Y Wing Style Proof for March 18, 2007

The following is an illustrated proof for the Tough Sudoku of March 18, 2007. The primary key used in this proof is Y wing styles.

You may need to refer to previous blog pages to understand this proof. Links to these pages are found to the right, under Sudoku Techniques.

At many times during this illustration, there are other steps available. It is not the goal of this page to show every possible step, but rather to illustrate steps that, taken together, unlock this puzzle.

The information on the following blog pages is required to understand this page:

The illustrations of forbidding chains used in this proof will share the same key:

  • black line = strong inference performed upon a set (strong link)
  • red line = weak inference performed upon a set (weak link)
  • black containers define a partioning of a strong set
  • candidates crossed out in red = candidates proven false
Strong and weak need not be mutually exclusive properties.

Puzzle at start


Puzzle start

A few Unique Possibilities are available here:

  • d7 = 6% cell (naked single)
  • f3 = 9% cell

Hidden Pair 59

Hidden Pair 59

As noted in previous blog pages, the symmetrical locations of the solved cells highlit above allow one to easily locate the Hidden Pair 59 at gi2. This technique is often best applied before entering the possibilities.

The Hidden Pair 59 at gi2 is shown as a continuous nice loop forbidding chain. One could write this step as a forbidding chain:

  • g2=5 == i2=5 -- i2=9 == g2=9 => ab2≠9, f2≠5, g2≠248, i2≠1248
No matter how one understands it, the eliminations allow one to advance the puzzle a bit further:
  • f2 = 3% cell
  • f1 = 5% cell
  • i8 = 1% column (hidden single, 1 is unique in column i)
  • i7 = 4% column
  • g3 = 4% column & box
  • b8 = 3% cell

Alternate to the last step: Naked Triple 128


Naked Triple 128

If one elects to fill in the possibilities before looking for eliminations, then it is probably more likely that one will miss Hidden sets and instead find Naked sets, such as the triple 128 illustrated to the left. This is really little more than a difference in solving style.

One may wonder: why do I seem to pound Hidden sets?

In a very real sense, a true and deep understanding of Sudoku can only be achieved by recognizing the interplay between Hidden strengths and Naked strengths. By this I mean:

A naked strength is a cell limited to some number of candidates, such as one sees in the naked triple

A hidden strength is a candidate limited to some number of locations in a house, as one sees in a hidden pair, or for that matter in Locked candidates

End of Digression

Locked Candidates 2 & 8

Some Locked Candidates

Illustrated above are two seperate steps. One could view it as one step, if one uses the aforementioned triple 128. Instead, I have illustrated:

  • Locked 2's (in box h2) at h23 => h68≠2
  • Locked 8's (in box h2) at h1,23 =>h689≠8
If one had just used the triple 128 previously noted, these steps are not required. However, to get to the same puzzle state, one would now need to use at least the 5s in box h2, else f12 would still be unsolved.

Y wing

Y wing

Illustrated above is the Y wing style that most sudoku solving sites discuss: the Y wing. The cell, f8=27, is the vertex of this Y wing.

  • f8=2 => f6≠2 => f6=6 => h6≠6
  • f8=7 => h8≠7 => h8=6 => h6≠6
As a forbidding chain or Alternating Inference Chain (AIC), one can also write:
  • h8=6 == h8=7 -- f8=7 == f8=2 -- f6=2 == f6=6 => h6≠6
After making this elimination, a few more cells solve:
  • h6 = 3% cell
  • d4 = 3% column & box
  • g9 = 3% row, column & box
  • d2 = 1% column
  • e3 = 8% cell
  • e1 = 4% cell
  • d6 = 4% column & box

Some steps not required

After solving the cells noted above, the following steps are neither required in this proof, nor are they illustrated.

  • Locked 9s at d89 => e9≠9
  • Locked 7s at h89 => i9≠7
  • Locked 6s at h89 => g8≠6
  • Locked 2s at g78 => g4≠2

Y wing Style that unlocks the puzzle

Y wing style

Illustrated above is a very easy to find Y wing style. It is easy to locate if one has followed previous blog pages regarding puzzle mark-ups, as cell b4 would contain circled 1 and circled 8, and h2 would also contain circled 1 and circled 8. This indicates one should look for a Y wing style on 18. Here the 1s in column b, limited to b14, are the vertex of this Y wing style:

  • b1=1 => h1≠1 => h1=8 => c1≠8
  • b4=1 => b4≠8 => c6=8 => c1≠8
Note how this representation is completely analagous to the Y wing illustrated previously. Hence the name that I have coined: Y wing Style.

As a forbidding chain, one can write:

  • h8=8 == h8=1 -- b1=1 == b4=1 -- b4=8 == c6=8 => c1≠8
Again, notice how similar this representation is to the forbidding chain representation of the Y wing.

Soapbox Digression

I supppose that the reason that Y wing styles, in all their many forms, are not treated as part of a greater whole at other sudoku solving sites is: A very unreasonable prejudice against Hidden Strengths in favor of Naked Strengths. As noted previously, Naked strengths are only 25% of the information in a Sudoku. Hidden Strengths are the other 75%. Regardless of whether one finds hidden strengths harder to find (hence the word, hidden), it is senseless to place less value upon them.

I somewhat favor using the following terms:

  • Hidden Strength = Strength in location
  • Naked Strength = Strength in cells
as it seems less prejudicial. Frankly, though, nomenclature is of little importance (to me).

The point that I am trying to drive home, time and time again, is that it is the interplay between these two strengths that one must truly understand in order to master Sudoku Solving. This is even the case in the simple task of locating Unique Possibilities!

Back to the puzzle

After c1≠8, the puzzle is reduced to naked singles (%cells) (singular strength in cells) to the end.

Done

Proof

  1. Start at 22 filled - the given puzzle. Unique Possibilities to 24 filled. (UP 24).
  2. Hidden pair 59 at gi2 forbids f2=5, ab2=9, g2=248, i2=1248 UP 30
    1. Locked 8's at h123 forbids h689=8
    2. Locked 2's at h23 forbids h68=2
    3. Y wing: h8=67, f8=72, f6=26 forbids h6=6 UP 37
  3. Y wing style: c6,b4=8, b41=1, h1=18 forbids c1=8 UP 81
  • Sets: 2 + 2(1) + 2(3) = 10
  • Max depth 3 at steps 3.3 & 4
  • Rating: .03 + 2(.01) + 2(.07) = .19

Post Blog note

One may note that in steps 3.3 and 4 above, I merely listed the strong sets considered. For every possible Sudoku elimination, such a listing is sufficient to prove the elimination. This is because:

  • Strength, regardless of type, is specific to the specific Sudoku puzzle & grid
  • Weakness, regardless of type, is universal to the specific Sudoku puzzle & grid

6 Comments
Indicate which comments you would like to be able to see
Steve  From Ohio    Supporting Member
Check out my page
Since I am a relative late-comer to the world of Sudoku solving, I hesitate to introduce my point of view at Sudoku Forums. Much of the work done by others at these forums towards solving Sudokus is very insightful and powerful work.

Nevertheless, when I have tried to post justifications for eliminations at some of these forums, my ideas are not well received. There seems to be a somewhat unreasonable prejudice at most Sudoku Forums towards justifying an elimination by using a technique that has a NAME.

For whatever reason, justifying an elimination using a technique that has an accepted name is viewed as eloquent. Justifying an elimination using a technique that is equivalent in logic and complexity, but lacks an accepted specific name, is viewed as lacking eloquence.

My point of view is radically different regarding eloquence. To me, an idea is more eloquent when it is expressed generally.
Specific and small applications of an idea are, to me, similar to a specific car. Regardless of how pretty and/or spectacular a particular vehicle may be, the true eloquence lies in the inspiration of invention. Thus, the idea of CAR is always more eloquent
than any car.

This is not to say that I regard my work in this blog as being eloquent. Nothing could be further from the truth. Instead, I recognize that the body of work that I am creating here leans entirely upon the work of those who first studied this simple puzzle. Eloquence is not relevant.

Instead, what is relevant is truly understanding the base inspiration for tackling these puzzles. Thus, when I wax mathematical in my approach, it is merely a reflection of respect for the body of logical analysis that predates my somewhat clumsy approach.

17/Mar/07 11:24 PM
JRD  From UK
Steve,

I use a slightly different method for finding hidden pairs. After making the easy moves, I fill in the possibilities where there are just places in the block,column or row. If there are more than two places I ignore them for the moment. The resulting diagram gives much the same information as your 'puzzle markup' and is much less cluttered. Any hidden pair is now revealed.
18/Mar/07 2:25 AM
Dave  From Minnesota
Your solving approach seems to threaten the establishment, Steve.

IMO the beauty of forbidding chains is their capacity to logically describe 'named' techniques, while adding an even more powerful tool, a language which can universally convey the proofs.

Don't be discouraged with the less than warm reception. Beauty is in the eye of the beholder!

BTW, speaking of names, I have been struggling with what I call 'confound chains', otherwise known as conditional or complex (compound?) chains. I posted a proof variation on the 11/Mar/07 page 2 blog that uses strong set {f3=3,f4=3,{pair 56 at f34}} in place of {f3=3,f3=5,f3=6} which produced the same eliminations.



18/Mar/07 5:12 AM
Vireya  From AU
Thanks for that, Steve. This is the first time I have understood the 'Y Wing' concept (the 1 8 one above). Although I don't yet know how you found it, I finally get what to do with it. I will try reading over your previous blogs to see what I have missed about how to find these critters.
18/Mar/07 9:57 AM
Steve  From Ohio    Supporting Member
Check out my page
Hi JRD!
I do not quite understand what you are trying to convey. If you could give me a picture, or an alternatve explanation, it would be much appreciated!
Better methods for revealing the information the puzzle is trying to unload is something I find very interesting!
BTW, you can email me at solidsudoku@yahoo.com.
18/Mar/07 3:39 PM
Steve  From Ohio    Supporting Member
Check out my page
Hi Dave!
Thanks for letting me know about the proof that you posted! I always enjoy studying other proofs!
18/Mar/07 3:59 PM
Please Log in to post a comment.

Not a member? Joining is quick and free. As a member you get heaps of benefits.

Join Now Login