Mines Calculator: Odds, Multipliers and RTP 2026

Mines odds, multipliers and RTP, explained end to end

Mines is the simplest game on a crash casino and the one players misread the fastest. A 25-tile grid, a handful of hidden bombs, and a multiplier that climbs every time you uncover a gem. The multiplier looks generous because the early picks look safe, and that is exactly the illusion the math dissolves. As of 2026 every major site prices Mines the same way, off a fixed formula, so the whole game is knowable in advance. This guide walks the formula in both directions: from the board to the payout, which is what every other Mines page does, and from the payout back to the operator's margin, which is the direction you can actually verify while you play.

How Mines odds actually work

Nothing on this page is simulated, sampled or estimated. Mines has an exact closed form, and once you have seen where it comes from you can reproduce every figure in the tables with a pocket calculator.

The board is decided before your first click

The 25 tiles are laid out and the mines are placed by the server at the start of the round, before you touch anything. From that moment nothing about the layout moves. Clicking a tile does not choose anything, it reveals something that is already there. This matters for the math because it means every pick is a draw without replacement from a fixed pool: the bombs stay put while the number of unopened tiles shrinks by one each time. It also settles the folklore in one line. Opening the corners, opening diagonally, opening slowly, opening in the order some video told you, all describe the order of reveal, not the layout, and the layout is what decides the outcome.

The formula, in plain English

For N mines and k tiles opened, multiply (25 minus N minus i) divided by (25 minus i), for i counting from 0 up to k minus 1. Read out loud on a board with 3 mines, that is 22 safe tiles out of 25 on the first pick, then 21 out of 24, then 20 out of 23. Multiply the three and you get 0.6695652, or 66.96%. That is the chance of opening three tiles in a row and touching nothing. Add a fourth pick and it drops to 57.83%, a fifth and it drops to 49.57%. By the fifth tile on a 3-mine board, more than half of all rounds are already dead. This is the number that surprises people, because the first click on that board is safe 88% of the time and feels like the whole game.

The same thing written as combinations

The product above telescopes into a ratio of binomial coefficients: C(25 minus N, k) divided by C(25, k). On the same 3-mine board opening 3 tiles, that is C(22, 3) over C(25, 3), which is 1540 over 2300, which is 0.6695652 again. The two forms agree exactly, and it is worth checking once yourself, because it is the fastest way to prove that a payout table you found somewhere is or is not built on the real distribution. This calculator uses the combination form, computed in exact integer arithmetic for the 25-tile board, so the figures do not drift at the extremes where a naive floating-point loop would.

The next tile and the whole run are different questions

Two probabilities get confused constantly, and confusing them is what makes deep cash-outs feel safer than they are. The first is the chance the very next tile is safe: gems left divided by tiles left. The second is the chance you survive the entire run from here to your cash-out point, which is the product of all the individual ones. On a 3-mine board the next-tile chance goes 88.00%, then 87.50%, then 86.96%, then 86.36%. Every single tile is safe more than four times in five, which is why a run of successes feels like momentum. But surviving ten of them in a row happens only 19.78% of the time. Note also which way the next-tile number moves: it falls with every success, because the mines never left and the pool of unopened tiles keeps shrinking. The popular advice to press on once a round is going well is exactly backwards relative to what the numbers do.

Where the multiplier comes from

Once you have the survival probability, the payout is not a design choice the casino makes freely. It is that probability inverted, then shaved by the operator's cut, and both steps are visible from the outside.

The fair multiplier is one divided by the odds

A payout is fair when the average return is exactly the stake: probability times payout equals one. Rearranged, the fair multiplier is one divided by the probability of surviving. On our 3 mines and 3 tiles board, that is 1 divided by 0.6695652, which is 1.4935x. This is the price a game with no margin at all would have to pay, and it is the reference every real quote is measured against. It depends on the board and on nothing else. Not on the site, not on your stake, not on your history, not on the time of day.

The real multiplier is the fair one minus the margin

Operators do not pay the fair price, and they are not pretending to. The multiplier they offer is the RTP divided by the survival probability, which is the same thing as the fair multiplier scaled down by the return. Take the same 3 mines and 3 tiles. At a 99% RTP, which is what the large provably fair sites typically run, the quote is 1.4786x. At 97%, a figure several Mines pages advertise, it is 1.4487x. The gap between 1.4935x and whatever your screen shows is the entire commercial content of the game, and on the fourth decimal it looks like nothing. Over a session it is the only thing separating the two sites.

What the ceiling looks like at each mine count

The maximum multiplier on a board is the one for clearing every gem, and it climbs violently with the mine count while the chance of getting there collapses just as fast. At a 99% RTP, one mine and all 24 gems cleared pays 24.75x, and so does 24 mines with the single gem opened once, because both survive exactly 4% of the time. Three mines cleared to the end pays 2277x. Five mines cleared to the end pays 52,599x. Ten mines cleared to the end runs into the millions. None of these are prizes anyone should plan around, and the table on this page shows the whole curve rather than the headline, precisely because the headline is the part that sells and the curve is the part that is true.

Reading the house edge off a live multiplier

This is the part that makes the page useful at the table rather than on paper. Most operators do not publish an RTP for Mines anywhere. You do not need them to.

Two multiplications and you have the answer

Because the fair price depends only on the board, a quoted multiplier carries the operator's margin inside it and simple arithmetic pulls it back out. Multiply the quote by the survival probability and you have the RTP; whatever is missing from 100% is the house edge. On our 3 mines and 3 tiles example, a site showing 1.46x is paying 1.46 times 0.6695652, which is 97.76%, so it keeps 2.24%. That is more than twice the cut of a 1% table for the identical board. The reverse check above does exactly this, and it does not clamp the answer: if a quote comes out above the fair price, it reports a negative edge rather than hiding it, because a table that overpays is a real and reportable finding.

Why the published figure is not enough

Where an RTP is published for Mines, it is usually one number for the whole game, and the payout table is a separate artefact that may or may not have been generated from it. Running the reverse check across a live table is how you find out. In the research behind this page we took one widely circulated Russian-language Mines payout table that sat directly beneath a claim of 97% RTP, and checked it cell by cell. Ten mines and five tiles opened was quoted at 12.2x against a fair price of 17.69x, which is a 68.96% return. The next cell along the same row, ten tiles, quoted a multiplier above the fair price, which is a return over 100%. Across the table the implied return ranged from roughly 69% to roughly 139%. A single number cannot be both. The table was not generated from the stated RTP at all.

Sample more than one cell

One reading tells you what a site pays for one board. Three readings tell you whether the site has a coherent table. Take the multiplier at a shallow cash-out, one in the middle and one deep, all on the same mine count, and run each through the reverse check. If the implied RTP comes back at roughly the same figure all three times, you have measured the site's real rate and you can write it down. If it wanders, the payout table was hand-entered rather than derived, and no single RTP claim about that game means anything. Either result is worth more than an advertised percentage, because you produced it yourself from what the game actually showed you.

Why more mines is not a worse deal

The single most common misreading of Mines is that low mine counts are safe and high ones are reckless. They are not different deals. They are the same deal at different volatility, and the algebra showing it is two lines long.

The probability cancels out

Expected return is the survival probability multiplied by the multiplier. The multiplier is priced as the RTP divided by that same survival probability. Multiply them and the probability cancels, leaving the RTP on its own. That result does not depend on the mine count, on the cash-out point, or on anything you do inside the round. One mine and twenty-four mines return the identical fraction of your stake over the long run. The expected-return tile in the results panel is there to make this visible: move any control on the page and watch it refuse to move.

What the mine count actually buys you

What changes is the shape of the ride. On three mines, one tile opened pays 1.125x at a 99% RTP and lands 88% of the time: a flat, grinding session where almost nothing happens per round. On ten mines, five tiles opened pays 17.52x and lands 5.65% of the time, roughly once in eighteen attempts: long dry stretches punctuated by a result that looks like a win of the week. Both are 99%. Neither is smarter. The choice between them is a choice about how much of your bankroll you are willing to see disappear before the distribution catches up, which is a question about you rather than about the game.

Where the curve turns

The cash-out ladder on this page exists so you can see the turn rather than feel for it. On three mines the chance of surviving is 88.00% after one tile, 77.00% after two, 66.96% after three, 57.83% after four and 49.57% after five. The halfway mark arrives at the fifth click. By the sixth, 42.13% of rounds are still alive. Nothing in the game announces this, and the multiplier climbing on screen argues the opposite case. Reading the ladder before a session is the cheapest thing on this page: it costs nothing and it replaces a feeling with a number.

What actually changes your result

Three things move the outcome of a Mines session, and none of them is a pattern, a sequence or a time of day.

The RTP of the site you play on

This is the only lever that touches your long-run return at all, and it is the one almost nobody measures because it is almost never published. The reverse check above measures it directly from the game's own screen. The difference between a table priced at a 1% edge and one priced at 2.24% is a doubling of what the game costs you per unit staked, for an identical board and an identical experience. If you take one thing from this page, take the habit of reading that number off a site before depositing rather than after.

How deep you cash out

Your cash-out depth does not change your expected return, but it sets your variance completely, and variance is what decides whether your bankroll survives long enough to experience the expected return. Shallow cash-outs give a flat, slow drift. Deep cash-outs give a mostly losing session with rare large results. Neither is a strategy in the sense of beating the game. Both are legitimate choices about what kind of session you want, and the ladder table lets you price that choice in advance instead of discovering it at tile eight.

Bankroll and stop rules

A game with a negative edge punishes long sessions arithmetically. The longer you play, the more reliably the result converges on the house edge, which means session length is not a neutral parameter. Bet sizing relative to bankroll decides how many rounds you can absorb before variance ends the session for you, and a stop rule decided in advance is the only version of a stop rule that works. This is the one area where a decision made before you start genuinely changes the distribution of outcomes you will see, which is exactly why it is the part nobody sells a tool for.

What is not on the list

Tile order, board position, corners versus centre, the diagonal, hot and cold streaks, waiting for a round to be due, increasing after a loss, and every published Mines pattern. All of these describe the order in which you uncover a layout that was fixed before your first click. None of them touches the layout. The Martingale variants deserve a specific mention because they are the most common: doubling after a loss changes the shape of your loss distribution, not its mean, and the shape it produces is one with a small chance of a very large loss, which is generally the opposite of what the person applying it wanted.

Why Mines predictors cannot exist

Mines has a large market in software sold as predictors, hacks and signal bots. The claim is not merely unproven. It describes an operation that has no input data.

The layout is committed before you play

A provably fair round works by committing to a server seed before the round starts, hashing it, and showing you the hash up front. The bomb layout is derived at that moment. The seed itself is revealed after the round, which is what lets you verify afterwards that nothing was changed while you played. During the round the layout exists on the server and nowhere else: not in the page you are looking at, not in your browser's memory, not in the network traffic your client sends or receives. A program running on your device therefore sees exactly what you see, which is 25 closed tiles. This is not a hard problem that better software would solve. It is a problem with no data, and no amount of computation retrieves information that was never sent.

What is actually inside the tools that are sold

The research behind this page opened four of the widely distributed Mines bots. The highlighting that marks tiles as safe comes from an ordinary pseudorandom number generator seeded from the device clock, so the same tool produces a different board on the same round if you run it twice. The accuracy percentage displayed on screen is computed from the clock as well, not from any result. The analysis progress bar is a fixed animation with no computation behind it. In one sample the publication schedule that gives the signals their air of authority was a hardcoded list of minutes, five, twenty, thirty-five and fifty past each hour, which is a line in a scheduler rather than a recalculation of anything. In every sample the payment funnel lived in the same file as the display logic. None of this requires trust in our reading of it, because the same check is available to you: a round of Mines starts when you press the button, not when the clock reaches a scheduled minute.

Verification is the thing that does work

There is a real check, it is free, and the sellers of predictors are uniformly silent about it. After a round, take the seed the server reveals and hash it, then compare it against the hash the server published before you played. If they match, the layout was not altered mid-round. If they do not, you have found something worth acting on. This is what a provably fair verifier is for, and it takes seconds. It is worth being precise about what it proves and what it does not: on a one-sided scheme with no client seed, verification shows the value was not rewritten after publication, and it does not show the value was chosen without regard to who was betting. That is a real limitation and it is still infinitely more than a signal bot gives you.

How to use this calculator

Set the board

Enter the mine count the game is using, from 1 to 24 on the 25-tile board, either by typing it or with the slider. The gem count updates underneath so you can see how many safe tiles remain, which is also the maximum number you can open. Then enter how many tiles you plan to open before cashing out. Anything impossible, such as opening more tiles than there are gems, produces a specific message rather than a blank or a wrong number.

Set the RTP you are testing against

The RTP field is deliberately editable and deliberately not preset to a site's advertised figure. Most operators do not publish an RTP for Mines, and putting a number in your mouth that we have not verified would be exactly the failure this page exists to correct. Use 100 to see the pure fair price, use 99 to compare against the tight end of the market, or use whatever the reverse check told you a specific site is really paying.

Read the results panel

The panel gives six figures: the chance of getting that far, the same chance written as one in N, the fair multiplier, the multiplier at your RTP, the chance the very next tile is safe, and the ceiling for that mine count. Beneath them is the expected return per unit staked, which is the number that will not move no matter what you change. That is not a bug in the calculator. It is the central fact about the game, displayed where it cannot be missed.

Work down the ladder and then reverse it

The table below the calculator runs the same math across every tile on that board, so you can find where the curve turns against you rather than guessing. Then scroll to the reverse check, type in the multiplier your casino displays for that same board, and read off the RTP it is actually paying. That last number is the one worth writing down before you deposit anywhere, and it is the only number on this page that comes from your screen rather than from arithmetic.

Mines terms, defined

Tile

One of the 25 squares on the grid. Each hides either a gem or a mine, decided before your first click.

Gem

A safe tile. With N mines on the board there are 25 minus N gems.

Cash out

Stopping and banking the multiplier earned so far, instead of opening another tile.

Multiplier

What your stake is paid if you cash out at that point. It rises with each safe tile because each one was less likely than the last.

Fair multiplier

The reciprocal of the survival probability, which is what a game with zero margin would pay. Every real quote sits below it.

House edge

The share of every stake the operator keeps, which is 100% minus the RTP. In Mines it is baked into the gap between the fair multiplier and the one shown.

Hypergeometric

The distribution that describes drawing without replacement. Mines is hypergeometric because the mines stay in place while the pool of unopened tiles shrinks, which is why the odds are a product of shrinking fractions rather than a single rate repeated.

Next-tile safety

The chance that the one tile you are about to open is a gem: gems left divided by tiles left. Distinct from the chance of surviving a whole run, and it falls with every success rather than rising.

Variance

How widely results scatter around the expected return. In Mines the mine count and the cash-out depth set the variance and leave the expected return untouched, which is the whole reason mine count feels like a strategy without being one.

Provably fair

A scheme where the server commits to a hashed seed before the round and reveals it after, so you can verify the board was not changed while you played.

Related tools on ToolsGambling

Mines sits inside a wider set of casino math. If you want the same honesty applied to other games, or the bankroll side of the same session, these are the pages that continue the thread.

• Mines game odds and cash-out strategy explained
• Provably fair verifier: check a revealed seed against its hash
• RTP calculator: turn a return figure into expected cost
• House edge calculator: compare margins across games
• Bankroll calculator: bet sizing and how long a session lasts
• Session simulator: watch variance play out over many rounds
• Gambler's fallacy demo: why nothing is ever due
• All casino calculators