Talk:Sliding puzzle

Untitled
"We will call sliding puzzles only those puzzles that are essentially two-dimensional in nature" ... well, okay ... but you'll have to break out "sliding block puzzles" to account for things like the Slide Cube

See also Talk:Tiling puzzle and Talk:String puzzle. Restored because:
 * Not listed in Votes for deletion/Karlscherer3, and therefore not properly deleted per VfD
 * Google shows that "sliding puzzle" is a legitimate generic term, and therefore Wikipedia should have a spam-free article on it. We cannot let a spammer decide for us that we can never have a Wikipedia article on a particular term which Google shows to be in relatively widespread independent use.

-- Curps 07:14, 7 August 2005 (UTC)

Vote for Deletion
This article survived a Vote for Deletion. The discussion can be found here. -Splash 02:54, 20 August 2005 (UTC)

Solutions?
Is there an algorithm for solving these puzzles like there is for the Rubik's Cube? —Ben FrantzDale 17:20, 20 February 2007 (UTC)

No, the problem is extremely hard to compute as there are too many possibilities for any computer to compute after a certain size. For a small size (like the 15 puzzle) you can solve it quickly, even large ones are solvable but not in optimal time. Keldon85 08:09, 19 August 2007 (UTC)

More specifically, determining whether a given sliding-block puzzle has a solution is PSPACE-complete. Refs:


 * Hearn, R. A. and E. Demaine. PSPACE-completeness of sliding-block puzzles and other problems through the nondeterministic constraint logic model of computation. Theoretical Computer Science, 343(1-2):72-96, October 2005.
 * Martin Gardner. The hypnotic fascination of sliding-block puzzles. Scientiﬁc American, 210:122–130, 1964.
 * Bobhearn (talk) 22:54, 27 November 2007 (UTC)

There is a general procedure for solving these puzzles - provided that they are soluble. I presume it's in PSPACE (since using it it and failing to find a solution would tell you the problem is insoluble), but it's nonetheless practical for 4x4 or 5x5 problems. I know this because at one point i knew what it was! As i recall, you solve row by row and column by column, using a sort of gradually decreasing circulating motion of tiles. However, i've forgotten the details. If anyone knows it, some info on the page would be good - i have a tile puzzle on my desk and can't solve it. -- T. A. —Preceding unsigned comment added by 128.40.81.22 (talk) 20:16, 13 March 2009 (UTC)

I agree with the above commenter. I used to solve these puzzles all the time as a kid, and I always used the same procedure, pretty much what he describes. I can't for the life of me figure it out now, though. Someone who knows it please please please post it! —Preceding unsigned comment added by 99.226.243.80 (talk) 15:57, 13 May 2009 (UTC)

Yes, there is a general procedure for solving them -- there's just not an efficient one. The only obvious general procedure is to brute-force search all possible non-looping move sequences, which is impractical for large puzzles. There are various tips and tricks one can use in practice on small puzzles; a nice list is found in Winning Ways vol. 4, pp 878-881. However, the method described above by T.A. is only applicable to puzzles like the Fifteen puzzle, where the blocks are all 1x1. These puzzles do indeed have efficient solutions. It doesn't work for more general sliding-block puzzles. The fact that general sliding-block puzzles are PSPACE-complete to solve means that effectively, you can build computers out of sliding-block puzzles. Here is another paper, more accessible than the above references, which shows this:


 * Hearn, R. A. The Complexity of Sliding-Block Puzzles and Plank Puzzles. In Tribute to a Mathemagician, pages 173-183. A K Peters, 2004
 * Bobhearn (talk)

Merge with Fifteen puzzle
This page and the page fifteen puzzle describe essentially the same kind of puzzle, save in different terms (the latter is more mathematically inclined). I suggest that these pages be merged. Qwertyus 17:55, 18 March 2007 (UTC)
 * I agree. Be bold and do it! --Bookgrrl holler/ lookee here  14:53, 24 September 2007 (UTC)

The page on the fifteen puzzle describes a particular instance of this problem with a significant history and mathematical properties, i believe this seperates it enough to justify a seperate article. 124.170.134.66 14:48, 13 October 2007 (UTC)
 * That's my opinion as well, couldn't say it better. What is really needed is a rewrite of *this* article in more general terms. There are many more sliding puzzles than just Loyd's Fifteen. More links, and probably an attempt to classify them somehow, are needed here. Misof (talk) 12:59, 13 December 2007 (UTC)

I agree, there are more sliding puzzles than just the unsolvable 15-block puzzle, indeed more than just the plain 15-block puzzle alone. There are many sliding-block puzzles that use uneven shapes and interesting concepts, such as Traffic Jam (cars are slid to release a target car), polyomino puzzles (irregularly shaped pieces) and including Klotski. The sliding puzzle has a very rich and interesting story of its own right. The fifteen puzzle I would consider a subset of this category. Sheldon Carpenter (talk) 16:50, 15 December 2007 (UTC)
 * Classic wood puzzles []
 * Modern sliding puzzles []

I have uploaded some images of sliding puzzles that are NOT of the numerical type for use in the Combination puzzles article. They are not the most wonderful pics in the world (might have another go one day) but you are welcome to use them here if you want. It would certainly stop the merger argument as the puzzles would be seen to be different.  Sp in ni ng  Spark  13:13, 6 January 2008 (UTC)

One more reason to merge with "15 puzzle": there are many more languages at the latter. — Preceding unsigned comment added by 184.148.234.162 (talk) 15:05, 24 January 2013 (UTC)
 * That is a really silly and irrelevant reason. Doubtless, all those other languages are not discussing sliding puzzles in general.  Spinning  Spark  17:43, 24 January 2013 (UTC)

redirection at rearrangement puzzles
"This property separates sliding puzzles from rearrangement puzzles." but the link redirect at the article. 216.86.113.233 (talk) 05:21, 17 March 2008 (UTC)

Missing Classic Examples
The 15-sliding puzzle was wildly popular in several versions employing letters. For example, the "Ro-Let" Buy This Word Game, and the more complex "Scribe-o", and "Lingo" games with more tiles. None of these seem to have been mentioned at all, but they do form a unique subcategory:

https://img1.etsystatic.com/000/0/6138521/il_570xN.207972547.jpg https://s3.amazonaws.com/bonanzleimages/afu/images/9242/5982/_b-gm_r_bmk___kgrhqj__gwezw56vugebm8zz4hp_w__0_12.jpg http://www.cs.brandeis.edu/~storer/JimPuzzles/SLIDE/CornellCrossword/KeithArticle2011.pdf — Preceding unsigned comment added by 64.85.43.249 (talk) 23:51, 3 July 2014 (UTC)

What Example?
From the introduction to the article: "As this example shows, some sliding puzzles are mechanical puzzles." No example seems to have been provided.