Friday, 2 March 2012

Friday Puzzles #146

Onwards with the theme, and we'll have ourselves a lovely Sudoku variant.  The rules are as follows: 1-8 go into each row/column/box, with two 2's in each.  Moreover, the cells containing the 2's are disjoint (so don't touch at either an edge or a vertex).  Think Sudoku crossed with Star Battle then.

I am quite pleased with how this one turned out, given that I started this on something of a whim.  Enjoy!

EDIT (14:18, 2/3/2012): fixing non-uniqueness.
    #177 Twodoku – rated medium
Previous non-unique version (see comment from rob):
All puzzles © Tom Collyer 2009-12.


  1. Hmm... not sure it isn't slightly broken, actually. I'm getting everything unique except the placement of the 3's 6's and 7's in the top third of the puzzle. I don't know if you can check that without breaking your lenten promise, but we'll see if anyone else agrees.

    Even if it needs tweaking, it's a very nice variant and had a lot of good logic along the way.

    - Jack

  2. I've checked my book, and it seems that it'd be impossible to permute those digits in just the top 3 rows without doing anything else lower down too.

    I'll upload a solution file.


  4. There seems to be an alternate placing of the twos around the top right... This is how far I get uniquely: What appears to be an alternate solution:

  5. rob - your first grid agrees with my solution, and as such I can't see anything wrong with the alternative second (and third) grids.

    My initial attempt at the puzzle had a 1 in R3C9 (and a 7 in R7C1). I thought this was too easy, and on taking out these digits I thought I could deduce what they ought to be from the remaining clues. Apparently not - the 1 seems to be key!

  6. 1 was a key in the last puzzle too. :P Nice one by the time I got here anyway.

  7. nice puzzle Tom.... lovely flow with the numbers being put in the grid... If you had not mentioned the issue... I would have solved it using the uniqueness method as a "solver" and come up with the solution... but since the issue was highlighted my solving was more as a "tester" and could see the issue.

  8. I think this variant has the advantage that it isn't as contrived as 90% of the other sudoku variants you see - the no touching logic integrates into the puzzle quite nicely, although is perhaps a little constrained for a classic grid (you can't have a 2 in the centre of a 3x3 box, for example). I'll perhaps give an irregular grid a go (as well as a harder classic grid version) when I get back to making puzzles - and I'll look forward to having a go at anything anyone else comes up with!

  9. Mellow Melon had something similar up (Puzzle #372), and noted there that the 9x9 sudoku regular grid was a big constraint on the positioning of the stars. I do think irregular grids could be the way to go, here.

  10. I'd say it wasn't just similar, it was exactly the same thing. Great minds and all that, haha :)

    I'd disagree that there are as few solutions as Palmer claims (for example, there are two on offer here already!) but most of the solving logic with the regular grid is checking the borders between two 3x3 boxes, and then working out what that means for the rest of those boxes.

    I'll definitely be pursuing an irregular version on my return to puzzling, and Rishi ( ) says he already has one in the works!


Contact Form


Email *

Message *