Define your session bankroll, then commit to it
Before you open a single slot, decide what amount of money you can afford to lose without affecting any other part of your finances. That figure is your session bankroll. It should be small enough that losing it would not change your week. For most casual players, that is $50-$200; for serious bankroll players, it might be $500-$2,000; for entertainment-only sessions, $20-$50 is plenty.
Whatever the figure is, deposit only that amount and stop when it is gone. The single most expensive habit in slot play is the "I’ll just deposit a little more" decision after busting an initial bankroll. The math does not improve in your favor on the second deposit. Whatever variance trapped your first bankroll will trap your second.
Most regulated operators offer deposit limits that you can set in advance. Set them at your session bankroll figure. The friction this creates the next time you want to "just add a little more" is the most important player-protection feature on the platform.
Bet sizing by volatility
The optimal bet size depends on the volatility tier of the slot you are playing. The basic principle: your bet should be small enough that the per-spin standard deviation of returns is a small fraction of your bankroll, otherwise variance will end your session before the math has a chance to play out.
A practical heuristic: bet 0.5%-1% of bankroll on Extreme volatility, 1%-2% on Very High, 2%-3% on High, 3%-5% on Medium, 5%+ is fine on Low. Translated: $200 bankroll, $1-$2 bet on Mental, $4-$6 bet on Bonanza Megaways, $10+ bet on Starburst. These are floors — many serious players go more conservative.
On Extreme volatility games specifically, the math justification is hit-rate-driven. With a 1-in-300 bonus trigger and an Extreme bonus distribution, you need at least 600-1,000 spins to have a reasonable chance of seeing two bonuses. At $1 bet on a $200 bankroll, that is 200 spins of guaranteed runway in the worst case (every spin pays zero), enough to reach the bonus at expected frequency.
Stop-loss and stop-win thresholds
Set a stop-loss threshold — typically 50%-70% of your starting bankroll — and stop the session when you hit it, regardless of how you "feel" about the next spin. This is hard to execute. The point is to internalize the discipline before the session, when your judgment is uncluttered, rather than during the session, when variance is pulling you in different directions.
Set a stop-win threshold too — typically 150%-300% of starting bankroll. If you double or triple your money, withdraw the original stake (or the entire balance, depending on your rules) and stop. The variance that produced the win can also take it back; banking the win locks in the realized return.
A useful framing: treat the session bankroll as a budget, not as capital to be grown. If you start with $200 and end with $500, that is a $300 entertainment subsidy from the variance distribution. If you start with $200 and end with $0, that is the cost of two hours of entertainment. Neither outcome is an investment thesis.
Game selection by bankroll size
Small bankroll ($20-$50): Low or Medium volatility, bet $0.10-$0.50, target steady play with frequent small wins. Starburst, Gonzo’s Quest, Sweet Bonanza at minimum bet, NetEnt classics generally. The goal is entertainment value per dollar wagered, not big wins.
Medium bankroll ($50-$200): High volatility tier, bet $0.50-$2. Modern Pragmatic Play and Play’n GO catalog. Megaways slots at minimum bet. The bonus is reachable and the variance is survivable.
Large bankroll ($200-$1,000): Very High or Extreme volatility, bet $1-$5. Push Gaming, Hacksaw, mid-tier Nolimit titles. Bonus Buy becomes reasonable. The math distribution is wider but the bankroll can absorb it.
Very large bankroll ($1,000+): Any volatility tier. Bonus Buy at standard cost. Multi-tier X-iter buys make sense. The bankroll is large enough to see the long-run statistics begin to converge over a single session.
What about progression systems?
No progression system improves slot RTP. Martingale (double on loss), Fibonacci, Labouchere, D’Alembert — none of them work on slots, because slot outcomes are independent. Doubling your bet after a loss does not make the next spin more likely to win; it just makes the loss bigger when it comes.
Some streamers practice "step-up" systems where they raise bet size after a big win and drop it after a loss, framed as "betting with the casino’s money." This is a behavioral preference, not a math edge — the expected value of every spin is the same regardless of bet size, and adjusting bet size only affects variance.
The closest thing to a real edge is bonus-hunting on positive-EV promotions. If a casino offers a 100% deposit match with reasonable wagering requirements (5x to 20x deposit, on slots that count 100%), the math can occasionally be slightly positive expected value across a long enough sample. This is real but rare, and most "advantage play" opportunities of this kind have closed in the past decade as operators tightened terms.
Time-based budgeting
Beyond money, set time limits. A typical online slot session at typical pace produces 400-800 spins per hour. At $1 bet on Extreme volatility, that is $400-$800 wagered per hour, with a long-run expected loss of $14-$28 per hour at 96.5% RTP. In a single session, your actual outcome can deviate by hundreds in either direction.
A practical pattern: set a 60-90 minute timer at the start of the session. When it goes off, take a 15-minute break — leave the room, get water, do anything else. After the break, decide whether you want to continue or stop. The break reliably interrupts variance-driven decision making, which is where most session damage happens.
Most regulated operators offer reality-check pop-ups (every 30, 60, or 90 minutes) that interrupt play with a session summary. Enable them. The friction is the feature.
Recognize when slots are no longer fun
Bankroll management is ultimately about keeping slot play in the entertainment category, not the financial-decisions category. If you find yourself increasing bankroll to chase losses, playing longer than you intended, lying about how much you have spent, or feeling worse after sessions than before — those are signals that the math has stopped being entertainment and started being something else.
Most regulated jurisdictions offer a self-exclusion option that locks you out of operators for a chosen period (24 hours, 7 days, 30 days, 6 months, or permanent). Use it without shame. The operator infrastructure is designed to make stopping difficult; self-exclusion is the strongest counter-tool available.
GambleAware (UK), GamCare (UK), Gambling Helpline (US), and BeGambleAware all offer free, confidential support if you want help thinking through your relationship with gambling. The math of slots is interesting, the design is engineered for engagement, and that combination can be problematic for anyone — there is no shame in needing tools beyond bankroll management.