Mines guide
Dragon Mines: the one-safe-tile board
Dragon is a page label, not a game feature. This page covers the most extreme setting the format allows: every tile but one is a mine.
A board that is almost all mines
With 24 mines on 25 tiles there is exactly one safe tile. The round has one decision: pick a tile. The chance of finding the safe one is 1/25 = 4.00%. The fair multiplier is 25x, and with a 1% edge the offer is 24.75x. A second reveal does not exist; the round is over either way.
This turns mines into a plain one-in-twenty-five draw. It has the same structure as a single number in some table games or a long-shot target in crash: rare success, large payout, same expected cost.
Near-extreme settings
| Mines | Safe tiles | Chance first pick is safe | Offered multiplier | SD per round (units) |
|---|---|---|---|---|
| 15 | 10 | 40.00% | 2.47x | 1.21 |
| 18 | 7 | 28.00% | 3.54x | 1.59 |
| 20 | 5 | 20.00% | 4.95x | 1.98 |
| 22 | 3 | 12.00% | 8.25x | 2.68 |
| 23 | 2 | 8.00% | 12.38x | 3.36 |
| 24 | 1 | 4.00% | 24.75x | 4.85 |
The standard deviation column shows the shape of the risk. At 24 mines one round swings by about 4.8 units, almost five times the stake. The expected net is still -1.00% per round.
Waiting for the safe tile
Since each round is an independent 4% chance, the number of rounds until the first success follows a geometric distribution.
| Round k | P(first safe pick exactly at k) | P(first safe pick by round k) |
|---|---|---|
| 1 | 4.000% | 4.00% |
| 5 | 3.397% | 18.46% |
| 10 | 2.770% | 33.52% |
| 25 | 1.502% | 63.96% |
| 50 | 0.541% | 87.01% |
| 75 | 0.195% | 95.32% |
The average wait is 25 rounds, but the spread is wide. A run of 50 misses in a row happens 12.99% of the time. To be 95% sure of at least one success you need 74 rounds. At one unit per round, a player chasing a single hit can easily stake far more than the 24.75 units that hit returns.
Comparing with a low mine count
Twenty-five rounds at 24 mines and twenty-five one-reveal rounds at 1 mine both stake 25 units, so both cost 0.25 units in expected terms. The difference is the path: steady small results against long losing streaks and an occasional big result. The edge does not care which path you choose.