Messages

1

Borland C++ Builder Contest / Re: Demina: a Minesweeper Solver/Helper

« on: November 14, 2006, 08:32 PM »

Hi Gianni,

here in Italy is called Campo fiorito (flowers instead of mines)

after reading your manual carefully, I found all the information. I have to localize my application as I didn't know that their are  different versions (testing with Windows Vista would be interesting too).

Also did You find the html file on which I detail some techniques used to solve the game ?
Do You use some different ones ?

I developed the algorithm mainly on my own (it is quite similar), doing a bit more about the theoretical stuff. I think I am the only one calculating the correct probabilites for a mine (thus solving more than 36% of expert games). But this could still be improved. 
I wanted to publish the algorithm but this went out of my focus. 
Maybe I will get this done in some time.

I will post a working version for the italian version of MineSweeper. It is good to get some feedback. 

2

Borland C++ Builder Contest / Re: Demina: a Minesweeper Solver/Helper

« on: November 12, 2006, 08:54 AM »

but i couldn't get it to work on my machine.
it starts the minesweeper (always in simple mode), and never plays it and just gives blank screen.

-mouser (November 12, 2006, 01:26 AM)

This is a difficult issue, as controlling the MineSweeper application is quite complex. I tested it on several systems - all working fine.
Does the MineSweeper Solver run (if not what OS do you use)?

I will try to make a version logging the errors which occur. Maybe that will give me a hint what is wrong.

Thx for trying anyway.

3

Borland C++ Builder Contest / Re: Demina: a Minesweeper Solver/Helper

« on: November 12, 2006, 12:34 AM »

Hi Gianni,

can you measure the odds of your solving algorithm?

After pausing one year I improved my solver and fixed some bugs. It does now 36.4% of expert games.
You can try it out: http://www.wsdh.org/?minesweeper

Mathias

P.S: I think I found and error within this board (having 10 bombs). The sum of all bomb probabilites should be 10 - but isn't: