Bankroll Management and Risk of Ruin

Most players lose because of bet size, not card play. Edge decides where you end up eventually; bet size decides whether you are still at the table when eventually arrives.

The gambler's ruin problem

Imagine a coin flip that pays even money and wins 51% of the time. That is a better edge than any casino has over you. Now bet your entire bankroll on one flip: you go broke 49% of the time, immediately, despite holding the advantage. Bet 1% of your bankroll per flip instead, and the probability of ever going broke collapses towards nothing.

Same edge. Same game. Wildly different outcomes — driven entirely by sizing. This is the gambler's ruin problem, and it is the reason professional gamblers talk about units rather than dollars. The question is never "how much do I want to win?" It is "how many losing bets in a row can I absorb and still be playing?"

Unit sizing

Convert your bankroll into units before you sit down. A common conservative standard is 1% per bet, or 100 units. Aggressive is 2%. Anything above 5% per bet is not gambling with an edge, it is a coin flip with extra steps.

$5,000 bankroll, 1% units: $50 a hand. You can survive a 20-hand losing streak and still have 60% of your stake.
$5,000 bankroll, 5% units: $250 a hand. A 12-hand losing streak — which happens regularly — leaves you crippled.

Streaks are longer than intuition suggests. At roughly even odds, a run of eight consecutive losses appears about once every 250 hands. Over a long night you will see several. Size so that the streak is an annoyance, not an ending.

Variance is not the same as edge

Two games with the same house edge can feel completely different. Baccarat's banker bet grinds: small swings, slow bleed, few dramatic moments. Video poker with a big royal payout has a similar long-run return but most of it is concentrated in a jackpot you may never hit, which means the day-to-day experience is losing steadily while waiting for one enormous outlier.

High-variance games need bigger bankrolls to survive the same expected return. If you are playing something with a jackpot component, cut your unit size roughly in half compared to a flat even-money game. If you are playing slots, accept that your bankroll is a budget for entertainment rather than a working stake.

Stop conditions beat willpower

Decide before you start, because you will not decide well later:

Loss limit. A fixed number of units. When it is gone, the session is over. No reload, no "one more shoe."
Win target. Less obvious but equally important. A target converts a paper win into a real one. Casinos do not fear losing a hand; they rely on you staying seated.
Time limit. Fatigue degrades decisions long before it feels like it does. Every hand you play badly costs more than the edge does.

Notice that none of these are predictions. They are commitments made while calm, to be executed while not calm. That is the entire skill.

Systems that do not work

Martingale (double after every loss) turns a steady small edge against you into a large probability of small wins and a small probability of total ruin. Expected value is unchanged. Table limits guarantee the doubling sequence eventually fails at the worst possible moment.

D'Alembert, Fibonacci, Labouchère — all the same idea with better branding. No sequence of bet sizes can turn a negative-expectation game positive, because each bet's expectation is independent of the ones before it. Progression systems reshape the distribution of outcomes; they never move its mean.

"Due" numbers — the gambler's fallacy. A roulette wheel that has hit red twelve times running is exactly as likely to hit red again. The wheel has no memory, no conscience and no sense of fairness.

Bankroll inside a one-night run

Last Night at Lucky's compresses all of this into a single night with a $10M target, which changes the calculus in one important way: you cannot pure-grind your way there. A 0.5% edge on a modest stake will not produce ten million dollars before the clock and the heat meters run out. The run is explicitly designed to force escalation.

The practical shape of a good run looks like this:

Early: small units on low-edge games. Build a working stake, keep every meter cold, and use the quiet time to talk to patrons and open leads.
Middle: raise unit size as the bankroll grows, but keep the ratio constant. Rotate tables so no single pit sees a long streak.
Late: take the high-variance shots — back-room games, big sports bets, jackpot machines — once you can absorb losing them and once the story requires the money faster than the tables can produce it.

Contraband complicates the picture further. Items bought from patrons appreciate the longer you hold them, but holding raises Police Pressure every few hands. That is a leveraged position with a carry cost, and it should be sized like one: never tie up money you need to keep playing, and never hold a hot item into an already-high meter.