Skip to content

Computes the luck needed to solve minesweeper in one guess

Notifications You must be signed in to change notification settings

JakobDegen/lucky-minesweeper

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Lucky Minesweeper

Simulation tool that answers too questions:

  1. How lucky do you need to be to solve a minesweeper board in one guess?
  2. Where should you guess if that's your goal?

This was a one (and a half) day project, so don't expect much. All board sizes are supported, but those and related parameters are const-generic (yay perf!) so if you want a different size, change the values in main.rs and re-compile. I use a couple more advanced const-generic features, so compiling requires nightly.

Should be decently (but not insanely) optimized. I'm not looking to optimize this much more right now unless someone has any suggestions for getting an order of at least an order of magnitude of performance out. Probably this will require some changed strategy.

Results

About one out of 70,000 beginner sized minesweeper boards (9x9 with 10 mines) can be solved in a single guess. Surprisingly, the best place to guess is to do this is in a corner of the board. A guess in the corner will give you about a one in 180,000 chance of instantly solving the puzzle. Guessing diagonal to a corner is the worst option, where you're reduced to about a one in 4,000,000 chance of solving the board.

I don't have enough compute to have found even a single instantly solvable board in Intermediate or Advanced sizes. Would love to hear from anyone who finds one though!

About

Computes the luck needed to solve minesweeper in one guess

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages