Casino games have a negative expected return — that's the entire business model. No bet sizing or progression system changes the math; the house edge applies to every wager regardless of how you sequence them. What bankroll management actually does is shape the distribution of outcomes within that expected-loss reality. Done right, it lets you play longer, take fewer ruinous sessions, and protect the wins you do hit. Done wrong, it accelerates loss without changing the average outcome.
The two numbers that matter
Set two numbers before each session and stick to them:
- Session bankroll: the amount you're putting into play this session. This is money you've decided you're comfortable losing entirely. If losing it changes your life materially, the number is too high.
- Bet size: a percentage of session bankroll per spin or hand. The conventional ceiling is 2% — meaning a $200 session bankroll allows $4 bets. Above 5% per bet, variance dominates the math and short sessions become more likely to end in either fast loss or fast win, not in playing long enough for the long-run math to play out.
The reason for the 2% rule isn't superstition — it's variance math. At 2% bets and 4% house edge slots, a $200 bankroll has a roughly 60% chance of lasting 100 spins; at 10% bets the chance drops to about 30%. Cutting your bet size in half doesn't double your expected return (the edge is the same), but it does roughly double how long you can play before busting.
Session caps that actually work
Hard caps protect you from the cognitive distortions that show up in extended sessions — chasing losses, anchoring on starting bankroll, switching from rational bet sizing to emotional bet sizing. The three caps worth setting:
- Loss cap. The percentage of session bankroll you'll lose before walking away. Common settings: 50%, 70%, or 100%. Below 50% feels too tight (variance can drop you that far on neutral expected value); above 100% means you've already broken your session bankroll discipline by adding more.
- Win cap. The amount above session bankroll you'll take and walk away. Less common to set, but useful for the discipline of leaving while ahead. A 50% win cap means if your $200 bankroll grows to $300, you cash out.
- Time cap. The maximum hours you'll play. Cognitive fatigue gets significant after 90-120 minutes; bet-sizing discipline degrades; session results worsen meaningfully. Two hours is a useful default ceiling.
Caps only work if you enforce them. Most modern operators support self-imposed limits (deposit caps, session-time limits) that lock you out for the period you set. Use them — willpower in the moment is unreliable, system enforcement is reliable.
Strategies that don't work
A few popular betting systems are genuinely zero-EV (no better or worse than random bet sizing), and several are actively worse than random:
- Martingale (double bet after every loss). Sounds clever, mathematically catastrophic. A short losing streak hits the table maximum or your bankroll cap. The system increases your probability of small wins at the cost of one very large eventual loss. Net EV identical to flat betting; variance much worse.
- Anti-Martingale (double bet after every win). Symmetrical opposite — increases probability of one large win at the cost of many small losses. Same flat EV, different variance shape.
- D'Alembert (raise bet by one unit after a loss, lower by one after a win). Slower version of Martingale. Same math, different speed.
- 'Hot' and 'cold' slot timing. Slot RNG has no memory. A slot that just paid out and a slot that hasn't paid in hours have identical probability of paying on the next spin. Timing based on recent history is meaningless.
- Pattern betting (always betting red after three blacks, etc.). Same fallacy. Each spin is independent.
What does work, in terms of changing outcomes: bet sizing as a percentage of bankroll, walking away at predefined caps, choosing low-edge games for serious play, and accepting that the house edge applies regardless. None of these change the expected value — they shape the variance around it.
Bankroll allocation across sessions
If you play regularly, the bankroll question gets bigger than one session. The pattern most disciplined regular players use:
- Set a monthly gambling budget — money explicitly allocated to entertainment, comfortable to lose entirely.
- Divide it into session budgets of consistent size — e.g., $1,000 monthly into ten $100 sessions or four $250 sessions, depending on play volume.
- Replenish only at the start of the next session. If you bust a session in the first 30 minutes, the session is over until the next one — no 'one more deposit' rule.
- Treat winnings as outside the bankroll. Withdraw promptly. Don't roll wins back into the session bankroll within the same period.
When to walk away entirely
The signals that mean stop, not just pause:
- Gambling with money that's allocated to other things (rent, savings, basic needs). The bankroll is no longer entertainment-only — it's destructive.
- Hiding the activity from people in your life. Active concealment usually means the play has crossed into territory you wouldn't endorse if you looked at it openly.
- Chasing losses across multiple sessions or multiple days. Recovery thinking corrupts bet sizing, session length, and game selection.
- Lying about play volume. The act of lying signals internal recognition that the activity is out of bounds.
If any of those apply to you or someone you know, help is available at BeGambleAware (UK), the National Council on Problem Gambling (US), GamCare, GamblingTherapy.org, or your country's equivalent service. Operators support self-exclusion — use it. Stop is always a valid move; the entire system is designed to make 'just one more' feel reasonable when it isn't.
The compressed bankroll discipline: session bankroll defined upfront, bet sizing at 2% or less, three caps (loss/win/time), no betting systems, no chasing, no hiding. None of this changes the math — but it changes whether the math gets a chance to work.