A bankroll is not the money in your wallet; it is the part of it that you have decided, before opening the app, is already spent. That distinction is the whole subject. Every number a betting platform publishes — the ₹10 entry floor, the ₹10,000 ceiling, the rate ladder, the draw schedule — is a statement about how fast money moves, and none of them is a statement about whether you should let it move that fast. This guide takes the arithmetic seriously instead: what the platform actually bounds, what it leaves unbounded, what a unit of stake costs on each rung, and what the two questions people actually ask — "how much should I bet?" and "how do I win it back?" — evaluate to when you work them out with the site's own rates.
Where these numbers come from. The entry floor and ceiling are betting.min_amount = 10 and betting.max_amount = 10,000, seeded per client and checked against every bid at placement. The payout multipliers are the defaults in the payout configuration: 9.5, 95, 140, 280, 700, 1,000 and 10,000. The draw schedule is the seeded market and slot data. The ruin and chase figures below were computed from those inputs and cross-checked by simulation; they are arithmetic, not forecasts, and they describe a long average rather than any evening you will actually have.
What The Platform Bounds, And What It Does Not
Start with the numbers that are enforced, because a bankroll plan that contradicts them is not a plan. Every bet you place is checked against a minimum and a maximum per entry, and the seeded pair is ₹10 and ₹10,000. Money moving in and out has its own bounds: a deposit must sit between ₹100 and ₹20,000, a withdrawal request between ₹200 and ₹50,000, and the withdrawal path refuses anything that would push your requests past ₹1,00,000 in the current IST calendar day — counting requests that are still pending, not only ones already paid.
| Boundary | Seeded value | Where it bites |
|---|---|---|
| Entry floor | ₹10 | Every bet, checked per entry before the wallet is touched |
| Entry ceiling, main markets | ₹10,000 | Every bet type, one global number for all of them |
| Deposit | ₹100 – ₹20,000 | Per request; the floor is ten times the entry floor |
| Withdrawal | ₹200 – ₹50,000 | Per request; below ₹200 nothing leaves at all |
| Withdrawal, per day | ₹1,00,000 | Rolling IST calendar day, pending requests included |
| Losses | none | No daily loss cap, session cap, cool-off or self-exclusion exists in the money layer |
Read the last row twice. The system has brakes on the way out and none on the way in — which is the opposite of what the word "limits" suggests when it appears on a settings screen. There is a pair of per-user market numbers in the configuration, a maximum bet count of 100 and a maximum total stake of ₹2,00,000, and they are published to clients and editable by an operator; nothing in the placement path reads them, so they describe a ceiling the app displays rather than a wall that stops a ticket. If a loss limit exists in your session, you imposed it. That is not a criticism of the software; it is the fact you have to plan around.
One more enforced structure worth knowing, because it is the operator's own stake-sizing opinion written as data. The Starline product does not use the flat global ceiling. It stores a rate row per bet type with its own minimum and maximum, and the seeded numbers step down as the game gets rarer: ₹50,000 on a single digit, ₹10,000 on a single pana, ₹5,000 on a double pana, ₹2,000 on a triple pana. Twenty-five to one between the flattest rung and the rarest. The book is telling you something about how much of each price it can carry; the main-market path, which applies one ₹10,000 ceiling to every game type including a full sangam, tells you nothing at all.
The Cost Of An Entry Is The Stake, Not The Odds
Most players size a bet by looking at what it pays. The correct quantity is what it costs, and that is not the stake either — it is the stake multiplied by how much of the stake the price takes back. The rate ladder has been taken apart elsewhere on this site, so the summary is short: a single digit at 9.5 is a 95 per cent price, a named pana at 140 is a 14 per cent price, and a full sangam at 10,000 is a one per cent price. The full derivation of those retentions, and the denominators they are the reciprocal of, is in the pana payout guide and the sangam guide; what matters here is what you do with them, which is multiply by money.
| Bet type | Multiplier | Chance a named entry wins | Long-run retention | Expected loss per ₹1,000 staked |
|---|---|---|---|---|
| Single digit (ank) | 9.5 | 1 in 10 | 95% | ₹50 |
| Jodi digit | 95 | 1 in 100 | 95% | ₹50 |
| Triple pana | 700 | 1 in 1,000 | 70% | ₹300 |
| Double pana | 280 | 1 in 1,000 | 28% | ₹720 |
| Single pana | 140 | 1 in 1,000 | 14% | ₹860 |
| Half sangam | 1,000 | 1 in 10,000 | 10% | ₹900 |
| Full sangam | 10,000 | 1 in 1,000,000 | 1% | ₹990 |
The last column is the bankroll column. It says that ₹1,000 pushed through single pana is a ₹860 expense and the same ₹1,000 pushed through an ank is a ₹50 expense, and that nothing about your read of the numbers changes either figure. It also explains the ceiling ladder two sections up: a book that lets you stake ₹50,000 on an ank is taking a different risk than one that lets you stake ₹10,000 on a pana, because the second is a much better price for it.
Two habits follow directly, and neither is about picking numbers. The first is to price a session in staked rupees rather than in bets, because "twelve entries today" tells you nothing and "₹4,200 of turnover today" tells you everything. The second is to notice that the cheapest-looking rung is the one you are least likely to be drawn to, since the same column that punishes a 14 per cent retention also advertises a 140-times headline. The vocabulary that keeps these separate — rate, stake, payout, net return — is defined in the four-terms guide, and the notation "10 ka 95" is unpacked in the rate-list explainer.
Sixty-Three Chances A Day, Whether Or Not You Planned For Them
Turnover is a rate, and rates need a clock. This platform declares a lot of results: twenty-five two-session main markets produce fifty declared outcomes a day, the Starline board adds thirteen hourly slots between 10:00 AM and 10:00 PM, and four jackpot boards declare once each late at night. Take the first two and you have sixty-three events a day that will happen with or without you, each one an opportunity to place a ₹10 entry — which is not a coincidence, since ₹10 is the floor and the floor is chosen to be small enough to feel like nothing.
| Habit | Turnover per day | Expected loss per day | Expected loss per 30 days | Days to halve a ₹10,000 bankroll |
|---|---|---|---|---|
| One ₹10 ank entry on all 63 events | ₹630 | ₹31.50 | ₹945 | ≈159 |
| One ₹10 single pana entry on all 63 events | ₹630 | ₹542 | ₹16,254 | ≈9 |
| One ₹10 ank entry on five events | ₹50 | ₹2.50 | ₹75 | ≈2,000 |
The comparison that matters is row one against row two: the same discipline, the same stake, the same number of events, and a fifty-fold difference in how fast the bankroll goes, entirely because of which rung was chosen. Row three is the other lever. Event count is the multiplier on every price in this table, so reducing how many of the day's sixty-three results you bet is arithmetically identical to reducing your stake — except that it is easier to do, and it does not ask you to bet on a worse price to compensate.
It is worth being honest about what these numbers are not. They are long-run means, and a mean over a few hundred plays is a weak description of any single evening; the variance around them is the subject of the next two sections. They also assume you bet every event, which is a choice, not a law — the sixty-three is what the schedule makes available, and the platform will not stop you at any of them. The independence argument that keeps a mean from becoming a prediction is made at length on the page for when a number is called "due", and it applies to your losses exactly as it applies to anybody's.
A Unit Is A Decision You Make Once
"Unit sizing" is where bankroll advice usually stops, because it is where the arithmetic gets unforgiving. The useful move is to define a unit as a fixed share of the bankroll you decided on, and then to notice what the platform's ₹10 floor does to that definition. If your bankroll is ₹1,000 and your unit is one per cent, your unit is ₹10 and the floor is irrelevant. If your bankroll is ₹400, one per cent is ₹4 and the smallest thing you can actually buy is two and a half times your unit. Below a ₹1,000 bankroll, percentage staking quietly stops being percentage staking, because the ticket size is fixed by the system and not by you. That is the first honest reason the entry floor matters more than the entry ceiling in a conversation about discipline.
The second is what a fixed unit does under a losing run. Staking the same ₹10 every play is arithmetic: a run of one hundred losses is one thousand rupees and then you are out, at which point the game stops for you whether or not you wanted it to. Staking one per cent of what is left is geometric: you never run out in a finite number of plays, you just become unable to buy anything meaningful. Both are worth quantifying, because both are sold as "safe".
| Staking rule | Rung | Typical bankroll after 1,000 plays | Average bankroll after 1,000 plays |
|---|---|---|---|
| 1% of what is left | Single digit | 41% | 61% |
| 1% of what is left | Jodi digit | 3.6% | 61% |
| 1% of what is left | Triple pana | under 0.1% | 5% |
| 1% of what is left | Single pana | about 0.01% | under 0.1% |
| Flat ₹10 | Single digit | busted or still standing; there is no third outcome | |
The gap between the last two columns on the single-digit row is the most instructive number on this page: 61 per cent average against 41 per cent typical. The average is held up by the rare sessions that got lucky and are not the experience you should plan around; the typical outcome is the one you should expect to live. Percentage staking does not turn a negative price positive — it converts a ruin into a long fade, which is better for your wallet and worse for your ability to notice that you are losing.
How Likely Is Doubling Before Zero
The question every player actually asks is not "what is my expected loss" but "what are my chances of leaving with twice what I came with". That one can be computed exactly rather than guessed, and it is the most useful number in this whole guide, because it is the number the interface never shows. Take a ₹1,000 bankroll, stake the ₹10 minimum on one named entry per declared result, and stop the first time you reach either ₹2,000 or less than one stake. Simulate it two hundred thousand times per rung, or solve it exactly as an absorbing chain; the two agree to a tenth of a percentage point.
| Rung | Net on one hit | Chance of doubling before zero | Expected plays until it ends |
|---|---|---|---|
| Single digit | +₹85 | 22.4% | ≈1,090 |
| Jodi digit | +₹940 | 39.4% | ≈150 |
| Triple pana | +₹6,990 | 9.5% | ≈96 |
| Double pana | +₹2,790 | 9.5% | ≈96 |
| Single pana | +₹1,390 | 4.5% | ≈105 |
| Half sangam | +₹9,990 | 1.0% | ≈100 |
| Full sangam | +₹99,990 | 0.01% | ≈101 |
Three things in that table are worth sitting with. First, no rung gets you to fifty per cent, and the best option — jodi, at about two in five — is not the best because it is generous but because one hit nearly reaches the goal, so most of the attempt is spent waiting rather than bleeding. Second, the single-digit row is the slowest way to fail and still fails nearly four times in five; length of session is not a safety property. Third, the sangam rows are not strategies at all: one per cent and one in ten thousand are the odds of a named pair arriving inside the hundred plays a ₹1,000 bankroll can afford, which is the same conclusion the sangam guide reaches from the other direction, in years instead of rupees.
Notice also what the expected-plays column does to the idea of a system. A single-digit attempt lasts about a thousand plays on average. At sixty-three results a day, that is a fortnight of betting every available event — long enough to feel like a career, long enough for a narrative about what worked, and short enough that the outcome was decided by which side of the coin the variance landed on, not by anything you learned.
What Winning It Back Actually Costs
Every losing session produces the same idea: stake a little more until the loss is covered. The chase is not a moral failing to be warned against; it is a calculation, and it can be priced exactly like anything else on this page. Size one entry so that a single hit nets you the ₹1,000 you want back, then repeat that entry until it lands. The expected amount you will have staked when it finally does is the stake divided by the hit probability, and the payout is the stake times the multiplier. Subtract.
| Rung | Entry needed to net ₹1,000 on one hit | Expected turnover until it lands | Expected net | As a share of the ₹1,000 chased |
|---|---|---|---|---|
| Single digit | ₹117.65 | ₹1,176 | −₹59 | −5.9% |
| Jodi digit | ₹10.64 | ₹1,064 | −₹53 | −5.3% |
| Triple pana | ₹1.43 → floored to ₹10 | ₹10,000 | −₹3,000 | −300% |
| Double pana | ₹3.58 → floored to ₹10 | ₹10,000 | −₹7,200 | −720% |
| Single pana | ₹7.19 → floored to ₹10 | ₹10,000 | −₹8,600 | −860% |
| Half sangam | ₹1.00 → floored to ₹10 | ₹1,00,000 | −₹90,000 | −9,000% |
| Full sangam | ₹0.10 → floored to ₹10 | ₹1,00,00,000 | −₹99,00,000 | −990,000% |
The top two rows are the only chases that are even worth discussing, and they cost about six per cent of the money you were trying to recover — a fee, not a disaster. Everything below them breaks for the same reason: the entry that would net ₹1,000 on a pana rung is smaller than the ₹10 the platform will accept, so the floor forces you to over-stake, and the waiting time is a thousand plays instead of ten. On a single pana, chasing ₹1,000 costs ₹8,600 on average. On a half sangam, ₹90,000. That is where the spiral actually lives — not in the size of the bet but in the denominator you have to survive before the bet pays.
This is the arithmetic version of a sentence this site has already written as advice: the safety and security guide states the rule that a stake should be treated as money spent and that chasing is the mechanism by which casual play becomes harm. It is worth restating which half belongs to which page. That guide owns the behavioural rules — the fixed entertainment budget, the stop-loss set before the session, taking profit out rather than rolling it. This page owns why those rules are arithmetic rather than temperament, and why the specific number ₹10 does the work that a resolution cannot: a floor of ten rupees is fine on a rung that pays in tens of plays and ruinous on one that pays in thousands.
The Ledger You Already Have
Almost nobody reads the passbook, and it is the one part of this system built for exactly the review a bankroll needs. The wallet ledger is a full double-entry record: every placement debits the stake at the moment the bet is booked — not when the result arrives, which is why "I'm only down ₹400 today" is usually wrong before you have even checked — every win credits the gross payout, and every deposit, withdrawal request, lock, refund and adjustment is its own row with an IST timestamp and a human description of what it was, in the form Single Pana 249 on Rajdhani Day (Open). Your balance is not a stored number you can argue with; it is the sum of that ledger, credits minus debits.
The ledger is filterable by direction, by category and by date range, which means the three questions a bankroll review actually asks are queries rather than memories. How much did I stake this week — sum the debits. What did the price cost me — debits minus win credits. Which rung did it come from — group the descriptions by bet type. Two honest caveats. A withdrawal request locks money in the wallet rather than removing it, so a pending request makes the balance look larger than the spendable amount. And when an admin voids a bet, the wallet rows for it are deleted, so the passbook is a record of what the system currently stands by, not a complete history of what you did — if you are auditing a period that had corrections in it, reconcile against the bet list as well.
None of this is a self-control feature. There is no spending summary, no trend line, no alert when a week's turnover doubles. The platform keeps the data and does not present it as a limit, which is a design choice you can undo for yourself in about ten minutes with a notebook, on the day rather than on the way home.
A Worked Budget, Start To Finish
Put the pieces together on a number a reader can actually use. Suppose ₹1,500 a month is the money you have decided this costs you — an entertainment budget, agreed with yourself, that you are prepared to see reach zero.
- Set the unit from the floor, not from the budget. ₹1,500 divided by thirty days is ₹50 a day. At the ₹10 entry floor that is five entries a day, which is a real constraint: the schedule offers sixty-three events, so you will spend five of them and ignore fifty-eight. Choosing which five is the whole of your "strategy"; the other fifty-eight are where a budget dies.
- Decide the rung before the day, because it is the price. Five ₹10 ank entries a day cost about ₹2.50 in expectation; five single pana entries cost about ₹43. Over a month, ₹75 against ₹1,290 — the same habit, the same stake, and the second one eats your entire budget on its own.
- Bank the day. Since the stake is debited at placement and the win credited at settlement, your passbook already separates the two. Write the daily net and stop at it; the arithmetic above says nothing about tonight, so a day that ends down is a day that ended down.
- Never chase inside the budget month. The chase table prices that move at roughly six per cent of the amount chased on the two cheap rungs, and between three hundred and nine hundred thousand per cent on everything else, because the ₹10 floor oversizes the entry and the waiting time is enormous.
- Move money out on a schedule. The withdrawal floor is ₹200 and the daily ceiling is ₹1,00,000, but nothing forces either. Money that stays in the wallet is spendable; money that has been requested is locked. That asymmetry is the closest thing the platform offers to taking a profit.
What Bankroll Management Cannot Do
It is worth being blunt about the ceiling on everything above. No sizing rule changes a retention figure. The ₹50, ₹300 and ₹860 in the cost table are properties of the price list, and dividing your stake more carefully, or more bravely, moves them proportionally and no further. A smaller unit buys a longer session, and a longer session is a real thing worth wanting, but it is not an edge — it is the same expected loss spread over more nights, which is why the honest description of bankroll management is that it converts a fast loss into a slow one and lets you stop.
The corollary is that any method claiming otherwise is selling you the variance rather than the price. The guessing-tricks review works through what the popular formula systems actually assume, and the accuracy-claim test sets out what a claim of that kind would have to show to be checkable. Neither is a bankroll argument; both are the reason a bankroll argument is the only one available.
The arithmetic in this guide is a description of prices, not a plan for any evening. Every figure is a long-run mean computed from the platform's own multipliers and schedules; none of the ruin probabilities above is a prediction about you. Matka is a game of chance and should be played only where it is lawful for you, with money you can afford to lose entirely. If betting has stopped being entertainment, the useful numbers are the ones on the safety page, and the useful action is to stop rather than to size better.
Frequently Asked Questions
How much should I start with?
Start with an amount you have already decided is spent, and make it large enough that the ₹10 floor leaves you room. A bankroll under about ₹1,000 cannot use percentage sizing, because one per cent of it is below the minimum entry the platform accepts, so every bet is a fixed and disproportionately large share of what you have. That is a mathematical floor on how small a bankroll can sensibly be, and it is set by the ticket minimum rather than by your nerve.
Is a fixed unit better than a percentage?
They fail differently. A fixed ₹10 unit means a bad run reaches zero in a countable number of plays and ends the session for you. A percentage of what is left never reaches zero in finite plays but decays until the entry floor swallows it — after a thousand single-digit plays at one per cent, the typical bankroll is about 41 per cent of where it started. Neither converts a negative price into a positive one; the percentage version simply makes the loss harder to notice.
Does the platform cap how much I can lose?
No. The enforced bounds are the per-entry minimum and maximum, the deposit and withdrawal ranges, and a ₹1,00,000 daily ceiling on withdrawals counted against pending requests. There is no daily loss limit, session limit, cool-off or self-exclusion in the money layer, and the per-user market numbers that appear in configuration are not read when a bet is placed. The only stop-loss in the system is the one you set before you start.
Why does the bet reduce my balance before the result?
Because the stake is debited at placement. The wallet records a debit the moment the ticket is booked and a credit only if the bet wins, at settlement — so the balance is always showing you money already committed, not money at risk. That is also why a session that feels break-even can look worse in the ledger: the losing stakes left and the winning payouts have not all arrived yet.
What is the cheapest way to bet every day?
Measured in expected loss rather than in stake, the cheapest rung on this site's price list is a single digit or a jodi, both at 95 per cent retention — ₹50 lost per ₹1,000 pushed through, against ₹860 for a named pana. Measured in total, the cheapest habit is the one with the fewest entries, since every price above is multiplied by your turnover and the schedule offers sixty-three results a day to bet into.
Should I take a bigger stake when I am due a win?
Nothing is due, and the sizing logic that assumes it is the most expensive idea in this article. The draws are independent, so the probability on tonight's ticket is exactly what it was on last night's, which is the argument set out on the "due" page. Increasing a stake after losses does not buy a better chance; it buys a bigger share of your bankroll riding on an unchanged probability, which is precisely the shape of the chase table above.
Do the payout ceilings tell me anything useful?
Yes, and they are the most underrated numbers on the site. The Starline product caps a single digit at ₹50,000, a single pana at ₹10,000, a double at ₹5,000 and a triple at ₹2,000 — a twenty-five-fold spread that mirrors how much risk the book is willing to carry at each price. Where a ceiling is flat, as in the main markets, it tells you nothing. Read both alongside the rate list, since the two together describe the game better than either alone.
Is any of this a way to win?
No. Every rung on the price list retains less than its own fair value, so the honest ceiling for a staking system is control over how fast and how much you pay for playing — which is genuinely worth having, and is not the same thing as an edge. If the goal is to end a month ahead, the arithmetic says the reliable way is to have staked less, not to have chosen better.