Beautiful. Elegant. Subtle.
[/spoiler]

You made it look like it was originally 1529 moves...

well originally it was 10, then I got the enlightment and upgraded it to 15, then I got uber-enlighted and uber-graded it to 29. be noted, this was not done by a computer. it's beautiful, and most importantly, possible for humans, in a finite, not long, time.

Tbh, I looked at the spoiler eventually. It's basically a gravity grid version of *** ********** or **** ** ***** (not to spoil).

lol, yes well, you can't expect this game to be fun. but hey, you have to appreciate this. This is really amazing in myopinion. Think about the idea behind the game. its simplicity makes it a real achievement to create such a level, the way I see it. You're welcoe to try - making a solvable level of more than 10 steps is very hard. this algorithm may be alittle cheating, but still, can you find anything like it?

I completed it w/ three moves left by doing a flip at the start.  

what?
but I checked that out, and I was 1 move missing *opens game*

oh well, but you get the idea. whaddya think?

I think I'm still waiting for that super-puzzle you promised us way back when you were making that program to solve them.

The program exists and it works fine, but there's a problem in creating an algorithm to produce the level-with-the-most-needed-moves in finite time that is good enough for my computer. see, there are many levels that need not to be tested in the brute force, and even telling the program how to check if a level should or should not be tested is a mind****. at the moment, I can tell the program suff like,

"build me a level with 4 colors, culumns of maximum 5 tiles high, with 12-35 gray blocks, solvable in 17 moves, but not in 16".

then it'll think, for this specific example, around 10 minutes, and then it'll give you a picture of the 9x9 grid with the level on, and a line with the solution at the buttom.

going for the best would require me to run one loop for the colors - 1 to 7, and in it, one move for the moves; gray blocks will be probably 0-77 automatically. clearly the 77 can be reduced by thinking alittle, and the 0 can be increased too, but we'll get to that later. the algorithm isn't ready enough. you're welcome to try and help, I can upload the files, it's around 8 files I think, including the applet, graphics, etc. there's also an option to run a void main, which will do exactly the same but in the Terminal, for stupid computers, such as mine and yours. If anyone has a good computer, the program is likely to run much better.

Lol, I'd actually like to see this program.

Could you post it/put it up for download Shos?

Perhaps I should try writing something like this in Matlab. A solver would probably be very easy to write.

http://www.2shared.com/file/VdRhHnmy/gravitygridwithanewtouch.html?


I'm *pretty sure* that this is the correct ZIP file. once again, it does not contain alot of stuff, but in general if you read everything, it should work. note the use of the random function. since I did not tell it to avoid stuff like checking both

011
122
200
and
411
133
344

etc.

Shos, I can't read this code at all. You need to use some damn spaces, 'cause otherwise it looks like just a random jumble of characters....

Open it in some programming program or something. I use BlueJ for that, I never code in text files lol.

It has all the tabs and stuff needed.. Although there are places where it's alittle messed up(alittle!), and the word 'height' is spelled without the E.

sorry for inactivity of late, any news here?