The microscope puzzle. As many of you might have predicted, even the non-programmers, it's one of the more difficult bits of the project. In fact, just as it was one of the more difficult bits of the game.
I'm going about it in a way that feels like cheating. Basically, I'm completely ignoring the disassembly for now, and implementing my own algorithm (with inspiration from this tool) play the game (hopefully well). The ideal thing being that once I've done this, and have a really good idea of how the game works, I then look at the disassembly and figure out more easily (though certainly not easily) how it works. But it is possible that this latter step may get... postponed. Depends how well my algorithm works (and how different its playing style is from the original's).
Subscribe to:
Post Comments (Atom)
5 comments:
It's beatable with that cell program, I think it makes the same moves every time as well. :) Better than the original game on modern machines in that respect.
But I believe the cell program could beat the original AI, so that doesn't bother me. But yes, it will make the same moves every time, unlike the original. I could add a random factor in there, but hopefully I'll get a chance to properly disassemble it (though if I re-implement that, I will make sure it takes into account the speed of modern machines ;-).
My only concern is that if you make Stauf's AI similar to that cell program, it'd be even more impossible to beat than before!
I have no idea how that cell program works, but I'm hoping there's a way to 'restrain' it. :P
Don't worry, I'll be sure to test it thoroughly! The algorithm I've used is very simple atm, but it has a "throttle" which can make it more or less clever. But I'm not happy with it. As Andrew pointed out, it's nothing like the original. I'm taking a break from it atm, and trying to fix some other bugs. When I come back I'll probably have to bite the bullet and figure out how the original did it...
Well with the original AI the cell program couldn't beat it even with the game running on a separate 233mhz laptop, I think the last machine that I know someone's beaten it on was a 100mhz (it was my sister playing and I was really young so I don't remember if she used the book or not). It's just plain evil on my 3.2ghz, that's like 152 times the difficulty compared to when it came out. Main thing is, it's a lot more fun to play now and that's awesome ;)
@Malorie
As far as I understand it, the cell program calculates all the possible moves within a certain number of turns (2 I think, I modified mine to 3) and chooses the move that has the greater chance of winning in a faster time.
Post a Comment