Hello there.
November 14, 2014 8:57 AM   Subscribe

 
I ground to a halt on my first 6x6 and was relieved to find out it was because I'd missed one of the rules: no two rows can be the same (presumably columns too).
posted by cortex at 9:14 AM on November 14, 2014 [1 favorite]


This game is the opposite of aspirin.
posted by vapidave at 9:17 AM on November 14, 2014 [3 favorites]


I find that once I've gone so far as to break a rule, it's hard to back track and make the changes needed to rectify it. (This is on the 8x8). I guess the solution is to be more meticulous and slow at the start so that I don't create problems for myself later on.

I wish they distinguished the locked squares. That would be helpful.
posted by tofu_crouton at 9:17 AM on November 14, 2014 [2 favorites]


Very similar to Simon Tatham's Unruly with the exception that in Unruly you can duplicate rows/columns. If you're an Android person, you can get Simon Tatham's puzzles as a free app, with Unruly and a bunch of other great games. Highly recommended.
posted by the dief at 9:25 AM on November 14, 2014 [3 favorites]


Also my brain keeps wanting to rewrite the title of this post as "Oh, Hello" and I wish it would stop because I'm really not ready for a puzzle game hosted by George St. Geegland.
posted by cortex at 9:27 AM on November 14, 2014 [2 favorites]


cortex: "I ground to a halt on my first 6x6 and was relieved to find out it was because I'd missed one of the rules: no two rows can be the same (presumably columns too)."

Yep, all in the tutorial.
posted by boo_radley at 9:33 AM on November 14, 2014


I have to second the dief's recommendation for Simon Tatham's puzzles. There are a large number of great puzzles that he offers up freely as windows and java/javascript programs on his website besides the android app. Simon Tatham's Portable Puzzle Collection
posted by jamincan at 9:36 AM on November 14, 2014 [3 favorites]


cortex: "I ground to a halt on my first 6x6 and was relieved to find out it was because I'd missed one of the rules: no two rows can be the same (presumably columns too)."

Yep, all in the tutorial.


It's also handy to note that the little eye at the bottom of the screen will move up and down if you've violated one of the rules.

This is a really fun, simple game. Thanks, boo.
posted by Ufez Jones at 9:38 AM on November 14, 2014 [3 favorites]


yeah, Simon Tatham's collection is fantastic, and it's great to see his games have javascript and java versions now. Last time I looked, they were all executables.
posted by boo_radley at 9:39 AM on November 14, 2014


Good news, iPhone people. Simon Tatham's app is also in the iTunes store.
posted by roger ackroyd at 10:07 AM on November 14, 2014 [2 favorites]


Conceptis Puzzles has this as well.
posted by still_wears_a_hat at 10:09 AM on November 14, 2014 [1 favorite]


Lots of fun. The first time that I loaded this in Chrome, it just stuck on the title screen. I assumed this was part of the logic puzzle and got frustrated that I couldn't solve it. Firefox had no problem taking me to the actual game, which is great. And not at all as frustrating as I initially thought.
posted by factory123 at 10:38 AM on November 14, 2014


I find that once I've gone so far as to break a rule, it's hard to back track and make the changes needed to rectify it.

Agreed. I think what it needs is a way to bookmark a board position and a particular square so you can fall back to that position when you run into a dead end. Basically this is all there is to the gameplay - make all the moves that are obviously required. When there aren't any, you just have to make a guess somewhere and then play forward to see if it works or not. So pretty much the only thing that makes this a challenge is keeping track of just where you had to make that guess.
posted by Naberius at 10:38 AM on November 14, 2014 [1 favorite]


So far it appears that the solution is entirely determined by their choice of locked squares, even on the 10x10. There must be a proof of how many locked squares are needed to completely determine the game...
posted by TreeRooster at 10:41 AM on November 14, 2014


Is there ever a need to have to guess? Every time I'm stumped I press the Eye tool for a hint and it shows me which row/column and what rule I've forgot to check.
posted by yeti at 10:41 AM on November 14, 2014


It reminds me of Minesweeper, where there are patterns that you can memorize and use to solve the puzzle. Any two of the same color in a row are between opposite colors, and any two of the same color with a space in between has the opposite color in between. There are probably more complex ones, too.
posted by Small Dollar at 10:43 AM on November 14, 2014


As it turns out, when you set up a custom game board in Unruly, you can specify to have unique rows and columns, which recreates this game precisely.
posted by TypographicalError at 10:56 AM on November 14, 2014 [2 favorites]


What I love about these kinds of puzzles is you start out clueless. Every move you have to carefully deduce why it's correct. Then you start to recognize patterns and meta-patterns which you can fill without thinking about it. In the better styles of puzzles there's actually gotchas there where what look like safe patterns aren't and you have to rethink some patterns you thought were safe. Because eventually you get to this state where you forget WHY your patterns are correct and you find yourself having to reprove them over again when you get stumped.
posted by aspo at 11:06 AM on November 14, 2014 [3 favorites]


When there aren't any, you just have to make a guess somewhere and then play forward to see if it works or not.

I haven't so far found a situation that actually required a guess, though I have run aground and needed hinting a few times on 10x10 boards and sort of given up on it as Not Quite Right For Me, because the no-guessing thing comes down sometimes to correctly visually identifying the two columns or rows out of the whole board that fulfill the requirement that they not be the same in whatever their finished form will be. Which is a lot of manual scanning and guess-and-check and is really hard for me to be actually thorough and efficient about.

Turns out I really enjoy the early trivial plop-plop-plop portion with the double-bounding and the triple-blocking, but not the attentive cleanup at the end, basically. Which is interesting, because I don't necessarily mind poring over a puzzle looking for That One Spot—I've spent a lot of time with nonograms and those come down to precisely that, with a whole bevvy of situation-specific heuristics you can apply—but in this form it just doesn't have the right balance of joy for me.
posted by cortex at 11:30 AM on November 14, 2014 [2 favorites]


The 10x10 board is a little tedious in checking all the options. The 8x8 is really the sweet spot for solving it without getting bored, but with a few sleeping policemen thrown in the middle.
posted by yeti at 11:35 AM on November 14, 2014 [1 favorite]


The web source code is now on GitHub if you're so inclined.
posted by JoeZydeco at 11:53 AM on November 14, 2014


Ya, I've played through about 10 games now, right up to the 10x10 board, and as far as I can tell, there is no guessing (as a few others have said now). First, play the obvious, easy-to-spot moves:

Same-colour tiles with a gap of one in between them get filled with the other colour.
Adjacent same-colour tiles get a different colour on each end.

When you run out of those, look for lines that have the maximum of one colour already set, and fill the rest in with the other. Then go back to the "obvious" set of rules and check again.

When you run out of those, you're left with the tedious part: checking for unfinished lines that are similar to finished lines. When you find one, in your mind, place one arbitrary tile to match the finished line and determine if it will force you to finish the row identically. If so, place the opposite tile. Now re-run the earlier rules.

This should allow you to finish any puzzle.
posted by WaylandSmith at 12:23 PM on November 14, 2014 [2 favorites]


It's weird to me that my brain has to do this verbally. It's really hard for me to compare two rows visually if they're not adjacent, but if I think "red red blue red blue blue red blue" I can do it.
posted by vytae at 12:27 PM on November 14, 2014 [3 favorites]


To compare two non-adjacent lines, I find it easiest to find two adjacent same-colour tiles in the line and scan perpendicularly to see if any other line also has two adjacent same-colour tiles parallel to it. I seem to be able to scan for matches with that method extremely quickly.

Incidentally, I just finished a 10x10 that did not even require me to compare lines to each other.
posted by WaylandSmith at 12:33 PM on November 14, 2014


The hardest part for me is the third rule-set, about identical columns and row. I can usually go around any particular board and deduce about 80% of it just from the rule about not having three in a row. After that it usually comes down to figuring out which two rows are in danger of being alike to one another, and using that to figure out some piece of information that breaks apart the whole puzzle.

Not super hard, but fairly relaxing. I have to say that I enjoy it.
posted by codacorolla at 12:41 PM on November 14, 2014


This tickles the exact same part of my brain that really enjoys doing sudoku puzzles to kill time. That thing aspo said about building up second-order heuristics that end up not being 100% correct and then having to modify them and prove them again is really enjoyable for me.
posted by spitefulcrow at 12:47 PM on November 14, 2014


I haven't so far found a situation that actually required a guess

You're right. I didn't realize clicking the eye would do that, but it does indeed always have a clear and logical reason for something to click next - mainly something abstract enough that I hadn't seen it, like "the last two in this column have to be red - blue, because if they were blue - red, they'd be the same as this other column three or four columns over. But it looks like there's always something.
posted by Naberius at 1:13 PM on November 14, 2014


I found this quite lovely; it scratches the same pattern-matching/deduction itch that sudoku does, but without any of the stuff I find frustrating about sudoku. And yeah, there's no need to attempt hypotheticals - each move is strictly determined. There's always a "thing you know next" somewhere on the board, so it's more of a searching game than anything.
posted by NMcCoy at 1:24 PM on November 14, 2014


This is lovely. It must be noted that the reigning masterpiece of this genre is Hexcells + sequels ($, Mac/Linux/Windows).

Persuasive gushing about said game
posted by Aiwen at 1:40 PM on November 14, 2014


I must not be smart enough for this. I get a title screen with a red box and a blue box and some words, but there's no actionable areas of the page.
posted by yesster at 4:48 PM on November 14, 2014


It's also handy to note that the little eye at the bottom of the screen will move up and down if you've violated one of the rules.

I'm pretty sure this is not so, unless I'm just seeing the eye bob as I click through red to get to blue.
posted by telegraph at 7:18 PM on November 14, 2014


This is excellent, thank you.
posted by jacquilynne at 9:09 PM on November 14, 2014


So far it appears that the solution is entirely determined by their choice of locked squares, even on the 10x10. There must be a proof of how many locked squares are needed to completely determine the game...

I'd guess I could figure it out for the 4x4 if I put my mind to it today but for the larger boards, while there certainly is a minimum number of locked squares that guarantees a unique solution, the verification of the exact number is likely to be a horrible brute-force endeavor. It took years - and an unholy number of core-hours of computation - to determine the minimum number of clues for Sudoku with proof.
posted by Wolfdog at 5:47 AM on November 15, 2014


Man, I really like this. Scratches and itch I didn't know I had.
posted by notsnot at 8:00 AM on November 15, 2014


So far I've only played one 10x10 where I wound up comparing lines. Wayland Smith's observations above will get you pretty far; once I run out of those, I look for lines with imbalances between the two colors filled in (e.g. 4 red 3 blue), and start figuring "if all the remaining blues go in this stretch, we'd have three in a row, so a blue needs to go in this other stretch." That approach seems to complete the board.
posted by adamrice at 11:39 AM on November 15, 2014 [2 favorites]


I wish they distinguished the locked squares. That would be helpful.

Late to party, but there's a new version up that will put a lock icon over all the locked squares if you click on any one of them. Click again and the overlay will go away. Some of the graphics were tweaked as well.
posted by vibratory manner of working at 12:37 PM on November 21, 2014


« Older MetaSorkin   |   The History Of The New York Times' Style Section Newer »


This thread has been archived and is closed to new comments