dailysudoku.com

[mtharp]

daily sudoku graded this one as too hard. I got stuck here:

[ZeroAssoluto]

Hi everyone,

first step: Finned Sworfish with number 1 in r3c69,r6c39,r7c36 finned in r2c9 and -1 in r3c7

[immpy]

mtharp, a personal favorite of mine....the finned XY-Wing solves this one.

Notice the 135 combo 35@r1c6; 13@r2c5; and 135@r2c9

...if this were a true XY-Wing r2c9 would include ONLY 15 as candidates, but alas, that pesky fin of 3!! One could say this is an "almost" XY- Wing....
It is indeed an XY-Wing with a fin (of 3). In this case the "fin" is in a pincer cell, but it can show up in the pivot as well. Here the pivot is 13, which sees the 35 and 135 as pincers.

In all cases, either the XY-Wing is true, or the fin is true....one must go about trying to solve using either the XY-Wing as a basis, OR using the fin to fill r2c9 as a basis.....

Upon trying the XY-Wing as if true, a contradiction arose midway thru the process of mock-solving....hence, the fin (3) was next to be plugged into r2c9....and this easily solved the puzzle...one still had to use basics in spots for common line, row, box eliminations; but it became basically stte.

Thank you mtharp.....very nice scheme, and a great puzzle.....

p.s. I usually avoid swordfish, they are too hard for me to spot, and there are other ways to crack most puzzles......

immpy

[mtharp]

Gianni, thanks. Elegant solution. I will need to work on spotting swordfish, included the finned variety.

Mike

[mtharp]

immpy, thanks for the help. I generally prefer to recognize a pattern, but as you point out swordfish are hard to see (at least for me). Mock solving seems a bit like guessing to me, but at least the finned XY-Wing gives you a logical place to start.

Again, thanks.

Mike

[Marty R.]

[mtharp]

Marty, it appears it can be reversed and played either way, albeit with a different set of reasonably basic moves. With your interpretation of the fin, the solution was:

chicago 1/5/18 v2

[ZeroAssoluto]

[keith]

I think that much of this is a waste of time. For years I have been solving Sudoku by looking for strong links in a single variable. Here is the puzzle:

[Marty R.]

[mtharp]

Keith, your solution seems beautifully straightforward.

As a relative newcomer, I am probably asking questions that have been answered before, but how do you pick which strong links to use for eliminations?

Why not use (what I understand to be) strong links r2c5, r8c5 and r6c3, r7c3 to eliminate the 1 in r8c1.

Thanks for wading in.

Mike

[immpy]

Keith's post is a priceless gem of information. Highlight it, copy it, paste it somewhere where you can read it often. It should be required reading for any and all Sudoku solvers. For me, this is going to shed a whole new light on some super hard puzzles. One great light bulb has been brightly illuminated, while about 10 other dimly lit bulbs (fish, loops, multi-coloring, Medusas, wraps, traps, etc.) can be left just as they are (dim at best).

Thank you Keith, I am heading down a whole new road enthusiastically.

immpy

p.s. After all, would I really be able to spot a finned Mutant Leviathan even it was screaming at me?? LOL

[keith]

Mike and immpy,

Thank you! I am in the process of writing this down.

You can make eliminations by pairing strong links in any order. The other eliminations will not go away. Make the elimination, modify the diagram, and look for more.

A good question, which I have already answered, is how do you know which candidates to draw the strong link diagrams for?

Take a look here:

http://forum.enjoysudoku.com/finding-the-possibility-of-single-candidate-eliminations-t32748.html

and study the example. In most puzzles, after basics, the possibility of strong link eliminations will lie with only one or two candidates. You draw strong link diagrams only for those candidates.

I have also written an article on how to do basics. It is here:

http://www.dailysudoku.com/sudoku/forums/viewtopic.php?t=8180&sid=cbf58c9d0597624a2013d1249ba2a1df

The idea of "sweeping" the floors and the chimneys as a first step is very useful. It also starts to look at the big picture: What is the pattern of each unsolved candidate?

More to come.

Keith

[keith]

[keith]

[keith]

From the Sudoku Player's Forum:

[immpy]

I just wanted to revisit this thread and comment that in the most recent months after this thread was originally posted, I have been able to solve some very difficult puzzles; due mostly to my ability to identify the STRONG LINKS! When it seems a puzzle offers nothing recognizable (as to the techniques I am familiar with and capable of implementing), identifying the strong links and making the eliminations they expose will eventually bring the "breakthrough" moment. Just one candidate elimination in a particular cell can make such a great difference. The strong links will reveal the proper elimination. It may not solve the puzzle in one move in and of itself in each case instance (although it is possible), but it surely can take a puzzle that I have stalled out on and allow me to resume my quest.

Thanks again to all of the community, and a special thanks to Keith. Without this source of knowledge, I would still be floundering on the difficult puzzles. Strong links are most definitely a "key" to my solving ability.

cheers...immp

[keith]

[keith]

I have, at last, written down my thoughts on Single-Digit Strong Links.

http://www.dailysudoku.com/sudoku/forums/viewtopic.php?p=39993&sid=b087e024632f261bcb9135dcfe4e77d1#39993

Thanks to immpy for the final nudge.

(That post will not show up as new, since I have edited a placeholder.)

Keith