How hard is this puzzle?

Well first of all I have no clue which level is last, if there is a last level. And Segfault's algorithm can't be up to much taking forever between 100 and 108... I mean I haven't even written a program yet, and I'm on Level 35, and the difficulty isn't exactly shooting up, at least in my eyes. Anyway, I've got my own idea for an algorithm. Complicated maybe, but if I can implement it, I should be in first place quite soon... Cool

Can anyone come up with a fiendishly difficult puzzle for this game? I'm not even sure what a really hard one is meant to look like so far.

i think this is one of the harder ones, personally. it scales up a lot when you get to the higher levels... let's see you beat segfault before declaring this one easy =)

adum

I intend to! I see the star system has been implemented on a per puzzle basis. Nice! Except, the stars should be shown everywhere where your name appears (i.e. on the front page, as well...)

Having trouble with data structures at the moment (as in which one to use...)

	ShardFire wrote:
	And Segfault's algorithm can't be up to much taking forever between 100 and 108...

Much of that is that I haven't been running it all the time.

I have a few ideas that might improve it somewhat, though I've been primarily concentrating on Modulo recently...

	Captain Segfault wrote:
	Much of that is that I haven't been running it all the time.

yes, whatever you say Laughing

ShardFire, Ha-ha... You have written so beautifully and dared, but still on level 36 is. Wink

let's go

Compared to modulo, where I cannot find a trick to improve the solver (except constant speedup by multithreading ... everything else was natural) ... brickolade seems to be reach of possible improvement paths compared to simple bruteforce. ... Move (in) dependency is what makes this puzzle really interesting.
Early cuts could improve bruteforce to some level, but I suppose dependency analysis and some kind of dynamic programming would be required to pass higher levels.

Hmmm, I did 76 manually faster than the simple code ...
... And the same with 103 Sad

and 108, 109 ...

I have recoded whole the solver to make it persistent as much as possible ... and seems I gained almost no speedup Sad

... let me hope the added statistics would sppedup the "early cuts", whose were slowing down the original version.

I have rewritten the program and removed wrong cut reasoning. ... And now the program is not able to solve even level 73. I would definitly want to see Yharaskrik's code...

I'm also stuck with Bricolage. It seems that all my approaches take too long to decide that a current state is unsolvable, thus navigating through a lot of useless solutions until it finds the right one.