Tuesday, 19 January 2016

Chinese Sudoku #1

One of the joys of being a member of the international puzzling fraternity is the variety of people from around the world that you get to meet.  Whilst at the World Championships held in Sofia last October I was very pleased to make the acquaintance of some of the Chinese team, a formidable group of solvers whose brilliance is all the more remarkable given their young age.

I was even more pleased to receive a thick wad of printed monthly tests that the BEET "Brain Extremely Exhausted Time" Puzzle Club have put together.  There really seems to be no end to the number of puzzles I could be solving before we all meet again.

Anyhow, we have maintained correspondence and I'm very pleased to have been given permission to share a recent monthly test with a wider audience.

The title of the contest is:
The 16th CSOC - Zodiac
- LINK -
With thanks to YangPei for authoring and Righthand for testing.  It's a lovely set of puzzles with an interesting challenge at the end to tie everything together.

Finally, whilst it seems a little fruitless to provide a link given that I'm sure the majority of my dearest readers won't be able to understand the website, and indeed my would-be dearest readers in China have various issues trying to reach any link on Google's blogger platform, I will do so anyway out of politeness:

http://www.sudokufans.org.cn

Enjoy!

Friday, 8 January 2016

Friday Puzzles #289

Better late than never - and even better when the puzzle isn't broken - here is a Sum Sandwich Sudoku.

Rules: Apply classic sudoku rules.  Additionally, the clues outside the grid indicate all numbers placed in the corresponding direction that are the sum of the numbers placed in the two neighbouring cells in that direction, e.g. 132.  If a number is not clued outside the grid, then it must not be the sum of its two neighbours in that direction.

Enjoy!
    #328 Sum Sandwich Sudoku – rated medium
All puzzles © Tom Collyer 2009-16.

Monday, 4 January 2016

Everything In Its Right Place

So one particular reason I have for keeping this blog going, rather than packing up shop and jumping fully aboard the GM Puzzles train, is that it allows me to maintain an independent voice.  Also, I'm pretty sure I've used that particular Radiohead song title as a post before, but really it's too good not to use for a post about the classification of placement puzzles.

It's always a pleasure to look back through the year's "Best Of" puzzles, however I thought it rather odd that a Scrabble puzzle was classified as a "Number Placement" puzzle.  The justification given was:
Creating a complete puzzle taxonomy is challenging. Here at GMPuzzles we limit ourselves to just five logic puzzle categories, but that means our “number” placement term may seem a misnomer when you see logic puzzles with letters or words like our first winner here. At GMPuzzles, whenever a transformation of letters into numbers or other symbols could leave a fundamentally identical puzzle, we consider “Number Placement” to be the proper categorization.
I think I'd like to challenge the thinking here.  Basically I think all scrabble puzzles are fundamentally object placing puzzles, where the objects being placed are words. Whilst the letters that make up the words are obviously important to placing the words, words have other properties, such as their length, which go beyond any property of their constituent letters which allow them to be placed.

Moreover, a number placement puzzle to me has its numbers constrained by the grid itself in some way - for example a latin square, or a marked inequality, or some outside clue.  I can't really see how the letters in a Scrabble puzzle are constrained by the grid. They are only constrained by the words that they are contained in. Of course some of the scrabble variants on GM puzzles do have additional grid decorations but I'd argue they are still primarily object placement in much the same way that I'd argue a Battleships Sudoku is still primarily number placement, even if a large part of the puzzle is indeed object placement.

I have a couple of corollaries to this line of thinking, which are hopefully food for thought:

1) A trivial spelling mistake in a Scrabble puzzle ought not to be judged incorrect where it is clear that the misspelled word uniquely represents the word which was to have been placed.

2) Once you have decrypted the clues, any crossword becomes an object placement puzzle.  Put this in contrast to Kakuro, which is alternatively known as "Cross Sums", which by virtue of its grid-inherent constraints on the numbers to be placed put it firmly in the category of number placement.

Friday, 25 December 2015

Friday Puzzles #288

That moment when you realise Christmas Day is also a Friday!

Nothing too taxing today as this was rather a last minute affair, but watch your step!  Each of the four shaded regions follows a different rule:

  • odd
  • even
  • fortress
  • renban group
Merry Christmas to all my dearest readers, and as ever, enjoy!

EDIT: I'd managed to leave a couple of clues out.  This should make things a little easier!

    #328 Mini-Matchmaker Sudoku – rated medium
All puzzles © Tom Collyer 2009-15.

Tuesday, 15 December 2015

2015 WSC and Group Theory

I've yet to post any sort of championship report for the WSC/WPC held in Sofia in October. I might still get round to that soon!  But the rationale for this post comes from finishing off the broken/unsolved puzzles from my booklets recently.  I came across this gem of a puzzle.


This puzzle is actually taken from the instruction booklet - however it's not the puzzle itself which interests me, rather the phrasing of the instructions.  The Twin Regions in question are effectively extra regions, which are related by a cyclic permutation - and here is the crucial bit - which might possibly be reversed in order.

Now given the cyclic nature of the constraint, it was reasonable to assume that if the twin regions were not quite parallel rows/columns, then they would at least both go from top edge to bottom or left edge to right.  But if this is the case, then that cyclic permutation couldn't possibly be reversed!

And why is that, you ask?

Answer 1: Any orientation reversing symmetry of the regular nonagon is a relfection whose axis passes through a vertex and the midpoint of the opposite edge.

Answer 2: Lets label the cells in the 1st twin region ABCDEFGHI.  Any order reversing cyclic permutation has a fixed point!  This is bad for Sudoku because that means you will end up with two of the same number in a column.

And why is there a fixed point?  Well, this is because any order-reversing cyclic permutation is in fact a reflection of the set X, and because X has 9 elements, when you start pairing off the elements of X that define the reflection, you will always be left with one left over.  The fixed point!

Don't believe me about the reflection claim?  Well let's take the easy case: the order reversing cyclic permutation.

ABCDEFGHI -> IHGFEDBCA

This pairs off (AI)(BH)(CG)(DF) and fixes E.  It's a reflection in E.

Now, we can get any other order-reversing permutation of X by taking this one and applying a (non-trival) order-preserving permutation.  This will move the fixed point from E, and rotate it round to another fixed point, which will then define another reflection.

Answer 3: The D_18 action on X has 9 distinct stabiliser subgroups!

Contact Form

Name

Email *

Message *