Breaking Arbiter for fun! :)

Area to discuss development and bugs of official clones (Clone, Arbiter, MSX, Viennasweeper)
Post Reply
dvereb
Posts: 4
Joined: Mon Oct 17, 2011 7:36 pm

Breaking Arbiter for fun! :)

Post by dvereb »

Hey everyone,

Just wanted to post a screenshot showing how I beat beginner in 0.00 seconds! I was bored and wanted to do some programming, so I googled & found a beginner board that could be won in two clicks. I then decided to make my program perform those two clicks, restart the level, perform those two clicks, restart the level, etc. I'm not sure how many tries it needed (I set it up to run 1,000,000 times overnight), but it was successful! I'll attach a screenshot of the error message that popped up. (Obviously this isn't a big deal since this scenario should never arise in normal play conditions)

Figured someone might be interested.

Dave
Attachments
beginner-0-seconds.PNG
beginner-0-seconds.PNG (30.18 KiB) Viewed 6509 times
EWQMinesweeper
Posts: 419
Joined: Sun Nov 30, 2008 11:50 pm

Re: Breaking Arbiter for fun! :)

Post by EWQMinesweeper »

yeah, "programming". :P

it's a neat idea you had there. you might want to read this:
http://www.minesweeper.info/articles/Lu ... dGames.pdf

seeing that the chances for a 2 3bv board are roughly 1 in 20 000, that the first click is always safe and that the second click has to done on one of the other 63 sqaures, your marco needs less than 130 000 attempts. when i tried this myself - before writing what is linked - it took me 5 minutes to find a nice prgram to record macros with and 10 minutes to edit said macro to suit my needs. since my macro was planned out to finish in around .015s and since my computer is far from new, i could only get it to around 200-250 iterations per minute. at this rate it should have taken me 9 to 11 hours to complete a game. needless to say that this is cheating. however, there is no need to worry, since the writing of the linked article led to a change in the rules for our rankings and such games would not be eligible to the rankings.
„Das perlt jetzt aber richtig über, ma sagn. Mach ma' noch'n Bier! Wie heißt das? Biddä! Bidddää! Biddddäää! Reiner Weltladen!“
qqwref
Posts: 125
Joined: Thu Sep 23, 2010 4:17 pm

Re: Breaking Arbiter for fun! :)

Post by qqwref »

Yeah, it's a good thing we have the lucky rule now, because otherwise macro games like this would actually be legit for record entries :) It's interesting that Arbiter breaks if you get a 0.00 on a nontrivial board - but yeah, this would never happen in real play. I know that when I got a 0.00 on Minesweeper X (on a 3BV=1 board, which is illegally easy, so other clones won't generate it) there was no error.
NF player. Best scores 1-10-39.
dvereb
Posts: 4
Joined: Mon Oct 17, 2011 7:36 pm

Re: Breaking Arbiter for fun! :)

Post by dvereb »

EWQMinesweeper wrote:seeing that the chances for a 2 3bv board are roughly 1 in 20 000, that the first click is always safe and that the second click has to done on one of the other 63 sqaures, your marco needs less than 130 000 attempts.
Wouldn't that require my program to actually analyze the board and pick the correct square, though? Mine is blindly clicking the same exact two squares. My next "programming" experiment was going to be trying to analyze the board and solve it automatically. After that I was going to try to make it move the mouse in a more realistic way. At that point, none of my low scores will ever matter again since it's obviously just me cheating. :)

9 to 11 hours seems crazy, though! My program was running 2,960 runs per minute.

qqwref wrote:It's interesting that Arbiter breaks if you get a 0.00 on a nontrivial board - but yeah, this would never happen in real play
I suppose I should say it still works just fine - "Breaks" isn't quite true. It just gives me a nice error message when I open it.
EWQMinesweeper
Posts: 419
Joined: Sun Nov 30, 2008 11:50 pm

Re: Breaking Arbiter for fun! :)

Post by EWQMinesweeper »

dvereb wrote:
EWQMinesweeper wrote:seeing that the chances for a 2 3bv board are roughly 1 in 20 000, that the first click is always safe and that the second click has to done on one of the other 63 sqaures, your marco needs less than 130 000 attempts.
Wouldn't that require my program to actually analyze the board and pick the correct square, though? Mine is blindly clicking the same exact two squares.
9 to 11 hours seems crazy, though! My program was running 2,960 runs per minute.
my estimation was for blindly guessing and the 9-11 hours for my slow macro. at close to 3000 iterations per minute your might have taken less than an hour.
„Das perlt jetzt aber richtig über, ma sagn. Mach ma' noch'n Bier! Wie heißt das? Biddä! Bidddää! Biddddäää! Reiner Weltladen!“
Post Reply