Puzzle 20: Masyu

masyu-corners

Another one from the Masyu series. This one I’m happy with. It just wouldn’t have made a good contest puzzle since no one would take the time to find the logical path.

EDIT: Solve at pzv.jp.

10 thoughts on “Puzzle 20: Masyu”

  1. Tualha's avatarTualha

    Very nice one, difficult. I think it would have worked for a contest, though. Took me maybe 20 minutes? But that’s on Inkscape, much easier to play what-if, and this has a lot of what-if.

    Reply ↓
  2. Tualha's avatarTualha

    Seems to me only the very simplest Masyu puzzles involve no what-if at all. I certainly can’t see any way to solve this one without some proof-by-contradiction lemmas.

    Reply ↓
  3. rob's avatarrob Post author

    Let’s say there’s a solution that’s very low on “what if”, assuming the Masyu basics. There’s one particular technique that needs to be applied in a couple of situations, I’ll try to give some hints of increasing explicitness (rot13’d):

    * Vg unf gb qb jvgu juvgr pyhrf.
    * Zhygvcyr juvgr pyhrf va n cnegvphyne neenatrzrag.
    * Gel gb trarenyvmr gur snpg gung n Znflh pna’g unir n fvatyr juvgr pyhr qverpgyl va n pbeare.
    * Ubj bsgra qbrf gur ybbc pebff n qvntbany bs juvgr pyhrf?

    Reply ↓
  4. Tualha's avatarTualha

    I’m sorry, I don’t see what you mean. There are many ways [gur cngu pbhyq pebff gur juvgr pyhrf va n qvntbany], aren’t there? It happens a few different ways in this puzzle.

    Reply ↓
  5. rob's avatarrob Post author

    Sorry for being unclear! You’re right, there’s multiple ways, but every white clue causes the line to be crossed exactly once.

    Check out this (broken) Masyu fragment: http://pzv.jp/p.html?mashu/10/10/0930909901900000000090010033030000

    The top left corner being valid, top right invalid is what I was trying to hint at first.

    Then, in the bottom left corner, there’s a non-obvious segment that can be placed (connect columns 3 and 4 in the last row). In the bottom right corner, you can mark an ‘x’ (between columns 8 and 9 in the last row.

    Reply ↓
  6. Tualha's avatarTualha

    I think I follow you — you’re saying it’s sort of a Jordan curve thing? Even numbers work, because if you start on the “inside”, you end up on the inside? But odd numbers don’t work; you need to pass through once more, even if there’s no white clue there.

    Still, it doesn’t seem like a hard-and-fast rule that can easily be applied. Maybe you can sometimes find an invariant, and either draw or rule out a line, but it seems like it would be hard to be sure of a conclusion.

    Reply ↓
  7. Tom Collyer's avatarTom Collyer

    Hi Rob, I’ve only just come across your blog. Enjoyed this puzzle a lot – took me about 5 minutes to crack in the end. In that respect I find that most hard masyu are a bit trial and error anyway – if I haven’t come across something very often (like this puzzle) then it’s at least semi obvious where I should be looking!

    I’ll look forward to playing with this theme in a puzzle of my own later this year :)

    Reply ↓

Leave a comment