Of course! I'll try my best to explain the process in steps: First, I wrote a function that will find all of the "edge" cells (aka the cells that border a revealed number) and stored them in an array with their row and column index and a mine status equal to null. Next we loop through the edge array and recursively run another function called "generateArrangements" which determines if a cell can contain a mine or if it can not contain a mine as determined by the neighboring numbers and other already placed mines. See below for more details. generateArrangements function logic: For each cell, if a mine can be placed in that cell, then we make a copy of the edge array and set that cell to a mine status equal to true. Then we re-run the generateArrangements function on the newly copied edge array with a new mine status. Also, if a cell has the possibility to not contain a mine, then we also make a copy of that edge array and set that cell mine status to false. Then we re-run generateArrangements on this new array as well. Finally, when generateArrangements reaches the last edge cell, then it will store the array as a possible arrangement. Overall, this will create an array of all possible arrangements for the edge cells. If this is still confusing here is a crude graphic representation with an edge array of 3 cells: Edge Cell # Possibility Tree (Y = mine) (N = not a mine) 1 Y N / \ / \ 2 Y N Y N | / \ / \ | 3 N Y N Y N Y Often times a previously placed mine or unplaced mine will cause the next cell to only have one possible mine status. In this hypothetical case, there would be 6 possible mine arrangements: Y Y N Y N Y Y N N N Y Y N Y N N N Y Hope this helps!

@@Mewok given that minesweeper is turing-complete, larger maps may use a lot of memory for all the trees?

@@ollllj Yes, you're probably right.Was just a challenge to see if I could code something. Definitely more efficiently ways to do it though.

Thanks for watching! Glad you found it helpful.

I hope it can help you improve your game!

I hope it does!

I'm glad you found it helpful!

Thank you! I'm glad you enjoyed it.

Unfortunately, minesweeper is NP complete, meaning that the larger the board (and the larger the mine density), the longer the program will take to calculate the probability. You may even have already noticed some delays on the 16x30 board due to this fact. Since my code is by no means completely optimized, it will struggle with larger boards, so I decided to limit the size of the grid. Nevertheless, I think that playing around with probabilities on a smaller scale may still be helpful to improve your best guess intuition which can be translated to larger boards. If I were to recommend a substitute probability calculator, I would direct you to check out one of the following other sites: davidnhill.github.io/JSMinesweeper/ mrgris.com/projects/minesweepr/demo/player/

Thanks!

Hope you got some use out of it.

Yes, that conundrum is what made me want to figure it out!

Thanks, I'm glad you think so!

Thanks for sharing Pedro! That sounds really cool. Is there a way for people to try it?

@@Mewok I've created a repository for it on GitHub today. I've edited my comment to include a link for it.

Hey, sorry for the late response. I explained it briefly at 1:34 in the video, but to expand: You calculate all possible mine arrangements around the revealed region of the board. Then, with each mine arrangement, you subtract the number of mines that that specific arrangement used from the total number of remaining mines. Then, you do a combinations probability calculation between the total number of unrevealed cells and number of remaining mines not used by that mine arrangement. Lastly, you add up all of the combinations calcuations to get the grand total.

I could... but that sounds like a lot of work. Here's a good website that does it already! mzrg.com/js/mine/probability.html

Thank you thank you!

thanks!

Hi! If you mean what algorithm I used to generate the probabilities, I wrote my own. You can see a rough logic explanation in the pinned comment on how it works.

Well I just manual, systematically tried out different possibilities and found out that only eight of them were possible. I don't know if that answered your question tho. Did it?

Craaaaaaaaaazy right?