commercial game of chance is engineered around one number: the house edge — the fraction of every wager the operator keeps on average. Players see prizes, payouts and near-misses; the operator sees a guaranteed long-run skim that the Law of Large Numbers converts into reliable profit. Probability gives the player exactly one honest weapon: the ability to compute the expected value of any bet and compare it against the cost. Most casino games yield a negative expected value, and a lottery ticket is the worst-priced bet of all. Yet people still play, rationally, because money does not translate one-for-one into satisfaction. The topic knits these threads together: the math of independent and dependent games, the arithmetic of lottery prizes, and the utility theory that explains why a -50 percent return can still be worth a pound.

Author's argument: Probability does not tell you whether to gamble. It tells you what you are paying for and how much. Whether the entertainment, the dream, or the rush is worth the price is a question only utility can answer.

Why it matters

Gambling is the cleanest laboratory for applied probability outside of insurance. The outcomes are well-defined, the payouts are public, and the operator's edge is a fixed parameter — so any disagreement between intuition and arithmetic is the player's misunderstanding, not the model's. Mastering the casino case study trains a habit you can carry into every other risky decision: separate the price of a wager (expected loss) from the shape of its outcomes (variance and skew), and then ask whether your personal utility function values the shape enough to absorb the price. That single move dissolves most "to play or not to play" arguments — including ones that have nothing to do with cards or dice.

How it shows up beyond casinos

Insurance, options trading, R&D portfolios, lawsuit settlements, and even career choices are all instances of paying a known cost for a probabilistic payoff with unfavourable expected value. Each one is a casino in disguise. Knowing the lottery is "a tax on people bad at math" is glib; knowing why a rational person might still buy a ticket is what makes the framework portable.

Key takeaways

Mental model — games sorted by skill and house edge

Copy sourceZoom

Casino mathematics — what you are paying for

Expected value, the only number that matters

For any bet with outcomes $x_1, x_2, \ldots$ occurring with probabilities $p_1, p_2, \ldots$, the expected value is $E = \sum p_i x_i$. The house edge is the expected loss expressed as a fraction of your stake. On a European roulette wheel, a £1 bet on red wins £1 with probability 18/37 and loses £1 with probability 19/37, giving $E = 18/37 - 19/37 = -1/37 \approx -£0.027$. That is the famous 2.7 percent edge. The American wheel adds a double-zero pocket; the same calculation now gives $E = -2/38 \approx -£0.053$, almost double the cost. Two pockets, one wheel, double the long-run skim.

Independent versus dependent games

Roulette spins are independent: the wheel has no memory. If red has come up twelve times in a row, the chance of red on spin thirteen is still 18/37. This is where the Gambler's Fallacy lives — the conviction that "black is due" — and it is the engine of doubling systems like the Martingale. Doubling after a loss eventually wins back one unit, but the bet sizes explode and the table limit eats the strategy. No sequence of independent bets with negative expected value can be combined into one with positive expected value. Period.

Blackjack is dependent: each card dealt changes the composition of the remaining deck. When the residue is rich in tens and aces, the player's edge over the dealer rises. Card counting tracks this drift with a running tally — high cards depleted, edge falls; high cards still in the shoe, edge rises. A disciplined counter can convert a 0.5 percent house edge into roughly a 1 percent player edge. This is the only widely known way to beat a casino game without cheating, and casinos respond with continuous shufflers, multiple decks, and the right to refuse service.

Why slots are the worst seat in the house

Slot machines are independent like roulette but typically run a much larger edge — 5 to 15 percent, hidden by complex paytables and the dopamine of frequent small wins. The arithmetic does not care how the loss is delivered; over a long evening, a 10 percent edge on £400 of total spin will cost £40 on average.

Lotteries — paying for a dream

The 50 percent rule

A 6/49 lottery returns roughly half of takings as prize money. The other half pays operations, retailer commissions, taxes and good causes. A £1 ticket therefore has an expected value around £0.50 — a -50 percent return on every pound. The probability of matching all six numbers is one in approximately 13.98 million. To make that tangible, Haigh notes the chance of a randomly chosen 40-year-old man dying within the next 35 minutes is comparable.

You cannot raise the chance — only the size

Because spheres do not remember, no choice of numbers is luckier than another. But other players are not random: they cluster on low numbers (birthdays), on the centre of the ticket, on diagonals and patterns. When the winning numbers land in a popular region the prize is split many ways; when they land in an unpopular region a single winner takes the whole pot. A rational lottery player therefore aims for high numbers (sum at least 177), spread into a few clusters, with a bias toward the ticket's edge — not to win more often, but to share less when they do. Avoid obviously clever picks like 1-2-3-4-5-6; thousands of people have the same idea, and in September 2009 eighteen Bulgarian tickets had to share a jackpot whose drawn numbers happened to repeat the previous week's.

Decision under risk — why anyone plays

Expected value is not utility

The expected value of a £1 lottery ticket is £0.50, yet millions of rational people buy one every week. The resolution is utility: the satisfaction money provides is not linear in the amount. Losing a single pound subtracts a negligible amount of utility from a working adult's life. Winning fourteen million pounds adds an enormous amount — far more than fourteen million times the utility of a single pound. Multiply the (tiny) probability of the (huge) utility, add the (large) probability of (near-zero) lost utility, and the sum can credibly come out positive. The other ingredient, often overlooked, is the dream — the entertainment value of imagining the win between buying the ticket and the draw. Haigh frames the trade plainly: 50p of expected return, plus 50p of fantasy, equals £1.

Risk aversion and risk seeking in the same head

Classical utility theory describes most people as risk-averse for gains: a guaranteed £100 is preferred to a 50 percent chance of £200. The same person will pay an insurance premium above its expected payout to avoid a rare large loss. Yet that same person will buy a lottery ticket whose expected value is sharply negative. There is no contradiction — only a utility curve that is concave over the bulk of the distribution (smoothing ordinary gains and losses) but convex at the tail (life-changing windfalls are valued out of proportion). The shape of the curve is the explanation; expected value alone never could be.

When the framing matters

The same person who will not bet £100 on a coin flip will happily buy £100 worth of lottery tickets over a year — because each individual stake is small enough to fall into the entertainment budget rather than the savings budget. Mental accounting is part of utility too: people partition money into pots with different risk tolerances.

Practical application — the four-question gambling audit

Example — the £2 scratch card

A £2 scratch card advertises a top prize of £100,000 with overall odds of "1 in 4 winning". Most wins are £2 or £5. Suppose the prize structure works out to an expected payout of £1.30 per card. The house edge is therefore 35 percent — worse than roulette, better than a lottery.

A risk-averse buyer asks: do I value the 70p of expected loss less than the brief thrill of scratching, plus the dream of the £100,000? If yes, buy. If no, do not. The arithmetic does not forbid the purchase — it simply prices the entertainment at 70p per card. Buy one knowing the price; do not buy a hundred thinking the next one is "due".

Caveats

Related material