The New Sudoku Players' Forum

[ronk]

Although I don't understand Sue de Coq, I've run enough SdC puzzles through my solver to notice that they often have common eliminations with a short continuous loop constrained to two boxes. I further observed this taletell pattern in many of the continuous loops.

The doubly-linked ALS xz-rule (aka ALS mutual exclusion rule) subsumes the Sue De Coq (SDC) technique ... and the distributed-disjoint-set (DDS) technique subsumes the doubly-linked ALS xz-rule.

Makes one wonder ... why all this attention currently on the SDC

[storm_norm]

Glyn: this is quite fascinating. obviously, my experience to this point with sudoku has me planted on the AIC side of the fence. to look for sets and ALS and su de coq, I will leave for a program to find which I don't have.

[re'born]

Although I don't understand Sue de Coq, I've run enough SdC puzzles through my solver to notice that they often have common eliminations with a short continuous loop constrained to two boxes. I further observed this taletell pattern in many of the continuous loops.

The doubly-linked ALS xz-rule (aka ALS mutual exclusion rule) subsumes the Sue De Coq (SDC) technique ... and the distributed-disjoint-set (DDS) technique subsumes the doubly-linked ALS xz-rule.

Makes one wonder ... why all this attention currently on the SDC

Doesn't subset counting subsume all of these techniques? Presumably, there is interest in SDC since it is a pattern that (with practice) a human can reliably spot in a short amount of time.

Incidentally, if you think in terms of subset counting, then it is completely trivial to decide the eliminations in the above SDC. There is no worrying about naked quads or anything else. Each of the candidates in a SDC has max multiplicity 1, so any cell outside of the SDC that sees every instance of a digit in the SDC cannot have that digit.

[susume]

StormNorm wrote:

obviously, my experience to this point with sudoku has me planted on the AIC side of the fence.

In the same way, I tend to see many things as Nice Loops because I picked that first to learn after basic techniques. After marking the strong inferences, what caught my eye was the strong links in box 1 on 4, 5, and 2, leading to this short continuous nice loop which subsumed the SDC [and I think is the same as Glyn's AIC]:

Code: Select all

[r1c2] =4= [r1c1] =5= [r3c3] =2= [r3c2] -2- [r7c2] -4- [r1c2]=
gives r5c2<>4, r1c1<>7,8, r3c3<>8, r2c1 OR r2c3 must be 8, etc.

Just another of many possible views.

[ronk]

After marking the strong inferences, what caught my eye was the strong links in box 1 on 4, 5, and 2, leading to this short continuous nice loop which subsumed the SDC:

Code: Select all

[r1c2] =4= [r1c1] =5= [r3c3] =2= [r3c2] -2- [r7c2] -4- [r1c2]=
gives r5c2<>4, r1c1<>7,8, r3c3<>8, r2c1 OR r2c3 must be 8, etc.

A technique which literally subsumes the SDC under discussion would use the strong inference sets of these five cells:

Code: Select all

 .      2479   .     
 789    .      789   
 .      2479   .
---------------------
 .      .      .     
 .      .      .     
 .      .      .     
---------------------
 .      24     .     
 .      .      .     
 .      .      .