Thursday, 28th March 2024
Puzzles Solved Yesterday: 124
Forum Index
 
Page 3 of 5<12345>
Masyu please!
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.10 21:29:19
Hmm. Right now if I use fixed position to test one option, and end up solving the puzzle it will reset the fixed position state to the solved state--so I can't go back and test the other option. Maybe it would be better if it wouldn't check the solution until I've clicked a "submit" button? Then I wouldn't have a problem with automatically getting a new puzzle.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.10 23:23:38
Unlike kwontom - my pages have full undo/redo - it works on fix position itself, so you can undo the fix position, and then reset.

The pages are also currently designed for solo puzzle format rather than competitive format, if I was to do something like kwontom I would remove the checking for solution entirely when working on a competition puzzle and indeed have a submit button.

Also, I already disabled the automatically generate new puzzle part, you may need to refresh your page.
Last edited by Tilps - 2010.06.10 23:24:14
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.10 23:55:12
Quote:
Originally Posted by Tilps
Unlike kwontom - my pages have full undo/redo - it works on fix position itself, so you can undo the fix position, and then reset.
Ah, I didn't notice that. Okay, so it's pretty much perfect, to me at least. And the "unlimited" puzzles are pretty challenging, too. Great stuff.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.11 01:07:02
On the topic of competitive puzzles - I'm willing to give writing such a site a go for Twist. (In silverlight - sorry html and I don't get along )  I don't want to steal from this site though, so I'll only do it if foilman doesn't want to do it for here.
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.11 01:27:29
My six puzzles, in Twist format: 1, 2, 3, 4, 5, 6
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.12 02:58:53
This puzzle is inspired by a slitherlink puzzle I made a few years back. The idea there was to have a large clueless space that the loop passed through. Here's the puzzle and here's a discussion with other examples.

Anyway, this is what I came up with: Masyu #7

I should warn you that the program took six or seven minutes to solve this (on my computer) after I finished the last line. I'm not sure why--it only takes about ten seconds to solve similar sized "unlimited" puzzles. I think the hardest puzzle I've come across is this one, which took quite a while for me to solve without using lengthy fixed positions, but which took only 25 seconds for the program to solve.

Edit: If I move one of the white circles in the upper left to get this, where the solution is identical and the reasoning is almost the same, now it only takes about a minute for the program to solve. (This doesn't matter to me at all, but I thought it might be useful to you, Tilps, as you optimize your program.)
Last edited by Brian - 2010.06.12 03:27:26
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.12 03:18:09
Ahh you've found a 'bug' which I had been hiding
When you load Twist with 'specific' puzzle, the puzzle solving for the final click is ... broken.  It will verify it eventually, but it first tries to solve it with a very wrong set of options which are probably the slowest possible solving mode (even slower than unlimited... at the same time as being less complete), if that fails it then tries to solve it with unlimited...
I'll look at getting that fixed...
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.12 08:47:29
I've put up a version which might be improved, if you want to give it a try.
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.12 09:58:15
Quote:
Originally Posted by Tilps
I've put up a version which might be improved, if you want to give it a try.
When I tried #7 it wasn't finished after ten minutes and I forced-quit firefox, as it gets locked up while solving. (I was watching a video at the same time, though.) The altered #7 finished in a minute and a half, a little longer than before. I always use Twist3, although when I post here I change the links to go the regular version.
Last edited by Brian - 2010.06.12 09:58:23
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.12 11:30:18
Here's the simplified version of #7--the center 5x5 square is empty. This one has an empty 7x7 square.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.13 01:56:57
I just solved #7 by hand using Twist, took me about ~30 seconds for it to do the final edge. Which is way too long, since the puzzle can be solved by simple trials leading to contradictions plus some basic counting, but it also isn't 10minutes.
Twist3 does appear to be a tiny bit slower, but it only took about a minute.
Last edited by Tilps - 2010.06.13 02:01:54
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.13 02:44:47
I've uploaded a new version which 'presolves' puzzles while you play them.  If you manage to solve the puzzle faster than it (...) it won't go all green until the top left corner says validated.
Btw I think I know why my times for last click were so much faster, even for Twist3. I am still using .Net, rather than mono runtime.  There are definite areas with mono which are not as fast as the official microsoft runtime.
Last edited by Tilps - 2010.06.13 02:54:25
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.13 07:37:11
Empty 7x9 and empty 8x8.
Last edited by Brian - 2010.06.13 07:37:22
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.13 07:56:17
I like empty 8x8, quite nice.
foilman
Kwon-Tom Admin
Puzzles: 3384
Best Total: 24m 6s
Posted - 2010.06.13 08:36:53
Quote:
Originally Posted by Tilps
On the topic of competitive puzzles - I'm willing to give writing such a site a go for Twist. (In silverlight - sorry html and I don't get along )  I don't want to steal from this site though, so I'll only do it if foilman doesn't want to do it for here.

Of course, please go ahead! I don't think I have the time to do such a thing right now (not properly anyway) and I'd be interested to see what you come up with. If you need any help or wonder how I've done things on this site, feel free to drop me an email.

I have started working on some code to automatically solve Masyu puzzles which I was going to experiment with to see if I could generate them too - but I'm quite happy to post them here or on your site if and when I get it working.
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.14 11:13:35
Torture. Made this while exploring some ideas. Haven't even attempted to solve it from scratch. Not recommended for anyone. I doubt there's any joy to be found in solving it, but I know some people love brutal challenges, so here you go.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.14 13:37:46
Hopefully the mass of new stuff I just posted still works on moonlight...

Before anyone asks, it is not done yet, not even close!  But I thought I would share this 'pre-alpha' version to show some general layout ideas (although even that is pretty dodgy...)
Also note the back-end server is a temporary host (themissingdocs.net is hosted on a unix machine and all my back-end code requires asp.net), a temporary host which will likely shut down at random times of the day... so if you find nothing is working at all (other than 'Just Play' which should work just like as if the rest wasn't there...) it is probably just the host is shut down.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.14 13:44:19
Quote:
Originally Posted by Brian
Torture. Made this while exploring some ideas. Haven't even attempted to solve it from scratch. Not recommended for anyone. I doubt there's any joy to be found in solving it, but I know some people love brutal challenges, so here you go.

Have you encoded the halting problem yet?   I don't even know where to start with that puzzle, and yet it seems my solver managed to work it out pretty quickly (probably using brute force recursion though...)
Brian
Kwon-Tom Obsessive
Puzzles: 4739
Best Total: 9m 6s
Posted - 2010.06.15 04:48:14
Quote:
Originally Posted by Tilps
Have you encoded the halting problem yet?   I don't even know where to start with that puzzle, and yet it seems my solver managed to work it out pretty quickly (probably using brute force recursion though...)
I don't think you can follow the usual process of slowly working your way to a solution. Instead you probably have to figure out all the possibilities for one corner, and then see how the corners can fit together. That's how I constructed it, at least.
Tilps
Kwon-Tom Obsessive
Puzzles: 6482
Best Total: 20m 22s
Posted - 2010.06.15 11:34:37
Didn't get much done tonight - but the login code is a bit more user friendly, and the leader board shows the cumulative total for the last week rather than just the times for the last puzzle.
For those who have bothered to create an account, I'm uploading a new puzzle.
Page 3 of 5<12345>

Forum Index