First we pulled half a million real Crash rounds and watched every strategy lose. Then we dropped a million simulated Plinko balls on every setting and itemized the bill. This week we went one step further, because Mines let us: we did not have to simulate our way to the answer. Mines can be SOLVED - fully, exactly, all of it. There are precisely 300 possible board states in the game (every mine count from 1 to 24, every number of safe picks), and every single one has a multiplier that follows one published formula. So we computed all 300, validated the math against the known payout tables, and then played fourteen million simulated games on top just to watch the theory come true.
What the solution says is more uncomfortable than any strategy guide will tell you. The expected value at every one of those 300 states is identical: 0.99. Cash out after one tile or after twenty, run one mine or twenty-four - the house edge does not move. There is no clever tile, no smart stopping point, no skill hidden in the grid. Every time you click continue, you are simply re-wagering your entire pot at the same one percent disadvantage. The only real choice in Mines is how fast you want that one percent to work on your money.
Below: the theorem, the full decision table, a starting eleven picked from fourteen million games of match data, the strategy with a 49.7 percent winning-session rate - the closest thing to a coin flip in the whole casino - and the chase that once lost 17,293 games in a row. Read it before your next board.
Mines, Solved: All 300 Board States, One Uncomfortable Answer
Three hundred board states. One formula. Fourteen million simulated games on top. Mines is fully solvable, and the solution says something no strategy guide will tell you: every cashout point carries the identical expected value. There is no skilled tile. There is only how fast you want the 1% to work.
The board
| ◆ | ◆ | ◆ | ◆ | ◆ |
| ◆ | ✹ | ◆ | ◆ | ◆ |
| ◆ | ◆ | ◆ | ✹ | ◆ |
| ◆ | ◆ | ◆ | ✹ | ◆ |
| ◆ | ◆ | ◆ | ◆ | ◆ |
25 tiles. Every safe pick multiplies the pot; every continue re-wagers all of it. The mines do not care which tile you feel good about.
The theorem
The decision table, abridged
| Mines ↓ Picks → | 1 | 2 | 3 | 5 | 8 | Full clear |
|---|---|---|---|---|---|---|
| 1 | 1.03x | 1.08x | 1.12x | 1.24x | 1.46x | 25x |
| 3 | 1.12x | 1.29x | 1.48x | 2.00x | 3.35x | 2,277x |
| 5 | 1.24x | 1.56x | 2.00x | 3.39x | 8.50x | 52,599x |
| 10 | 1.65x | 2.83x | 5.00x | 17.52x | 166.40x | 3,236,072x |
| 24 | 24.75x | — | — | — | — | 25x |
Multipliers from the published formula, validated against nine known table landmarks. Note the symmetry the formula hides in plain sight: surviving ONE pick against 24 mines and clearing ALL 24 tiles against one mine pay the same 24.75x - because they are the same bet: one specific safe path through 25 tiles.
The full strategy card
| Strategy | Pays | Win rate | Winning sessions* | Median session* | Worst streak |
|---|---|---|---|---|---|
| 1 mine · cash @ 1 | 1.03x | 96.0% | 15.9% | -$5 | 4 |
| 1 mine · cash @ 5 | 1.24x | 80.0% | 31.9% | -$5 | 7 |
| 1 mine · cash @ 12 | 1.90x | 52.0% | 40.9% | -$5 | 31 |
| 3 mines · cash @ 1 | 1.12x | 88.0% | 28.2% | -$5 | 6 |
| 3 mines · cash @ 3 | 1.48x | 67.0% | 36.2% | -$5 | 12 |
| 3 mines · cash @ 5 | 2.00x | 49.6% | 41.8% | -$5 | 20 |
| 3 mines · cash @ 8 | 3.35x | 29.6% | 43.1% | -$4 | 38 |
| 5 mines · cash @ 2 | 1.56x | 63.3% | 37.3% | -$6 | 14 |
| 5 mines · cash @ 4 | 2.58x | 38.3% | 43.0% | -$4 | 26 |
| 5 mines · cash @ 10 | 17.52x | 5.7% | 47.8% | -$10 | 190 |
| 10 mines · cash @ 2 | 2.83x | 35.0% | 44.9% | -$5 | 27 |
| 10 mines · cash @ 3 | 5.00x | 19.8% | 49.7% | -$5 | 49 |
| 24 mines · cash @ 1 | 24.75x | 4.0% | 44.5% | -$5 | 239 |
| 3 mines · FULL CLEAR | 2,277.00x | 0.0% | 20.1% | -$500 | 17,293 |
* sessions of 500 one-dollar games, 2,000 sessions per strategy. Green row: the captain. Red row: the banned legend.
The starting XI
Fourteen strategies auditioned across a million games each. Eleven made the squad. Every player on this list has an expected value of exactly 0.99 - the formation only decides the shape of the losing.
The shot-stopper. Saves 96 of every 100 - the best win rate in the entire game - and still finishes just 15.9% of sessions in profit. The grinder never wins big enough to outrun the edge. Safe hands; losing scoreline. It is the Plinko goalkeeper all over again.
The steady right back. An 88% win rate at 1.12x, a -$5 median night, and a longest losing run of just 6. Loses slowly and with excellent positioning.
The overlapping fullback: five tiles deep on the safest board, 80% success at 1.24x, and nearly double the goalkeeper’s winning-session rate (31.9%). Proof that even cowards profit from a little ambition.
The centre half. Two picks against five mines, 63.3% survival at 1.56x, 37.4% winning sessions. Solid in the air, nothing fancy.
The sweeper. The classic beginner setting played sensibly: 67% wins at 1.48x, 36.2% of sessions in profit. The back line’s most balanced defender.
The box-to-box engine. A true coin flip per game - 49.6% survival at 2.00x - and 41.8% winning sessions. Runs all day; charges the same 1% either way.
The playmaker. 2.58x at 38.3%, a 43.0% winning-session rate, and the shallowest median in midfield (-$3.70). Quietly excellent distribution.
The attacking mid. Eight tiles deep at 3.35x, 43.1% of sessions in profit - but the losing runs stretch to 38 straight. Creative, streaky, occasionally invisible.
The destroyer. Ten mines on the pitch and still a 44.9% session rate at 2.83x. Wins the ball once every three games and makes it count.
The Superstar Diva. One tile, one chance in 25, and the famous 24.75x. Wins 4% of the time and STILL takes more sessions to profit (44.5%) than the goalkeeper who wins 96%. That statistic alone is the whole article. Longest sulk: 239 straight misses.
THE CAPTAIN. The best seat in the game: a 49.7% winning-session rate - one tick from a genuine coin flip - on a 5.00x that lands one time in five. If Mines has an honest configuration, this is it. Its expected value is 0.99, exactly like every other player on this list. The armband changes nothing; it just loses with the most dignity.
On the bench: 1 mine cash@12 (a respectable 40.9% session rate nobody talks about), 5 mines cash@10 (17.52x bait with 190-game droughts), and the banned legend - the 3-mine FULL CLEAR at 2,277x. It hit 451 times in a million attempts, exactly the theoretical 1-in-2,300, and its longest losing streak was 17,293 consecutive games. Its median session is everything, gone. Same expected value as the goalkeeper. That is the theorem laughing at all of us.
The grinder paradox
High scores
README.TXT
Method: the published multiplier formula (0.99 divided by survival probability), validated against nine known payout-table landmarks, enumerated across all 300 possible states, then stress-tested with a million simulated games per strategy - seeded and reproducible. What this proves: the complete price of every decision in Mines. What it disproves: the existence of a clever one. Companion reading: our 500,000-round Crash strategy audit, and the Crash verifier that checks any round in your browser - because Mines results, like Crash results, are provably fair and independently checkable. That transparency is why we can solve the game at all.
18+ · Mathematics and simulation, not prediction. Only bet what you can afford to lose.
