The New Sudoku Players' Forum

[Cec]

Do not confuse my use of (*) for elimination cells with Cec's use of (*) to mark (X) cells. Unfortunately, two different display formats use the same symbol to mean different things. Also note that it would have been better if Cec had used (-) instead of (#) for the elimination cells because convention reserves that symbol to mean something else. "

Both points taken and I will endeavour to conform to convention now that I know what symbols to use. That said, I suspect there are many members, like myself, who are not mathematicians nor familiar with reserved convention of such symbols. To such members, unfamiliarity of symbols (without clarification) can make explanation of solving techniques difficult to follow with possible reluctance by readers to not seek clarification of the meaning of these symbols. Because I stated what the '*' and '#' were related to I'm not sure why there would be confusion when reading my previous post.

I'm trying so hard to get the "Swordfish" but it just ain't clicking. I'm hopin Big or someone can dumb that one down for me..

Because of no response, either to Dogguk's above request nor to Bigtone's subsequent reply and accompanying puzzle to identify a "Sword-Fish", I decided to familiarise myself to better understand this technique. I also assumed Bigtone posted his puzzle for someone to explain how the "Swordfish" pattern appears on a grid and it's relevance to solving a puzzle.

It was not my intention and I apologise if my post gave the impression of trying to improve on the far superior grid presentations and explanations of more experienced members, including the excellent angusj description which was included as a link in my previous post to Dogguk. I should also have mentioned Scanraid as being equally informative in explaining this technique.

Cec

[Bigtone53]

I also assumed Bigtone posted his puzzle for someone to explain how the "Swordfish" pattern appears on a grid and it's relevance to solving a puzzle.

Actually, I posted the puzzle so that Dogguk could have a go at it in the knowledge that there was a swordfish in it somewhere but I am glad that it was of use to Cec in his explanation. Here are two more from the same source in case anyone wants to practise their swordfish skills.

Code: Select all

 *-----------*
 |5..|...|..4|
 |.89|..3|.7.|
 |.67|2..|...|
 |---+---+---|
 |..3|.7.|...|
 |.2.|.6.|.9.|
 |...|.1.|8..|
 |---+---+---|
 |...|..9|43.|
 |.4.|8..|65.|
 |8..|...|..1|
 *-----------*

 


Code: Select all

 *-----------*
 |.8.|...|61.|
 |1..|.9.|2..|
 |2..|..5|..4|
 |---+---+---|
 |9..|...|.7.|
 |...|326|...|
 |.4.|...|..1|
 |---+---+---|
 |6..|8..|..3|
 |..3|.1.|..5|
 |.17|...|.9.|
 *-----------*

 
 

[Cec]

"...Actually, I posted the puzzle so that Dogguk could have a go at it in the knowledge that there was a swordfish in it somewhere but I am glad that it was of use to Cec in his explanation ...."

Struth! I've put my foot in it again!... I'm waiving my white flag and getting out of here.

Cec

[eleven]

I see y-wing easier than swordfish. First puzzle:

Code: Select all

 +---------------------------------------------------+
 | 5    13   12  | 67    89   67    | 239   18  4    |
 |*124  8    9   | 145   4-5  3     |#25    7   6    |
 | 134  6    7   | 2     89   145   | 359   18  359  |
 |---------------+------------------+----------------|
 | 6    159  3   | 59    7    8     | 15    4   2    |
 | 17   2    8   | 345   6    45    | 1357  9   357  |
 | 79   579  4   | 359   1    2     | 8     6   357  |
 |---------------+------------------+----------------|
 |*127  17   56  | 156  #25   9     | 4     3   8    |
 | 39   4    12  | 8     23   17    | 6     5   79   |
 | 8    39   56  | 4567  345  4567  | 79    2   1    |
 +---------------------------------------------------+


y-wing 25, link 2 in col 1, r2c5!=5

[Dogguk]

Thanks!! I get it!!

I'm lovin this damn game.

[Cec]

"I see y-wing easier than swordfish. First puzzle:
......y-wing 25, link 2 in col 1, r2c5!=5"

Hi eleven,

In your above grid which are the three cells that form the Y-Wing pattern.
TIA.

Cec

[eleven]

Now, when you say that, i saw in google, that people also call xy-wing Y-wings. But what i mean, was called y-wing here by "udosuk". Cannot find it again, but found many in puzzles.
When i see a pair in 2 different boxes (25 here), i look, if i can connect them with two 2's (or two 5's), which are the only ones in a house (here col 1). Then one of the pairs must be 5.

[tarek]

[Erased]

tarek

[daj95376]

Cec, I believe what eleven is using is called a Y-Wing Styles

http://forum.enjoysudoku.com/viewtopic.php?p=42297#p42297

It's analogous to a Y-Wing. It's also called a W-Wing -- especially when more than three strong links are involved. I think it's even called a Semi-Remote Naked Pair.

Definitely a nice technique that I plan to add to my new solver as soon as I get the chaining algorithm written.

[tarek]

ah, I see it now daj95376,

My initial post was wrong.

the technique involves ALSs joined by strong links.

I can see it now (ALS-xy meets strong links)......

The basic idea is a number of cells that provide "sticky ends" these sticky ends can go on & "stick" on compatible "sticky ends" of another group of cells.

the cells at the discontinuity could interact by eliminating a candidates in cells "seen" by all of these free ends.

If I'm not mistaken Mike Barker has utilised this fact in his solver with great success.

tarek

[Cec]

Now, when you say that, i saw in google, that people also call xy-wing Y-wings. But what i mean, was called y-wing here by "udosuk". Cannot find it again, but found many in puzzles.
When i see a pair in 2 different boxes (25 here), i look, if i can connect them with two 2's (or two 5's), which are the only ones in a house (here col 1). Then one of the pairs must be 5.

Thanks eleven for reply but still confused. Now about 2am (daylight saving time over here) so will study further tomorrow when my head is clearer. Briefly, some comments confuse me in the terminology used...eg. I referred to the solving technique as "y-wing" only because that's what you called it.

I can understand the logic of the "xy-wing" pattern though, like you, the reason why it is sometimes called "y-wing" further adds to my confusion. You also refer to a "house" which I'm unfamiliar with?

Thanks also Tarek for your explanation but I'll need to read up the "ALS" rules again - and probably again and again - to hopefully understand this technique. I have to confess that the advanced solving techniques beyond say colouring and x-wing have over-taxed my brain but lately I've become a bit more enthusiastic to hopefully learn more.

PS. I've just noticed daj95376 and Tarek have offered further help which is appreciated and I've certainly got some homework tomorrow. Again thanks to you all for helping.

Cec

[daj95376]

[Withdrawn]

[Cec]

Because my previous post contained incorrect assumptions - I should say "rubbish" - I decided to delete it and resubmit this revised post. I apologize for any inconvenience caused to members.

Cec, I believe what eleven is using is called a Y-Wing Styles

As mentioned by daj95376 the Y-Wing Styles pattern is a nice technique and relatively "easy" to identify by the initial criteria of a Naked Pair appearing in two different "units". I'm assuming a "unit" is what eleven meant when referring to a "house".

At the moment I still can't get my head around the "ALS" techniques which Tarek has mentioned but I'll continue to read about these.

Thanks again to all for help.

Cec

[tarek]

At the moment I still can't get my head around the "ALS" techniques which Tarek has mentioned but I'll continue to read about these.

Almost Locked sets are simply subsets that fulfil all but one requiremnt of the "locked subsets".....

I'm sure that you already know Locked subsets like naked triples have "in 3 cells only 3 types of candidates". In an almost locked naked triple you would have "in 2 cells only 3 types of candidates".

This characteristic gives these almost locked sets some applications in our sudoku solving strategies... examples are the ALS-xz rule (where 2 ALSs interact together) & the ALS-xy rule (where 3 ALSs interact together).

Now, I mentioned that almost means all but 1... "almost almost" then becomes "all but 2" & this can go as long as you want.... ALS, AALS AAALS. where each "almost" roughly translates into (-1 cell).

the examples commonly in use are for almost naked subsets.... with the same logic, they can be extended into almost hidden subsets.

All of these are tough to spot, however, they have been used to crack MANY puzzles that were considered EXTREMLY tough & UNSOLVABLE in the past.

tarek

[Cec]

Thanks Tarek for your comments. Your analogy of Almost Locked Sets (ALS) to the situation of an "Almost Locked Naked Triple" where three different candidates are confined to only two cells provides some "insight" but then again I still fail to grasp how an "Almost Locked Naked Triple" can lead to exclusions of candidates with only two cells occupying the three candidates.

You mentioned ALS-xy and ALS-xz rules which I presume are those defined in this Thread. I have read these explanations many, many times and, no offence to the author, I simply cannot comprehend or understand the explanations. Having a common birth year to the opening of the Sydney Harbour Bridge perhaps it's my dwindling brain cells but if there is a "simpler" or alternative explanation I'd be glad to try and digest it.

Cec