croco-puzzle

[nickdeller]

How do people solve Killer Sudoku so quickly? They make my brain ache, for me I did that one relatively fluently and still ended up way down the list.

I've no idea what's obvious and what isn't in sudoku solving, but have you managed to train your brain to see/hear 51492 and INSTANTLY respond with 3678? It speeds things up no end if you can look at what's present and immediately know what's missing.

Specific to solving killers: there are certain cages which DEFINITELY contain a specific set of numbers, much as in kakuro - if you see a two-cell 3 interacting with a two-cell 4, for example, you're likely going to be able to fill one of those two cages with definites straight away. A two-cell 3 all in the same row with a two-cell 7 tells you the contents of both cages.

The 'one row sums to 45, two rows sum to 90' so looking for cages that exactly fill every cell bar one/every cell plus one is helpful for more complex Killers, but tends to be an unnecessarily slow step in the croco-killers I've seen so far. When that technique IS useful though, it's quicker to do the sums modulo 10; if you're only looking for the content of one cell, you already know that you're looking for a one-digit number.

[ronaldx]

The 'one row sums to 45, two rows sum to 90' so looking for cages that exactly fill every cell bar one/every cell plus one is helpful for more complex Killers, but tends to be an unnecessarily slow step in the croco-killers I've seen so far. When that technique IS useful though, it's quicker to do the sums modulo 10; if you're only looking for the content of one cell, you already know that you're looking for a one-digit number.

I think it's probable that I automatically went to this step first when it was unnecessary, although the columns lined up nicely for that modulo 10 method.

[PuzzleScot]

When that technique IS useful though, it's quicker to do the sums modulo 10

Dang - I never thought of doing that - thanks!

I'll repay that with one of my own 'shortcuts' (that everyone probably knows anyway). Sum 1-n = n(n+1)/2. So if you have a 1-7 grid, you know sum 1-7 = 28 right away. [Obviously either n or n+1 will be even]

[ronaldx]

I just noticed that uvo took 21 minutes on today's puzzle... and 27 out of 36 are yet to solve... gosh!! What's the rule when fewer than half solve?

[Calavera]

I just noticed that uvo took 21 minutes on today's puzzle... and 27 out of 36 are yet to solve... gosh!! What's the rule when fewer than half solve?

Then the time of the last solver is taken as median which often results in very high points for those who finished.

[ronaldx]

Good news then! One hour and five minutes Thanks Calavera!

[nickdeller]

That. Was. HIDEOUS!!!!

[ronaldx]

How much of the last 3 std, 25 min, 34 sek should you have been sleeping for, Nick?

P.S. Thanks for bumping my score up too

[nickdeller]

How much of the last 3 std, 25 min, 34 sek should you have been sleeping for, Nick?

Ooh, about 2std, 25 min, 34 sek at the very minimum!

P.S. Thanks for bumping my score up too

I'm guessing it'll get bumped up further than that before you're finished. I'm hoping that mine'll get bumped up further before we're finished, too... though I have my doubts!

There's a lesson to be learned from this experience, but where would the fun be in that?

[Calavera]

Well, as there are always the ones starting directly after midnight times over 10 hours are quite possible. So the only thing you have to worry about is that more than half of the people have enough time to solve this one .


Calavera (puzzle addict who spent the night until after 3 to solve the Inception-Skyscaper)

[PuzzleScot]

A 'lightning' 100m from me! (4 attempts).

[nickdeller]

(puzzle addict who spent the night until after 3 to solve the Inception-Skyscaper)

Good grief, that's an extraordinary looking construction! I hope I've got enough common sense not to attempt to solve it...

[Calavera]

You could start with it's little brother. If you generally like skyscrapers also other puzzles from the user "Hausigel" are recommended (14x14 skyscaper for example ). And - like most puzzles in the portal with such high ratings - they are solvable with a completely logical solving path. Although it is at times really well hidden and might take several attempts.

[david mcneill]

I don't think the solution to the "little brother" puzzle is unique. Anyone else found thi sproblem or am I missing something.

David McNeill.

[Calavera]

After 44 solvers without any comment like that I'm pretty sure it has a unique solution. But you can either post a comment there -"Kommentar schreiben" (with limited visibility for solvers and the author only - "alle, die das Rätsel gelöst haben, sowie für den Autor" in the DropDown) or describe it here. Although it might take longer here as I'd have to solve it again to answer .

[PuzzleScot]

I understand the rules to be:
1) Six 4x4 latin squares (effectively the faces of an imaginary cube)
2) numbers in circles indicate the sum of towers visible from that edge point.

PRIME_BBCODE_SPOILER_SHOW PRIME_BBCODE_SPOILER: Walkthrough for 'little brother' - Written as solved, so there's probably better ways to do it

Cols A-L. Rows 1-16. A1 = TL. grids 1-6 in reading order.
6 between H5 & L9 -> neither of these 2 cells contain a 4.
3 between D1 & H5 -> D1=4.
3 between A1 & L13 -> A1=3; L13=4.
6 between C1 & L10 -> L10 <> 4. (already know C1 <>4)
4 between A1 & L12 -> L12 <> 4 -> L11 = 4 -> G5 <> 4.
Other 6s & 7s -> 4 cannot be in A6/K16; E12/I16; H11(/I11); D5/D6
3 between D5 & E5 -> E5 = 4.
J16 = 4.
B6 or C6 = 4, so K14 = 4. -> I15=4 to complete 4s in grid 6, and C6=4. Also, A6<B6 and K16<K15.
6 between E12 & I16 -> E9/10/11/12 = 4/3/2/1 in that order. -> C8<C7.
4 between D1 & L9 -> K9 <>4.
3 between H9 & I9 -> H9=3; I9=4.
5 between H8 & H9 -> H6=4 and H8<H7.
D5 <> 3. Also D5 <> 1 -> D5=2 -> C5=1; B5=4 and A5=3.
5 between G8 & G9 -> G10 <>4.
4 between H8 & I9 -> G8 <>4 -> G7=4 -> F8=4 and H8<G8.
6 between C1 & L10 -> C2 <>4...
C7=3;C8=2 (inequality not filled in before...) -> D6=3 (-> A6=1;B6 =2.) and D2=3;D3=2;D4=1.
Grid 2 can be completed now.
4 between A8 & E12 -> F12 <>4. (struggling...)
H5 must be 1 or 2 -> H7=3.
I'm wondering at this point if there's another rule missing - like "same sized buildings can't share an edge"? [definitelyisn't the case...]
Oh, Grid 3 Col F has at least 2 (viewed from top) as does grid 1 row 3, so both must be 2 -> F5=3; A3 <>4. Also, C3<>3 -> B3=3;C3=4;A3=1;C4=3
In the same way, the 6 between C1 & L10 -> both have 3 visible -> C2=2 (grid 1 can be completed) and J10=4 (with L10<K10) -> K12=4.
5 between A5 & L16 -> L16 <>3. 4 between A4 and L16 -> K16<L16 -> I16=3 and K16=1;L16=2.
6 between H11 & I11: no more than 3 visible in 4 -> no less than 3 visible in grid 5 -> J11 or K11 is the 3. Since J9 or K9 is also a 3, and K10>L10: I10=3; L12=3; K10=2; L10=1; L9=2. So G5=2;H5=1 -> H8=2 -> G8=3 and grid 3 can be comleted.
(3 grids to finish off - struggling again...)
In Grid 4: there are only 2 possible arrangements for the remaining 4s.
If G11 is a 4: F10=4; H12=4. F11=3; H11=1; G12=3;H10=2. But now we can't satisfy the 5 between G8 & G9. So G11<>4.
G12=4; H10=4; F11=4. Grid 4 can now be completed.
In Grid 6: I14<J14.
If I14=1: J14=2;L14=3;L15=1->J13=1;J15=3. The 5 between J12 & J13 cannot be satisfied.
So, I14=2 and grid 6 can be completed.
J12>J11 -> J12=2; J11=1 and last grid can be completed.

I think this proves there is a unique solution.
Calavera - I hope you don't mind us publishing this...

[Calavera]

It's not my puzzle. I'm the guy with the slitherlinks . Don't know how the puzzle author thinks about this. My personal opinion is, that everyone reading the spoiler before at least trying to solve it for himself simply misses the fun. So no harm done except some too curious people might lose the opportunity to enjoy thinking for themselves .

Nevertheless the comment function of the portal would be even better for this in my eyes. Especially for the more complex ones or some puzzles which require guessing of the rules it is used quite frequently and also in English. For example one of the top solvers is from South America who doesn't speak German and communicates through web-translation from Spanish to English as his English isn't that perfect either . So if you find another puzzle you're interested in and have problems understanding the rules just use the comments and I'm pretty sure you'll get an answer.

[david mcneill]

Thanks Calavera and Alan for responding to my query. Alan's walkthrough shows me what I had overlooked and I am happy that there is a unique solution.

David.

[nickdeller]

I see that in amongst all the Fehlversuche today, one solver has 1 Kühlschrank to their name. As far as I can ascertain, that translates as "refrigerator". Dare I ask what one needs to do to end up with a Kühlschrank?

[Calavera]

You have to be a friend of berni . She told him that her refrigerator got delivered in the middle of solving the puzzle...

[sf2l]

Hi Stefano!

The magic combination is pressing "F" (shift is not required - I just made it capital so that it's easier to read) while being on a cell with definit symbols that is adjacent to a one with double symbols. Also the double symbol can be entered by pressing "." and +, - and N result in plus, minus and black cell (half)s. There are one or two other shortcuts in the applets all of which can be found (in German unfortunately) under "Anleitung". Another really handy one is "#" in any puzzle with cells in which you have to enter letters or digits. This switches on and off a candidate list for this cell.

Calavera

Thank you very much -
Stefano

[PuzzleScot]

I've discovered why my time was so slow today - I went round filling in every border (like you'd need to do in a WPC contest), but replaying it, I discovered you only need to fill in the digits. Is there a general rule on what's required for a 'complete' puzzle?

[Calavera]

No general rule I guess... As the applets are currently updated I'm not sure how long this information is correct but some puzzles where it might be interesting are:

Fillomino: When finished the applet automatically adds borders between different numbers and then fills empty fields with numbers equal to the size of there area.

Laser: Empty fields don't have to be marked and mirrors don't have to be drawn. But as far as I remember wrong mirrors produce a Fehlversuch.

Magnetplatten: Also empty fields must be marked.

Zahlenspirale, Buchstabensalat, Hitori, Heyawake, Rundweg, Tapa, Masyu, Schlange, Sternenhimmel, Doppelstern, Pillen, U-Bahn: Empty/Unused fields/borders don't have to be marked.

But as I said new applets should be there quite soon and might have a different behaviour (borders between different Fillomino-numbers might be drawn in realtime etc.).


Calavera

[nickdeller]

I'm amazed and awed by some of the times posted today - I'm pretty sure that even if I were actually looking at the answer grid, I wouldn't be able to reproduce it in the puzzle grid in under 20 seconds! Quite a contrast from yesterday...!

[ronaldx]

I think I would have taken about 10 times as long if I hadn't seen the high score table before I started. Feels a bit like cheating !

It's the 5:16 time that impresses me - to get a Fehl, correct it, and still finish with a top 10 time (but for the penalty)... that's good work!