Key Takeaways
- No staking plan beats the house edge. Your expected loss is always total amount wagered × house edge; a plan can change how that loss is distributed, never its size.
- In the short run luck dwarfs the edge. At ₹50 a bet, 60 rounds have an expected loss of about ₹90 but a typical swing of about ₹390 either way — which is why a losing system can look like a winning one for a while.
- Set your unit at 1–2% of the session bankroll. Smaller units make a session last; larger units make a quick result likelier in either direction. Choose deliberately which you want.
- Martingale fails on arithmetic, not bad luck. Seven straight losses turn a ₹100 opening bet into a ₹12,800 stake after ₹12,700 already lost — and in 100 rounds there is a 37% chance a run that long occurs.
- The Kelly criterion, the famous "optimal bet size" formula, says to bet zero on any game with a negative expected return. Every game on the platform has one, so the goal is to spend a fixed amount, not to grow money.
- Your strongest tools are set in advance: a session budget, a stop-loss, a win goal and a time limit. Decisions made before the first bet are better than any made during a swing.
Ask experienced players what separates people who enjoy real-money gaming for years from those who burn out in weeks, and it is rarely game knowledge. It is money management: how much they bring, how big each bet is, when they stop, and whether they stick to it. That set of habits is called bankroll management, and it is the most practical skill in this whole hobby.
It is also the most misunderstood. Many players hope a clever staking plan will convert a losing game into a winning one. It won't, and this guide shows exactly why, using numbers instead of opinion. What bankroll management can do is just as valuable: it can control how big a bad session gets, how long your budget lasts, and how much of your decision-making survives the pressure of a losing streak. Every probability quoted here is calculated, not estimated by feel, and the tables are yours to reuse.
What "Bankroll" Actually Means
Most players use "bankroll" to mean "whatever is in my wallet on the app". That is the first mistake. A useful definition has three layers:
| Layer | What it is | Who sets it |
|---|---|---|
| Monthly gaming budget | The total you are prepared to lose in a month — taken from discretionary income, after essentials and savings | You, calmly, on payday |
| Session bankroll | The slice of that budget you bring to one sitting | You, before you open the app |
| Unit | Your standard bet, expressed as a small percentage of the session bankroll | Your budget and the game's volatility |
Your bankroll is money you have already decided you can afford to lose entirely — never rent, EMIs, savings or borrowed funds.
The distinction matters because the layers protect you from different failures. The monthly budget stops one bad week from becoming a bad year. The session bankroll stops one bad hour from becoming a bad week. The unit stops one bad bet from ending the session. Together they replace vague "I'll be careful" intentions with numbers you can check.
What Bankroll Management Can and Cannot Do
Being honest about the limits up front makes the rest of the guide easier to trust.
What it cannot do
- Turn a negative-expectation game into a positive one. Every game on Come Game has a published RTP below 100%, so the average result of every bet is a small loss.
- Predict or influence outcomes. Staking decisions happen after the game's outcome is already determined by its own rules; they can't reach back and change it.
- Guarantee a profitable session. Some sessions end ahead by pure chance — that is variance, not a plan working.
What it can do
- Cap the size of any bad session in advance, which is the most valuable protection there is.
- Make your bankroll last for the entertainment you actually want.
- Remove in-the-moment decisions — the ones most likely to be made badly during a run of losses.
- Let you compare games fairly by turning "how much fun" into "how much per hour".
The Cost of Play: Expected Loss Versus Luck
Start with the two numbers that govern every session. The first is the expected loss: total amount wagered × house edge. The second is the swing — the typical distance your actual result lands from that average. For a series of even-money bets, the swing is roughly the stake multiplied by the square root of the number of bets. The table below uses a 3% house edge, which is close to the published edge of games like Aviator (3.0%) or Andar Bahar (2.9%).
| Stake | Bets | Total wagered | Expected loss | Typical swing (±) |
|---|---|---|---|---|
| ₹20 | 60 | ₹1,200 | ₹36 | ₹155 |
| ₹50 | 60 | ₹3,000 | ₹90 | ₹387 |
| ₹100 | 60 | ₹6,000 | ₹180 | ₹775 |
| ₹20 | 300 | ₹6,000 | ₹180 | ₹346 |
| ₹50 | 300 | ₹15,000 | ₹450 | ₹866 |
| ₹100 | 300 | ₹30,000 | ₹900 | ₹1,732 |
3% house edge, even-money bets. "Typical swing" is one standard deviation, so roughly two-thirds of sessions finish within that distance of the expected result.
Look at the first ₹50 row. Over 60 bets you are expected to lose ₹90, yet the ordinary swing is about ₹390. Luck is more than four times bigger than the edge. That is why short sessions feel random and why so many players are convinced they've found something that works: a system tested over a few dozen bets is measuring luck, not the plan.
Over time the balance shifts, because the swing grows with the square root of the number of bets while the expected loss grows with the number itself. The probability of finishing a session ahead falls as you play more:
| Number of ₹50 bets | Expected loss | Typical swing | Chance of finishing ahead |
|---|---|---|---|
| 60 | ₹90 | ₹387 | about 36% |
| 300 | ₹450 | ₹866 | about 28% |
| 1,000 | ₹1,500 | ₹1,580 | about 16% |
3% edge, even-money bets, estimated by simulation (20,000 trials per row) and checked against the normal approximation. The longer you play, the more the edge asserts itself.
Because the edge is a long-run force, your best chance of finishing ahead is to play fewer rounds, not more. Conversely, if your goal is a long session of entertainment, you accept a higher expected cost. Neither choice is wrong, but you should know which one you're making.
Unit Sizing: How Big Should Each Bet Be?
The standard guidance is to size one unit at 1–2% of your session bankroll. That gives you 50–100 units to work with, which is enough to survive ordinary swings in a low-to-medium-volatility game. The table converts a few example session bankrolls.
| Session bankroll | Unit at 1% | Unit at 2% | Units available at 2% |
|---|---|---|---|
| ₹1,000 | ₹10 | ₹20 | 50 |
| ₹2,000 | ₹20 | ₹40 | 50 |
| ₹5,000 | ₹50 | ₹100 | 50 |
Illustrative arithmetic only. Your session bankroll should come from your monthly gaming budget, not from the size of your wallet.
Adjust for volatility. A high-volatility game — a slot with a rare large multiplier, or a single-number roulette bet — pays out so unevenly that 50 units may not be enough to reach a win, so either use a smaller unit or bring a larger cushion. Our game comparison guide gives rule-of-thumb budgets by volatility class, and the RTP and volatility guide explains the underlying idea.
The Maths of Ruin: Why Small Stakes Last Longer but Rarely Win Big
Gamblers' ruin is a classic result: play an even-money game with a small house edge until you either hit a target or lose everything, and the odds of each outcome depend heavily on how many units you start with relative to your target. The table shows the chance of doubling your bankroll before going broke at a 3% house edge:
| Starting bankroll | Chance of doubling first | Chance of ruin first | Interpretation |
|---|---|---|---|
| 10 units | 35.4% | 64.6% | Big bets relative to bankroll — a quick, decisive result |
| 20 units | 23.1% | 76.9% | Moderate |
| 40 units | 8.3% | 91.7% | Small bets — a long session, a low chance of doubling |
| 100 units | 0.2% | 99.8% | Tiny bets — very long play, doubling almost never |
Even-money bets with a 3% house edge, play continues until the bankroll doubles or hits zero. In a fair game the chance of doubling would be 50% at every bankroll size.
The pattern is counter-intuitive and worth sitting with. Smaller bets make it less likely you will double your money, because they force you to play more rounds and give the edge more time to work. If a single decisive result is what you want, larger units get you there faster, at the cost of a high chance of ruin. If a long, entertaining session is what you want, smaller units deliver it — while accepting that a large profit is unlikely. Neither approach changes the expected loss; they change what you get for it.
Five Staking Plans Compared
Staking plans are rules for changing bet size based on recent results. There are hundreds of named systems; these five cover the ideas behind most of them.
| Plan | How it works | What it changes | Main risk |
|---|---|---|---|
| Flat betting | The same stake every round | Nothing — the baseline; the loss is proportional to the number of bets | Boredom; none mathematically |
| Percentage betting | Stake is a fixed percentage of the current bankroll | Bets shrink automatically as you lose, so you can't hit exactly zero | Stakes fall so small that recovery is pointless; bets grow after wins, raising risk |
| Martingale | Double the stake after every loss; return to base after a win | Turns many small wins into a rare huge loss | Catastrophic loss on a long losing run; table and wallet limits |
| Paroli (reverse Martingale) | Double after a win, up to a set number of wins, then reset | Concentrates risk into "profit" money rather than your base stake | A long winning run is required; one loss wipes out the accumulated gain |
| D'Alembert | Add one unit after a loss, remove one after a win | A gentler progression than Martingale | Stakes still climb during a bad run; no change to the expected loss |
The essential fact about every plan on this list: total expected loss = total amount wagered × house edge, regardless of how you size the bets. A progression plan changes only when you wager more, so it can reshape the pattern of wins and losses — many small wins with a rare disaster, or many small losses with a rare jackpot — but it can't alter the average. A plan that seems to "win" for weeks is simply sitting in the frequent-small-win half of its distribution, and the other half is waiting.
Martingale Under the Microscope
Martingale deserves its own section because it is the most popular system and the most persuasive. The logic seems airtight: if you double your stake after each loss, the first win recovers every earlier loss and adds one base unit of profit. And at close to 50/50, a win is always "just around the corner".
The flaw is the speed at which the required stake grows. Starting with a ₹100 base bet and a 3% house edge (win probability ≈ 48.5%), here is what a losing streak costs:
| Consecutive losses | Total lost so far | Next stake required | Chance of that streak from a given start |
|---|---|---|---|
| 3 | ₹700 | ₹800 | 13.7% |
| 5 | ₹3,100 | ₹3,200 | 3.6% |
| 6 | ₹6,300 | ₹6,400 | 1.9% |
| 7 | ₹12,700 | ₹12,800 | 0.96% |
| 8 | ₹25,500 | ₹25,600 | 0.49% |
| 10 | ₹1,02,300 | ₹1,02,400 | 0.13% |
Each step doubles the stake, so the total lost after n losses is ₹100 × (2ⁿ − 1). The final column is the chance of n losses in a row beginning at a given round.
A 1-in-100 chance sounds safe. It isn't, because you get many chances. A "streak starting at any round" is far more probable than a streak starting at a particular one:
| Run length | In 50 rounds | In 100 rounds | In 200 rounds | In 500 rounds |
|---|---|---|---|---|
| At least 7 losses in a row | 19.5% | 36.8% | 61.0% | 90.9% |
| At least 8 losses in a row | 10.2% | 20.6% | 37.8% | 70.2% |
| At least 10 losses in a row | 2.7% | 5.7% | 11.6% | 27.0% |
Probability of at least one losing run of that length, with a 48.5% chance of winning each round.
Put the two tables together. In just 100 rounds there is a 37% chance of a run of seven losses — the point at which a ₹100 base bet has already cost ₹12,700 and demands ₹12,800 more. If you have hit your wallet limit or a table's maximum bet by then, the system has no way to continue. And when it does end in a win, the reward is exactly one ₹100 unit; recovering a single ₹12,700 collapse takes 127 successful cycles. Martingale has a very high chance of a small win and a small chance of a crushing loss, and the average of the two is always the house edge.
💸 Set Your Limits Before You Play
A deposit limit is a bankroll rule enforced for you. New members also get a ₹10,000 welcome bonus on their first UPI deposit — read the wagering terms first, because they change how much you effectively play.
Why the Kelly Criterion Says "Bet Nothing"
Serious bettors and investors often cite the Kelly criterion, a formula for the bet size that maximises long-run growth when you have an edge. It is sometimes quoted in gaming circles as if it justified a particular stake. It doesn't, and the reason is instructive.
In its simplest form, the Kelly fraction is your edge divided by the odds. It only produces a positive bet when the edge is positive — when the game pays better than its true probabilities. With a house edge, your edge is negative, the formula returns a negative number, and the instruction it gives is to bet nothing at all. That is not a quirk: it is the mathematical way of saying that repeated betting on a negative-expectation game shrinks money over time.
The practical reading is that real-money gaming isn't an investment and no staking method turns it into one. Treat the money as the price of entertainment, the way you would treat a cinema ticket or a day out — and use bankroll management to make sure the price is one you chose in advance.
Stop-Loss and Win-Goal Rules That Work
A stop-loss is the amount you are willing to lose in a session before you stop; a win goal is the profit at which you stop while ahead. Both are decided before you begin, and both are more useful than any staking plan.
Setting the stop-loss
Your stop-loss is usually your session bankroll — you stop when it is gone. A softer version is to stop at a fraction of it, say 50%, which protects the second half for another day. Treat it as a purchase price: "I am paying up to ₹2,000 tonight for this entertainment." When you reach it, the evening is over.
Setting the win goal
A win goal is harder emotionally, because stopping while ahead feels premature. But the arithmetic supports it. Suppose you start a session with 40 units and decide to stop at either +20 units or −20 units. Your chance of reaching the win goal first depends on the edge:
| Game edge | Chance of hitting +20 units before −20 units |
|---|---|
| 0% (a fair game — a benchmark, not a game on the platform) | 50.0% |
| 3% | 23.1% |
| 6% | 8.3% |
Even-money bets, symmetrical ±20-unit band around a 40-unit starting bankroll. Even a modest edge makes reaching the upper limit much less likely than reaching the lower one.
That is why a win goal is less about optimising and more about banking. If you get to it, treat it as a rare event and take it: withdraw the winnings to your bank account rather than leaving them in the balance to be re-wagered. A common approach is to withdraw half of any profit as soon as you're 50% ahead, and play on only with the rest.
A stop-loss that moves is not a stop-loss. If you feel the urge to increase it after a losing run, that urge is the signal to stop for the day. Come Game's deposit limits are deliberately slower to raise than to lower for exactly this reason — see the responsible-gaming setup guide.
A Complete Session System
Here is the whole method as a routine you can repeat every time. Nothing in it needs software or a calculator beyond a phone.
Fix the budget layer by layer
Decide the monthly gaming budget from discretionary income, then the session bankroll from it. Only that amount goes into the wallet.
Choose the game, then the unit
Pick a game whose volatility your bankroll can absorb, then set the unit at 1–2% of the session bankroll. The unit follows from the budget, not from how the game feels.
Write down the exits
Note the stop-loss, the win goal and a time limit. Setting a session reminder in the app makes the time limit automatic.
Play flat
Bet one unit every round. If you want variety, change the game between sessions, not the stake within one.
Exit when any exit is reached
Whichever comes first — stop-loss, win goal or time — ends the session. Withdraw any winnings.
Log it
Record the date, game, total wagered, result and how the session felt. After a month, the log tells you more than any memory.
| Date | Game | Unit | Bets | Wagered | Result | Stopped because |
|---|---|---|---|---|---|---|
| Example | Andar Bahar | ₹40 | 75 | ₹3,000 | −₹280 | Time limit |
A simple session log. "Wagered" is the number that matters — multiply it by the game's house edge to see your expected cost, and compare with what you actually lost.
Building the Monthly Budget Around It
Bankroll rules are only as strong as the budget behind them. Work from the top down:
- Start from discretionary income. Subtract rent or EMIs, food, utilities, transport and committed savings. What remains is the maximum pool from which gaming can come — and most people should use only a small part of it.
- Split it into weekly or per-session amounts. A monthly budget of ₹2,000 becomes four ₹500 sessions, not one ₹2,000 evening.
- Never top up from essentials. If a bad session leaves you tempted to use bill money "just to get back", the budget has already done its job by telling you to stop.
- Automate the ceiling. Set the equivalent monthly and weekly deposit limits inside the app so the platform enforces what your willpower might not. The deposit limits guide shows the setup.
- Handle bonuses separately. A bonus has its own wagering requirements that can push you to play more than planned; check the arithmetic first, as covered in our welcome bonus guide.
Tilt, Chasing and the Biases Behind Them
Bankroll rules exist mainly because human judgement deteriorates under a swing. Four well-documented biases explain most of the damage:
| Bias | What it sounds like | Why it's wrong |
|---|---|---|
| Gambler's fallacy | "Six reds in a row — black is due" | Independent rounds have no memory; the next result has the same probability as any other |
| Hot-hand belief | "I'm on a streak, keep going" | A winning run is normal variance and tells you nothing about the next bet |
| Sunk-cost thinking | "I've already lost ₹1,500, I can't stop now" | Money already lost is gone; the only question is whether the next bet is worth its cost |
| Loss chasing | "One big bet will fix tonight" | It replaces a planned loss with a larger unplanned one and is the fastest route to breaking every limit at once |
"Tilt" is the emotional state these biases produce: frustrated, impulsive, betting to change how you feel rather than to play. The defence is to build the exit into the routine so that leaving doesn't depend on being calm. Two habits help: a 24-hour rule after any session that felt bad (no deposits until the next day) and a strict rule that a stop-loss, once hit, ends the day. If you notice that you keep breaking the rules you set, that is worth taking seriously — the responsible-gaming page lists free helplines in India.
Playing Responsibly
Everything above is a set of ways to spend a fixed, affordable amount deliberately. It is not a route to income, and any suggestion that a staking plan can make gaming "pay" should be treated with suspicion. Set your limits in advance, treat losses as the price of entertainment, and stop if gaming stops being fun or starts affecting money you need. Real-money gaming is restricted in Andhra Pradesh, Assam, Odisha, Telangana, Nagaland and Sikkim, and is strictly for adults 18 and over everywhere else. For support tools and helplines, see the responsible gaming page.
The Bottom Line
Bankroll management doesn't beat the house — it makes you a better customer of it. The cost of playing is fixed by the arithmetic: what you wager times the edge. What you control is everything around that number: how much you bring, how big each bet is, how long you stay, and whether you leave when you said you would.
Keep the routine short. Set a monthly budget from money you can afford to lose. Size your unit at 1–2% of the session bankroll. Bet flat. Write down a stop-loss, a win goal and a time limit before the first bet, and treat them as final. Skip progression systems, since the numbers above show they concentrate risk rather than remove it. Do that consistently and the rest of the maths works quietly in your favour — not by changing the odds, but by keeping every session inside a size you chose.
