The New Sudoku Players' Forum

[daj95376]

When I was first learning about Sudoku chains, I had a nightmare of a time adjusting to Nice Loop notation. Today, bat999 posted a chain in Nice Loop notation, and I decided to review my notes on Nice Loop notation. After all these years, I've come to the conclusion that the format for Nice Loop notation is just wrong. Here's my notes ... and the chain bat999 posted:

Code: Select all

(SI)  e.g.: ( bilocation  [a]=n=[b]  ) or ( bivalue cell  m-[c]-n )
(WI)  e.g.: ( peers       [d]-n-[e]  ) or ( ?-value cell  m=[f]=n )


bilocation    [a]=n=[b] :  if [a] is not 'n', then [b] is     'n'
bivalue cell  m-[c]-n   :  if [c] is not 'm', then [c] is     'n'

peers         [d]-n-[e] :  if [d] is     'n', then [e] is not 'n'
?-value cell  m=[f]=n   :  if [f] is     'm', then [f] is not 'n'


[r3c1]=7=[r3c7]-7-[r9c7]-9-[r7c9]-1-[r7c5]-9-[r1c4]=9=[r1c1]-9-[r3c1] => r3c1<>9


Here's my updated Nice Loop notation ... and the updated chain bat999 posted:

Code: Select all

(SI)  e.g.: ( bilocation  [a]-n=[b]  ) or ( bivalue cell  m-[c]=n )
(WI)  e.g.: ( peers       [d]=n-[e]  ) or ( ?-value cell  m=[f]-n )


bilocation    [a]-n=[b] :  if [a] is not 'n', then [b] is     'n'
bivalue cell  m-[c]=n   :  if [c] is not 'm', then [c] is     'n'

peers         [d]=n-[e] :  if [d] is     'n', then [e] is not 'n'
?-value cell  m=[f]-n   :  if [f] is     'm', then [f] is not 'n'


[r3c1]-7=[r3c7]=7-[r9c7]=9-[r7c9]=1-[r7c5]=9-[r1c4]-9=[r1c1]=9-[r3c1] => r3c1<>9


Exploded View: alternating strong/weak inferences

Code: Select all

[r3c1]-7=[r3c7]                                                       SI
         [r3c7]=7-[r9c7]                                              WI
                7-[r9c7]=9                                            SI
                  [r9c7]=9-[r7c9]                                     WI
                         9-[r7c9]=1                                   SI
                           [r7c9]=1-[r7c5]                            WI
                                  1-[r7c5]=9                          SI
                                    [r7c5]=9-[r1c4]                   WI
                                             [r1c4]-9=[r1c1]          SI
                                                      [r1c1]=9-[r3c1] WI  =>  r3c1<>9


_

[bat999]

Hi
The notation from sudopedia e.g. =5= and -7- makes sense to me when I look at a picture such as this one attached below...
but if you have a different/better notation then that's OK.

discontinuous.png (7.94 KiB) Viewed 1571 times

[daj95376]

Hi
The notation from sudopedia e.g. =5= and -7- makes sense to me when I look at a picture such as this one attached below...
but if you have a different/better notation then that's OK.

Nice picture. I'm glad that you are satisfied with 5-[r6c5]-9 representing a strong link/inference in the bivalue cell. I'm not.

_

[SteveG48]

One of the things that I've come to like about the Eureka notation is that the alternating strong-weak inferences are always clear. It's easy to spot typos or errors in logic. The Nice notation leaves out the inferences internal to the cells, and makes it hard to follow, IMO.

[ronk]

When I was first learning about Sudoku chains, I had a nightmare of a time adjusting to Nice Loop notation.
...
Here's my updated Nice Loop notation ... and the updated chain bat999 posted:

Code: Select all

(SI)  e.g.: ( bilocation  [a]-n=[b]  ) or ( bivalue cell  m-[c]=n )
(WI)  e.g.: ( peers       [d]=n-[e]  ) or ( ?-value cell  m=[f]-n )


bilocation    [a]-n=[b] :  if [a] is not 'n', then [b] is     'n'
bivalue cell  m-[c]=n   :  if [c] is not 'm', then [c] is     'n'

peers         [d]=n-[e] :  if [d] is     'n', then [e] is not 'n'
?-value cell  m=[f]-n   :  if [f] is     'm', then [f] is not 'n'


[r3c1]-7=[r3c7]=7-[r9c7]=9-[r7c9]=1-[r7c5]=9-[r1c4]-9=[r1c1]=9-[r3c1] => r3c1<>9


So "[a] -n= [b]" represents a strong bilocation link, at least when read left-to-right. Now a strong link would still be a strong link when read right-to-left, but the notation is not symmetrical. I think that's a design flaw for your notation.

[daj95376]

So "[a] -n= [b]" represents a strong bilocation link, at least when read left-to-right. Now a strong link would still be a strong link when read right-to-left, but the notation is not symmetrical. I think that's a design flaw for your notation.

Hello Ron, ___

Yes, I was expecting that reply. So, it appears that I have two choices for NL notation:

1) Accept an NL notation that is consistently confusing in two directions, -or-

2) Use an updated NL notation that makes sense from l-to-r and ignore the fact that it needs editing when read from r-to-l.

Since I doubt if there are very many people who even care to read notation from r-to-l, I'm sticking with (2).

Myth Jellies opened a Pandora's Box when he mentioned that AICs were bidirectional. Turns out, people think their notation must be bidirectional as well. For NL notation, that's a stumbling block. For Eureka notation, all is not a perfect world, either. Just look at some of the horrific chains Ruud created in order for them to be bidirectional. If the notation makes sense from l-to-r, then I'm happy!


Regards, Danny

_

[David P Bird]

One of the things that I've come to like about the Eureka notation is that the alternating strong-weak inferences are always clear. It's easy to spot typos or errors in logic. The Nice notation leaves out the inferences internal to the cells, and makes it hard to follow, IMO.

SteveG48, yes, and that’s one of the reasons I would strongly advise any newcomer to stick with the Eureka notation, and another is the rules are simpler. For example look at how much effort it takes to differentiate between three types of 'discontinuous Nice Loop' in <Pauls Pages> because the internal links in cells aren't shown.

Further, showing the internal links in single cells then extends to other patterns such as Group Nodes, ALSs, URs etc. which the Eureka system can accommodate fairly easily.

[sultan vinegar]

I second DPB, and I can speak from experience because I am reasonably new to sudoku. Maybe back in the day when the only inferences available were ungrouped native sets then you could get away with nice loop notation. Now we have grouped links, split nodes, SK loops and derived inferences through ALS, UR, UL, fish, exocet etc. How the hell are you meant to write them in nice loop notation! The beauty of AICs in that you can mix and match inferences any way you want as long as you alternate strong and weak. Why techniques such as ALS XY wing are still referenced is beyond me. It's just an AIC where all the strong links are through the ALS.

[bat999]

... then you could get away with nice loop notation. Now we have grouped links, split nodes, SK loops and derived inferences through ALS, UR, UL, fish, exocet etc. How the hell are you meant to write them in nice loop notation!...

I suppose that Nice Loop notation is intended to be used with Nice Loops.

On paulspages he shows how to use Complex Nice Loops which will do the same thing as some ALS etc.

But I agree, AICs with Eureka notation will be better/necessary for other techniques such as those you mentioned.

It is "Horses for courses".

[SteveG48]

... then you could get away with nice loop notation. Now we have grouped links, split nodes, SK loops and derived inferences through ALS, UR, UL, fish, exocet etc. How the hell are you meant to write them in nice loop notation!...

I suppose that Nice Loop notation is intended to be used with Nice Loops.

On paulspages he shows how to use Complex Nice Loops which will do the same thing as some ALS etc.

But I agree, AICs with Eureka notation will be better/necessary for other techniques such as those you mentioned.

It is "Horses for courses".

Yes, but since Eureka does things that Nice notation won't do, and the converse appears not to be true, why bother with Nice notation?