The New Sudoku Players' Forum

[Luke]

I enjoyed this one, thought others might too. A nice break from all the grueling diabolicals around, and not bad for a published puzzle.
(From the Set 26, 2008 Mercury News in NoCal.)

Code: Select all

......8...5.6..91.63..1.......8....2.9..6.58.4..5.3.......3..41.16..7.5...7......

 *-----------*
 |...|...|8..|
 |.5.|6..|91.|
 |63.|.1.|...|
 |---+---+---|
 |...|8..|..2|
 |.9.|.6.|58.|
 |4..|5.3|...|
 |---+---+---|
 |...|.3.|.41|
 |.16|..7|.5.|
 |..7|...|...|
 *-----------*
 

[udosuk]

(From the Nov 26, 2008 Mercury News in NoCal.)

Hopefully, that piece of newspaper also tells you who is the next US president and who has won the World Series. Plus all the info about the share prices etc will make you a billionaire!

Anyway, for the puzzle, it took me a rare vwxyz-wing to crack.

After SSTS (including a turbot fish on 7):

Code: Select all

+----------------------+----------------------+----------------------+
|#1279  *27     1249   | 3      24579  2459   | 8      267    4567   |
|-278    5      248    | 6      2478   248    | 9      1      3      |
| 6      3      2489   | 2479   1      24589  | 24     27     457    |
+----------------------+----------------------+----------------------+
| 157    67     15     | 8      479    149    | 1346   3679   2      |
|*12     9      3      | 1247   6      124    | 5      8      47     |
| 4      2678   128    | 5      279    3      | 16     679    679    |
+----------------------+----------------------+----------------------+
| 2589   28     2589   | 29     3      6      | 7      4      1      |
|*239    1      6      | 249    2489   7      | 23     5      89     |
|*239    4      7      | 129    2589   12589  | 236    2369   689    |
+----------------------+----------------------+----------------------+

vwxyz-wing:
r1c1 from {1279}
=> r1c2+r2589c1 can't be [72139|72193]
=> r2c1 can't be 2

Then an xy-wing, a few singles, lock candidates plus a skyscraper on 2 solves it.

Not too shabby a puzzle from a newspaper two months into the future!

[Draco]

Luke451 -- this must've come from the paper (and not the website), yes? The website uses UClick and I've never seen a puzzle this good out of their standard batch (shared amongst many newpapers for online play).

As to a solution, not too different than udosuk, except I hit a different square with a different technique after STSS to get the same result:

Code: Select all

r4c1=5 r4c3=1 r5c1=2
+
r4c3=5 r7c1=5 r2c1=8 ==> r2c1<>2, r7c1<>2


The r2c1 cancellation is the critical one but since r7c1 is also hit by the same chain, may as well show it. Doesn't affect the end result though. Not surprisingly, since we knock out the same critical value, I follow the same final steps as udosuk (though I use single color chaining instead of a skyscraper to knock out the 2's at the end -- same net effect).

Cheers...

- drac

[Luke]

(From the Nov 26, 2008 Mercury News in NoCal.)

Not too shabby a puzzle from a newspaper two months into the future!

Hmm, lemmee check that issue for something you could use...ah, here we go: Chelsea blanked Newcastle United 2-nil last Saturday. You'll thank me come November! (switching off the flux capacitor now and...)

....Back to the Present....

Draco, the puzzle was indeed published in the paper. San Jose Mercury News, Fri Sept 26, 2008. I'll check tomorrow as to who's their distributor.

Udo, I've never heard of a vwxyz-wing, altho I've no doubt it exists (Lord knows, *nothing* gets by you! )

But it seems to be just an ALS, x= 7 and z=2, not exactly rare. I assume that what is rare is this particular type of ALS pattern.

[Draco]

While you're at it, here's a puzzle I generated to, more or less, match the level of the one you posted Luke451...

Code: Select all

.7.....95..5...6.2...4.6......1..8.7.9..6..1.5.1..7......7.4...2.6...1..94.....2.

     7    348 | 238  138  1238 | 34  9   5
1348 138  5   | 389       1389 | 6   348 2
38            | 4         6    |     38   
--------------+----------------+-----------
34   36   234 | 1    349  2359 | 8   456 7
     9    248 | 258  6    258  | 245 1     
5    368  1   | 2389 3489 7    | 249 46  69
--------------+----------------+-----------
138  1358 38  | 7         4    | 359 356 69
2    358  6   | 3589 389  3589 | 1         
9    4        |      13   135  | 35  2     


As with the SJM News puzzle, nothing too diabiolical.


Cheers...

- drac

[udosuk]

Hmm, lemmee check that issue for something you could use...ah, here we go: Chelsea blanked Newcastle United 2-nil last Saturday. You'll thank me come November!

Mmm, believe it or not that result isn't as certain as say Obama winning the election or one of Angels/Dodgers becoming the champ. For the moment I'll say Chelsea 4 Newcastle 1. Let's come back by that time and see who will be more correct!

As for Draco's nice puzzle, I can offer 2 different solving paths:


After singles, locked candidates, plus a skyscraper on 5 (and another single):

Code: Select all

+-------------------+-------------------+-------------------+
| 6     7     348   | 238   138   1238  | 34    9     5     |
| 1348  138   5     | 389   7     389   | 6     348   2     |
| 38    2     9     | 4     5     6     | 7     38    1     |
+-------------------+-------------------+-------------------+
| 34    6     234   | 1     349   239   | 8     5     7     |
| 7     9     248   | 258   6     258   | 24    1     3     |
| 5     38    1     | 238   348   7     | 249   46    69    |
+-------------------+-------------------+-------------------+
| 138   1358  38    | 7     2     4     | 359   36    69    |
| 2     35    6     | 3589  389   3589  | 1     7     4     |
| 9     4     7     | 6     13    135   | 35    2     8     |
+-------------------+-------------------+-------------------+

Ignoring a useless x-wing on 4, there are 2 ways to crack it:


1. The short but tricky way:

r1c45+r2c46 from {12389} must have 1 or 2
=> r6c245+r9c5 can't be [8231]
=> r6c5 can't be 3

-or-

ALS-xy-wing:
ALS A: r1c45+r2c46={12389}
ALS B: r6c24={238}
ALS C: r9c5={13}
restricted common between A,B: x=2
restricted common between A,C: y=1
common between B,C: z=3

Therefore r6c5, seeing r6c24+r9c5, can't be 3.

That opens up an xy-wing @ r4c1+r6c25 which solves the puzzle.


2. The long but easy (or easier) way:

r4c13+r5c3 from {2348} must have 3 or 8
=> r3c1+r17c3 can't be [383]
=> r1c3 can't be 8

-or-

ALS-xz:
ALS A: r34c1={348}
ALS B: r457c3={2348}
restricted common: x=4
common: z=8 @ r3c1+r57c3

Therefore r1c3, seeing r3c1+r57c3, can't be 8.

=> r1c37={34} (naked pair @ r1)
=> Hidden pair @ b2: r2c46={39} (naked pair @ r2)

Turbot fish (kite):
3 @ r1 locked @ r1c37
3 @ c8 locked @ r37c8
Now r1c7+r3c8, both @ b3, can't be both 3
=> r7c3, seeing r1c3+r7c8, can't be 3, must be 8

=> Hidden single @ b4: r6c2=8

r1c45 from {128} must have 1 or 2
=> r6c4+r9c5 can't be [21], must have 3
=> r6c5, seeing r6c4+r9c5, can't be 3, must be 4

-or-

ALS-xz:
ALS A: r16c4={238}
ALS B: r19c5={138}
restricted common: x=8
common: z=3 @ r6c4+r9c5

Therefore r6c5, seeing r6c4+r9c5, can't be 3, must be 4.

And then the puzzle can be solved via naked singles.

[Draco]

Nice udosuk!

I'll save posting my solution (only 1) until I see if Luke or others want to try. Thanks for pointing out that the X-Wing isn't needed. After all my work on solvers (and my lazy reliance on mine), I rarely notice that a SSTS technique isn't needed!

Cheers...

- drac

[Luke]

Nice puzzle, Draco. I'm still nibbling away as time allows, but it's about to topple (so far an x-wing, an AUR [r2c1<>3], two UR's, an X-cycle on 5, and it looks like it's ripe for an XY chain now.)

BTW, SJ Merc puzzles are from "The Mepham Group" and distributed by Tribune Media Services. The Fri 4 star is the only published puzzle I ever look at.

Udosuk wrote

For the moment I'll say Chelsea 4 Newcastle 1. Let's come back by that time and see who will be more correct!

Let's hope the we won't have to dodge the punters if we're wrong or Newcastle fans if we're right!

I'm wondering, is your solution generated by your solver? I usually don't score an ALS unless I trip over it.

[udosuk]

I'm wondering, is your solution generated by your solver? I usually don't score an ALS unless I trip over it.

I wouldn't say "generated" by a solver, but I did have JSudoku as my "apprentice" (to keep track of pencilmark etc). All the ALS I need to make the best presentation myself.

[Draco]

...
BTW, SJ Merc puzzles are from "The Mepham Group" and distributed by Tribune Media Services. The Fri 4 star is the only published puzzle I ever look at.

Thanks for the reference; Mephram apparently supplies the LA Times and Daily Telegaph with puzzles as well. I've coded up downloaders for many different puzzle sites (including these), though apparently I need to revisit my DT code. I'm not getting the same puzzles as Mephram's site shows . The price of reverse engineering puzzle pages I suppose.

Ah well if it was always easy it wouldn't be as much fun, eh?

Cheers...

- drac

[Draco]

My solution (after SSTS, which includes a multi-color exclusion) to get to:

Code: Select all

6    7    348 | 238  138 1238 | 34  9   5
1348 138  5   | 389  7   389  | 6   348 2
38   2    9   | 4    5   6    | 7   38  1
--------------+---------------+-----------
34   6    23  | 1    349 239  | 8   5   7
7    9    248 | 258  6   258  | 24  1   3
5    38   1   | 238  348 7    | 29  46  69
--------------+---------------+-----------
138  1358 38  | 7    2   4    | 359 36  69
2    35   6   | 3589 389 3589 | 1   7   4
9    4    7   | 6    13  135  | 35  2   8

A short chain cracks it to singles:

Code: Select all

9c5=1 r1c6=1 r1c4=2
               and
r9c5=3 r9c6=1 r9c7=5 r8c2=3  ([r6c5<>3] + [r6c2<>3]) = r6c4=3
               ergo
r6c4<>2

Cheers...

- drac
[Edit]: PS - my SSTS steps do include the X-Wing.

[udosuk]

Hmm, lemmee check that issue for something you could use...ah, here we go: Chelsea blanked Newcastle United 2-nil last Saturday. You'll thank me come November! (switching off the flux capacitor now and...)

....Back to the Present....

Now that the future has come and gone how wrong Luke's info was. Luckily I didn't lost money on that! (My prediction was much worst but at least I have a track record for guessing completely wrong results in premier league matches. )

[Luke]

Half wrong. Chelsea did indeed blank Newcastle!

[udosuk]

Half wrong. Chelsea did indeed blank Newcastle!

Too bad you can't bet on how many goals one team allows in a match (at least not the bookers I know of). Perhaps a much more useful info would have been that in the round none of the top 5 teams scored a single goal and none of them won a single match.

[Luke]

Too bad you can't bet on how many goals one team allows in a match (at least not the bookers I know of).

The closest you can get stateside is "Proposition betting."

(I think the mods will ashcan this whole thread into "Off Topic" at any moment! )