Using X-Wing & Swordfish in the wild

Advanced methods and approaches for solving Sudoku puzzles

Using X-Wing & Swordfish in the wild

Postby katymeredith » Thu Sep 15, 2005 7:22 pm

Hi !

I think I understand the more advanced techniques like X-Wing and Swordfish in theory - I just can't see how you would spot when to use them on a real puzzle.

Do you have to have candidates marked out already and just run through each column in turn, or are there visual clues in the cells already completed that say 'hmm, look out for a swordfish here!'

The reason for asking is that my current way of solving is to do some basic completion with naked singles and minirows and minicols first, then quite early on go through the entire grid and mark candidates. But I've got so bored doing that I want to start trying to solve moderate puzzles without any candidate marking at all.

I managed today's Torygraph moderate without using my pen except for completing the cells, and I'm so geed up with that I want to try more!

Thanks !
katymeredith
 
Posts: 5
Joined: 15 September 2005

Postby PaulIQ164 » Thu Sep 15, 2005 7:42 pm

It ought to be possible to do puzzles in the papers without much/any candidate marking - it's how I like to do it. You certainly shouldn't ever need a Swordfish or an X-Wing (except perhaps the odd X-Wing in the Sunday Times). Only problem is that as I understand it Daily Telegraph ones sometimes require trial and error (or extremely advanced techniques that are hard to manage in newsprint), but as long as it's not one of those puzzles, it should be possible to do them with just techniques you can do in your head - though probably not without more than a little practice.
PaulIQ164
 
Posts: 533
Joined: 16 July 2005

Postby Nick67 » Fri Sep 16, 2005 3:27 am

Do you have to have candidates marked out already and just run through each column in turn, or are there visual clues in the cells already completed that say 'hmm, look out for a swordfish here!'


Alas, I don't think there are any easy ways to spot patterns
like X-Wing and Swordfish using pencil and paper.
One thing you could do would be to mark all candidates
in all cells, and then carefully highlight every candidate, using
a different color for each number. Then, for example,
you could see all the cells with the candidate 6, and you
could check if these cells formed a pattern. Of course,
the marking and coloring would be extremely tedious!

But I've got so bored doing that I want to start trying to solve
moderate puzzles without any candidate marking at all.


Yes, candidate marking is a drag!
This leads me to think you might be better off using
a Sudoku program (which marks candidates automatically),
instead of doing puzzles with paper and pencil.
Nick67
 
Posts: 113
Joined: 24 August 2007

Postby emm » Fri Sep 16, 2005 4:45 am

Nick67 wrote:Yes, candidate marking is a drag! This leads me to think you might be better off using a Sudoku program (which marks candidates automatically), instead of doing puzzles with paper and pencil.


I have won that fight with the purist in me that demanded that I personally enter every pencilmark with my own hand and that everything be done without hints, solvers or filters, but now I find it frustrating that some of these solvers do not recognise my higher abilities as far as the pencilmarks go.

They are only too happy to enter all the candidates (this is helpful) but refuse to delete them unless that number is a ‘given’ in that row or column (this is not). If I delete pencilmarks because of a higher strategy that is evolving in my head, my decision is overridden and they are reinserted at the next move resulting in a frustrating battle of wills which of course I cannot win.
emm
 
Posts: 987
Joined: 02 July 2005

Postby katymeredith » Fri Sep 16, 2005 5:19 pm

Thanks all:D

Paul - thanks for the info about X-Wing & Swordfish - I can sleep a bit easier now knowing that I can probably live without really having them available.

Nick & Em

It's not so much the actual writing the candidates down that I found boring; it's the fact that when I was doing that the solution turned into a mechanical process of going through the units one by one. So I could always solve a fiendish just by spending long enough checking for locks(?) and pairs(?) <apologies, haven't really got the terminology sorted yet:!: >

It's that that I found boring.

However, now I've started trying to solve them without marking candidates, it's become a much more satisfying experience. True, I'm only half way through the Fiendish in today's Times but that just means I've got half a challenge ahead of me, and it was even intellectually satisfying completing the mild.

I'd been thinking about just forgetting the duko but this has really reignited my interest.

:D:D

I don't want to use a solver program, as I've got no willpower at all & I'll end up using it when I should just prod the puzzles with my grey bits, and I'll lose that intellectual satisfaction I'm just regaining. Plus, after a day at work gazing at a screen, the last thing I want to be looking at on the train home is another computer:)

Thanks for the replies all:D
katymeredith
 
Posts: 5
Joined: 15 September 2005

Postby emm » Fri Sep 16, 2005 6:47 pm

katymeredith wrote:apologies, haven't really got the terminology sorted yet

A good place to go for terminology is the sticky called Basic Terms in the General/Program section of this forum.

KatyM, like you I have no willpower (sigh!). I too resisted the solvers (don't they sound like cheating!) but the truth is that the very hard puzzles require so much tedious donkeywork that I'd never get past the filling in stage if I didn't use them - unless you can hold 100's of pencilmarks in your head (personally I don't like to stress my grey matter to that extent, it makes my brain hurt). So far I have absolutely resisited the temptation to use a hint, though I do get lots of them off people in this forum (but that's different.)

PS : If I had any money I'd bet $$ that one day you will want to try swordfish & Xwing!:)
emm
 
Posts: 987
Joined: 02 July 2005


Return to Advanced solving techniques

cron