## Diabolical puzzle that has JSudoku stumped

Post the puzzle or solving technique that's causing you trouble and someone will help
champagne wrote:First of all, my AICs are following the general rule agreed here for AICs (thanks to Mike Barker who helped me).

Thanks for the detailed response, and I see where you're coming from. The "my" of "my AICs" probably explains my misunderstanding. When someone uses the AIC term, I expect the notation to follow the broadly accepted rules for AIC notation. Where is the "here" of "agreed here"

For user more familair with full tagging, I would have written :
Code:
[] Fp PM mC cE eF 'F' is dead. (which is exactly the same).

From a tagging POV, this makes sense. I've used it myself in advanced coloring. From an AIC POV, however, it ignores that portion of the chain that places the 'F' tags. This would be avoided if the "tagging AIC" were of the form ...

f = F - p = P - M = m - C = c - E = e - F = f, implying 'f' is true
ronk
2012 Supporter

Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

ronk wrote:
champagne wrote:First of all, my AICs are following the general rule agreed here for AICs (thanks to Mike Barker who helped me).

Thanks for the detailed response, and I see where you're coming from. The "my" of "my AICs" probably explains my misunderstanding. When someone uses the AIC term, I expect the notation to follow the broadly accepted rules for AIC notation. Where is the "here" of "agreed here"

I never posted a message in English out of that forum... And Mike (I think he is reliable) gave me the rules to apply to AIC's....

My AICs are AIC is that sense that we have strictly a sequence "strong link" ; "weak link"; strong link" ; "weak link".

They are not a sequence of bi-values tightly connected as were the first AICs, but you have for long passed that step as well.

It seems to me that each one introducing a new type of AIC has to create a new printing form. I had to do it when I introduced ACs, super candidates and memory function. I am prepared to rally any standard although I have a preferred form to express AICs nets that have an important place in "hardest puzzles".

I will consider carefully any change improving the communication. For the time being, I am not sure many guys studied my proposal for "hardest puzzles"

ronk wrote:[ This would be avoided if the "tagging AIC" were of the form ...
f = F - p = P - M = m - C = c - E = e - F = f, implying 'f' is true

You are right .. in that example, but hidden sequences can be anywhere in the chain and the length of the hidden sequence is not limited.

Another point is that I only use the form of AIC ending by two weak links.
This fits with the general logic applied in full tagging method and limits the risk of error for players.

At level one in ful tagging, your chain has an equivalence thru the so called "OR" clearing.
champagne
2017 Supporter

Posts: 6250
Joined: 02 August 2007
Location: France Brittany

champagne wrote:My AICs are AIC is that sense that we have strictly a sequence "strong link" ; "weak link"; strong link" ; "weak link".

They are not a sequence of bi-values tightly connected as were the first AICs, but you have for long passed that step as well.

Today's preferred term is inference, not link.

It seems to me that each one introducing a new type of AIC has to create a new printing form.

There have been additions due to grouped single digits, ALSs, the strong inferences of AURS and other uniqueness patterns, multi-inferences, etc. ... but no one has proposed "jumping" or chopping off one end of an AIC before.

I had to do it when I introduced ACs, super candidates and memory function.

Are your ACs still AHSs, as I recall from prior posts I have no idea what you mean by "super candidates" and "memory."

I am prepared to rally any standard although I have a preferred form to express AICs nets that have an important place in "hardest puzzles".

I will consider carefully any change improving the communication.

That's encouraging.

For the time being, I am not sure many guys studied my proposal for "hardest puzzles"

I saw your early posts on this forum ... and gave up because of your notation. I suspect others did too. I give you credit for improving the notation to where it is now.

champagne wrote:
ronk wrote:[ This would be avoided if the "tagging AIC" were of the form ...
f = F - p = P - M = m - C = c - E = e - F = f, implying 'f' is true

You are right .. in that example, but hidden sequences can be anywhere in the chain and the length of the hidden sequence is not limited.

I don't know what you mean by hidden sequence either.

Another point is that I only use the form of AIC ending by two weak links.

That was ... and is ... one of my objections. Traditional AICs end with strong inferences.

At level one in ful tagging, your chain has an equivalence thru the so called "OR" clearing.

That has no meaning for me. You might as well have written it in Greek.
ronk
2012 Supporter

Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Hi ronk

I am hesitating on how to go ahead with our discussion.

1) It seems that full tagging offers a solving potential that worth considering it: efficiency, full coverage of known puzzles..
If it had not been the case, I would have stopped for long writing posts on that forum.

Full tagging has its own coherency and is working on tags, not on candidates. (tags are containers of “equivalent” candidates);

This means that for about 2.5 years, we (users of that method) are producing AICs with jumps and often “chopping off one end”.
We are also using AIC’s with weak inferences at each end . This is quite normal. You have to keep in mind that no tag is defined (before entering level 4) without having a strong inference. So when you can write
F - p = P - M = m - C = c - E = e - F ,
You know that you have also
f = F - p = P - M = m - C = c - E = e - F = f,
On the other way, it is exactly the same to say ‘F’ is dead or ‘f’ is qualified. I was not waiting for a difficulty here.

The second form is normally preferred in that situation
f = F - p = P - M = m - C = c - E = e which gives a clearing potential thru the condition (f or e).

2) As long as we stay with relatively simple puzzles, differences between full tagging and other methods are relatively small. I thought I could use directly the results of my solver to communicate quickly. Your reactions show that I was wrong.

3) I thought you had a sufficient knowledge of the method to use some key words. Again I was wrong, so forget all Greek and Chinese parts of the post..

I think I have still to answer a small point

ronk wrote:Are your ACs still AHSs, as I recall from prior posts

Unless I misunderstood what I red, yes. I kept my name AC to stay coherent with AC2, the entity used at level 4 (Chinese again sorry)

To close that post on a positive view, I developed a solution for Golden Nugget and other “Hardest” puzzles. I find these solution boring and having very little chances to be used by players.

Reversely, I worked on Metcalf and strcmk puzzles at the top of JPF list. These have a good chance to be studied one day by players not using a computer;
I hope I’ll have finished to publish Metcalf solution within few days. Strcmk is very similar;
champagne
2017 Supporter

Posts: 6250
Joined: 02 August 2007
Location: France Brittany

Code: Select all
` *--------------------------------------------------------------------* | 28      2789    5     | 179     179     3    | 127     6       4   | | 2346    234679  2369  | 14579   14679   1457 | 1237    1238    138 | | 1       3467    36    | 8       467     2    | 357     35      9   | |-----------------------+----------------------+---------------------| | 23456   1       23689 | 345     248     458  | 2349    7       36  | | 7       234     238   | 6       1248    9    | 1234    123     5   | | 23456   234569  2369  | 13457   1247    1457 | 8       1239    136 | |-----------------------+----------------------+---------------------| | 9       368     4     | 2       5       68   | 13      138     7   | | 2358    2358    1     | 479     4789    478  | 6       59      38  | | 568     568     7     | 19      3       168  | 59      4       2   | *--------------------------------------------------------------------*`

1) [r4c3]=9=[r4c7](=4=[r5c7]-4-[r5c2])-9-[r9c7]-5-[r9c12]=5|3=[r7c2](-
-3-[r5c2]-2-[r4c3])-3-[r7c78]=3=[r8c9]-3-[r46c9](-6-[r4c3])-1-[r5c78]
=1=[r5c5]=8=[r5c3]-8-[r4c3], => r4c3<>2,6,8.

2) [r5c5]=1=[r5c78]-1-[r6c9]=1=[r2c9](-1-[r2c5])=8=[r8c9]-8-[r8c456]-
-9-[r9c4]-1-[r6c4], => r2c5, r6c4 <> 1.

3) [r7c8]-3-[r7c2]=3|5=[r9c12]-5-[r9c7]=5=[r8c8]-5-[r3c8]-3-[r7c8], => r7c8<>3.

4) [r2c9]=8=[r8c9]=3=[r7c7]-3-[r7c2]=3|5=[r9c12]-5-[r9c7]=5=[r8c8]-
-5-[r3c8]-3-[r2c9], => r2c9<>3.

5) [r12c7]-1-[r2c9]-8-[r8c9]-3-[r7c7]-1-[r12c7], => r1c7, r2c7<>1.

6) [r9c6]=1=[r6c6]-1-[r5c5]=1=[r5c7]=4=[r4c7]=9=[r6c8](=2=[r5c8]-2-
-[r5c5])-9-[r8c8]=9=[r9c7]-9-[r9c4]-1-[r9c6], => r6c5<>1, r5c5<>2, r5c7<>3, r4c7<>3,

7) [r5c8]=3=[r5c2]-3-[r7c2]=3=[r7c7]-3-[r23c7]=3=[r3c8]-3-[r5c8], => r2368c2<>3, r6c8<>3.

8) [r2c3]=2=[r6c3]-2-[r5c2]-3-[r4c3]-9-[r2c3], => r2c3<>9.

9) [r2c3]=2=[r6c3]=9=[r6c8]-9-[r8c8]-5-[r3c8]-3-[r3c3]-6-[r26c3], => r2c3, r6c3<>6, and the puzzle is solved.

Thanks Udosuk for this nice puzzle.
Carcul

Posts: 724
Joined: 04 November 2005

Hi Carcul,
Sorry... but I can't stop to "say hi"... You are one of famous named here (togethter : Jeff, Mike, Ronk, tarek, Myth, Ocean, re'born...)
Very happy to see you again on this Site... Hope, you are live here long long times...
ttt

Posts: 185
Joined: 20 October 2006
Location: vietnam

Hello Ttt. Thank you very much.
Carcul

Posts: 724
Joined: 04 November 2005

Previous