The New Sudoku Players' Forum

by **ronk** » Sat Aug 06, 2011 1:15 pm

champagne wrote:
ronk wrote:champagne, does your skfr get a 7.7 rating for any of these?
here is the rating of skfr for that lot

Thanks ... and that meets my needs without further analysis on your part. I'll add details to this post with an edit a bit later.

by **daj95376** » Sat Aug 06, 2011 4:20 pm

Another perspective of champagne's networks: based on the assumption 8r7c6

Code: Select all: +-----------------------------------+ | . . . | . . . | 8 . . | | . . . | . 8 . | . . . | | *8 *8 . | . . . | . . . | |-----------+-----------+-----------| | 8 8 @8 | @8 . . | . . @8 | | 8 . @8 | @8 . 8 | . . @8 | | *8 *8 . | . . -8 | . . . | |-----------+-----------+-----------| | . 8 -8 | . . +8 | . . . | | . . . | . . . | . 8 . | | *8 *8 . | -8 . . | . . . | +-----------------------------------+ 8r7c6 => -8r6c6,r9c4 => 8r369c12 (* invalid) -or- 8r7c6 => -8r7c3,r9c4 => 8r45c349 (@ invalid)

Either way, we get: r7c6<>8 => r9c4=8 => r45c4,r9c12<>8

by **champagne** » Sat Aug 06, 2011 4:39 pm

daj95376 wrote:Another perspective of champagne's networks: based on the assumption 8r7c6

8r7c6 => -8r6c6,r9c4 => 8r369c12 (* invalid)
-or-
8r7c6 => -8r7c3,r9c4 => 8r45c349 (@ invalid)

Either way, we get: r7c6<>8 => r9c4=8 => r45c4,r9c12<>8

This is the kind of quick and nice exit i would add to my solver, but it's requiring time.

may be later!!!!

champagne

by **daj95376** » Sat Aug 06, 2011 9:06 pm

FWIW: My solver used a finned Swordfish and/or a grouped 2-String Kite as the first step(s) in all of these puzzles.

Update: One puzzle used an Empty Rectangle as the first step. However, finned Swordfish were also present.

by **champagne** » Sun Aug 07, 2011 6:29 am

daj95376 wrote:FWIW: My solver used a finned Swordfish and/or a grouped 2-String Kite as the first step(s) in all of these puzzles.

Update: One puzzle used an Empty Rectangle as the first step. However, finned Swordfish were also present.

This is not a surprise.

Here below the file of 7.7 skfr diamonds I stored in my last pattern game.

As far as i can see, they all have very similar pattern.

Anyway, to get a 7.7 or 7.8 rating, it must be a relatively simple "Nishio pattern"

champagne

PS I'll show in a separate post a slightly different path

Hidden Text: Show

by **champagne** » Sun Aug 07, 2011 6:35 am

Here is the second puzzle in the list.

In that puzzle (rated 7.8 by SE), skfr finds 8 eliminations rated 7.7.

A quick check through the log file shows either 4 eliminations or 8 eliminations for each puzzle of the file given in the previous post.
(subject to a deeper check)

contrary to daj's one, my solver does not look for more than the basic fishes, so it would start with a chain.

champagne

Code: Select all: 2 1459 1345 |3479 1346 13679 |8 15679 3579 1459 7 1345 |2349 8 1369 |12369 12569 2359 189 189 6 |2379 123 5 |12379 1279 4 -------------------------------------------------------- 1689 12689 128 |2389 123 4 |5 2679 2379 14569 3 1245 |259 7 19 |2469 8 29 4589 24589 7 |6 235 389 |2349 249 1 -------------------------------------------------------- 3 24568 2458 |1 456 678 |2479 24579 25789 145678 124568 12458 |4578 9 678 |1247 3 2578 14578 1458 9 |34578 345 2 |147 1457 6 chain 1 seq 3r1c9 ~3r1c3 ~3r4c9 3r2c3 3r6c7 ~3r2c6 ~3r6c6 3r1c6 chain 2 seq 3r1c9 ~3r1c6 chain 1 seq 3r2c9 ~3r2c6 ~3r4c9 3r6c7 ~3r6c6 3r1c6 chain 2 seq 3r2c9 ~3r2c3 3r1c3 ~3r1c6 chain 1 seq 3r4c4 ~3r9c4 ~3r4c9 3r9c5 3r6c7 ~3r3c5 ~3r3c7 3r3c4 chain 2 seq 3r4c4 ~3r3c4 chain 1 seq 3r4c5 ~3r3c5 ~3r4c9 3r6c7 ~3r3c7 3r3c4 chain 2 seq 3r4c5 ~3r9c5 3r9c4 ~3r3c4 chain 1 seq 8r6c1 ~8r3c1 ~8r6c6 8r3c2 8r4c4 ~8r9c2 ~8r9c4 8r9c1 chain 2 seq 8r6c1 ~8r9c1 chain 1 seq 8r6c2 ~8r9c2 ~8r6c6 8r4c4 ~8r9c4 8r9c1 chain 2 seq 8r6c2 ~8r3c2 8r3c1 ~8r9c1 chain 1 seq 8r7c6 ~8r7c3 ~8r7c9 8r8c9 ~8r8c3 8r4c3 chain 2 seq 8r7c6 ~8r6c6 8r4c4 ~8r4c3 chain 1 seq 8r8c6 ~8r8c3 ~8r8c9 8r7c9 ~8r7c3 8r4c3 chain 2 seq 8r8c6 ~8r6c6 8r4c4 ~8r4c3

by **ronk** » Sun Aug 07, 2011 9:30 am

champagne wrote:Here is the second puzzle in the list.

In that puzzle (rated 7.8 by SE), skfr finds 8 eliminations rated 7.7.

It's more of the same, Explainer finds finned (row\col) swordfish and your skfr finds finned franken swordfish.

Code: Select all: 2.....8...7..8......6..5..4.....45...3..7..8...76....13..1.........9..3...9..2..6 Explainer: (3)c367\r126 + fin r7c3 skfr: (3)c36b6\r126 + fin r4c9 ==> r12c9<>3 Explainer: (3)r369\c457 + fin r6c6 skfr: (3)r39b6\c457 + fin r4c9 ==> r4c45<>3 Explainer: (8)r349\c124 + fin r4c3 skfr: (8)r39b5\c124 + fin r6c6 ==> r6c12<>8 Explainer: (8)c349\r478 + fin r9c4 skfr: (8)c39b5\r478 + fin r6c6 ==> r78c6<>8

Since these puzzles are diamonds (7.8/7.8/7.8 or 7.7/7.7/7.7), there is little need to post pencilmarks.

by **daj95376** » Sun Aug 07, 2011 9:01 pm

Withdrawn !!!

by **ronk** » Mon Aug 08, 2011 3:15 pm

daj95376, today is a slow day for me.

Relative to your withdrawn post, would the changes in green seem satisfactory to you?

Sudoku Explainer wrote:Nishio Forcing Chains

With this solving technique, we will prove the two following assertions:
If R1C9 contains the value 3, then R2C6 must contain the value 3
If R1C9 contains the value 3, then R2C6 cannot contain the value 3
...
Chain 1: If R1C9 contains the value 3, then R2C6 must contain the value 3 (View 1):
(1) If R1C9 contains the value 3, then R3C7 cannot contain the value 3 (the value can occur only once in the block)
...
(6) If R1C6 does not contain the value 3 and R6C6 does not contain the value 3 (4), then R2C6 must contain the value 3 (only remaining possible position in the column)

Chain 2: If R1C9 contains the value 3, then R2C6 cannot contain the value 3 (View 2):
(1) If R1C9 contains the value 3, then R1C3 cannot contain the value 3 (the value can occur only once in the row)
...
(3) If R2C3 contains the value 3, then R2C6 cannot contain the value 3 (the value can occur only once in the row)

by **daj95376** » Mon Aug 08, 2011 4:44 pm

ronk wrote:daj95376, today is a slow day for me. Relative to your withdrawn post, would the changes in green seem satisfactory to you?

Sudoku Explainer wrote:Nishio Forcing Chains

With this solving technique, we will prove the two following assertions:
If R1C9 contains the value 3, then R2C6 must contain the value 3
If R1C9 contains the value 3, then R2C6 cannot contain the value 3
...
Chain 1: If R1C9 contains the value 3, then R2C6 must contain the value 3 (View 1):
(1) If R1C9 contains the value 3, then R3C7 cannot contain the value 3 (the value can occur only once in the block)
...
(6) If R1C6 does not contain the value 3 and R6C6 does not contain the value 3 (4), then R2C6 must contain the value 3 (only remaining possible position in the column)

Chain 2: If R1C9 contains the value 3, then R2C6 cannot contain the value 3 (View 2):
(1) If R1C9 contains the value 3, then R1C3 cannot contain the value 3 (the value can occur only once in the row)
...
(3) If R2C3 contains the value 3, then R2C6 cannot contain the value 3 (the value can occur only once in the row)

Yes !!!

by **champagne** » Mon Aug 08, 2011 5:24 pm

daj95376 wrote:
ronk wrote:daj95376, today is a slow day for me. Relative to your withdrawn post, would the changes in green seem satisfactory to you?

Yes !!!

I guess that bug is known by all users of SE. It does not change the rating, so we are used to work with it

champagne

by **ronk** » Fri Aug 12, 2011 12:36 pm

ronk wrote:It's more of the same, Explainer finds finned (row\col) swordfish and your skfr finds finned franken swordfish.

After switching to puzzles rated ED=7.9/7.9/7.9, it didn't take long to find a fish equivalent (to an Explainer v1.2.1 Nishio move) that uses a mix of strong inference sets in rows, columns and boxes.

Code: Select all: ..1.2.3...4...5.1.2.....6.5...7...8.3...9.1...7...8...8.2.6...1.3.4...6...4...9.. At the start: *7 . . | . . *7 | . *7 *7 7 . 7 | . 7 . | *7 . 7 . . 7 | . 7 7 | . 7 . ---------+----------+---------- . . . | 7 . . | . . . . . . | . . . | . 7 7 . 7 . | . . . | . . . ---------+----------+---------- . . . | . . *7 | @7 *7 . *7 . #7 | . 7 7 | *7 . -7 *7 . . | . 7 7 | . 7 7 mutant jellyfish(7)r17c7b7\c16b39 + fin: r8c3 + endo-fin: r7c7 ==> r8c9<>7

by **civiliza** » Wed Aug 17, 2011 1:44 pm

I have noticed that while Direct Claiming has a SE rating of 1.9, it does not appear in Options-> Solving Techniques.

Since Pointing and Claiming appear as a pair in said menu, is it safe to assume that the Direct Pointing option also toggles whether or not Direct Claiming takes place?

by **ronk** » Wed Aug 17, 2011 2:20 pm

civiliza wrote:Since Pointing and Claiming appear as a pair in said menu, is it safe to assume that the Direct Pointing option also toggles whether or not Direct Claiming takes place?

I doubt it. Nicolaus Juillerat concluded that exclusions due to Direct Claiming do not exist.

The New Sudoku Players' Forum

Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Re: Oddities of Sudoku Explainer

Direct Claiming not in Solving Techniques

Re: Direct Claiming not in Solving Techniques