Sudoku

Cylindrical Sudoku 1

I’ve been threatening to do this for some time, and I finally got around to it. Here’s how it works.

This isn’t exactly a cylinder; it’s a regular nonagonal prism. That means the cross section is a regular nonagon, a nine sided polygon with equal sides and equal angles. The above picture shows 5 of the 9 columns. The other 4 are on the other side of the prism, hidden from view.

“How can I possibly solve the puzzle when I can’t see four of the columns?”

Here are the other four columns, the other side of the prism.

“Why is the prism colored red, green and blue?”

If the prism was colored white and gray like a normal Sudoku grid, then there would be two gray 3 X 3 boxes touching each other, or two white 3 X 3 boxes touching each other, and that wouldn’t look pretty.

“Why is the back different from the front?”

Because nine is an odd number, and if I’m going to show all the columns, one side has to have 5 columns, and the other side has to have four.

“Why are some of the columns horizontally squished?”

The only column that you see straight on (perpendicular to your line of sight) is the center column of the front view. All the other columns appear to be squished because they are slanted away from you, some of them a little bit, some of them a lot.

Other than that, it’s a normal Sudoku puzzle. Fill in the grid with the numbers 1 through nine so that there is no duplication in a row, column or 3 X 3 box. You can tell what’s coming next, can’t you? Yes, cylindrical Sudoku in Chinese!

Technorati: Sudoku

A Cypriot Syllabary Sudoku

The Cypriot syllabary is a syllabic script used in Iron Age Cyprus, from ca. the 11th to the 4th centuries BCE, when it was replaced by the Greek alphabet. But it seems to have enough interest to have its own Unicode chart

Here is the solution to the Coptic Sudoku of Nobember 25

Rethinking Timing TBLR Sudoku

Somehow, in the course of seeing if I could solve Sudoku puzzles top-to-bottom, left-to-right (TBLR) in 30 minutes or less, I stopped having fun. So instead of seeing what percentage of the time I could solve a puzzle TBLR in under 30 minutes, I’ve decided to just do the puzzles and see how much time it takes, and see if I can improve my time. The not being under time pressure seems to help. I will keep doing one post a week, unless, of course, I stop having fun doing it this way.

Thursday, 12/10/09: 4-star, 45:00
Friday, 12/11/09: 5-star, 39:58
Saturday, 12/12/09: 6-star. I got tired around 40 minutes and I made a choice without following the consequences far enough. Fail.

A Somewhat Coptic Sudoku

When I created this puzzle, I asked for two random numbers between 1 and 9, and received 9 and 8, again not the random numbers I wanted.

The last time I did this, I never got around to actually solving the puzzles I created. I have done so this time. Here is the solution to the Slightly Armenian Sudoku from November 12. The little numbers in the top right corner of some of the squares signify the order in which I filled up the grid.

On further review, the character I used came from the Armenian ligature Unicode set. In this context, ligatures are characters which stand for multiple characters. I was confused by the Unicode website. I got the impression that the Armenian ligatures set was the main Armenian character set, but it isn’t. I shall have to make another Armenian Sudoku eventually. I didn’t know what to call this Armenian ligature when I was solving the puzzle. But it reminded me of a Hebrew letter, so I called it “Gimel.” I later discovered that there isn’t a Hebrew letter that looks like this symbol. Since there were no 2’s given in this puzzle, I could have decided that Gimel stood for 2 instead of 1. But when I was composing the puzzle, and I asked for a random number between 1 and 9, I received the number 1. So Gimel stands for 1. Got that?

One thing I notice is that to solve these puzzles with unusual symbols, I have to make the unusual symbols part of an ordering of all the symbols. In this case, my ordering was “Gimel-2-3-4-5-6-7-8-9.” Will this Coptic puzzle be significantly harder than the Armenian one? We’ll find out soon enough!

Timing TBLR Sudoku, 11/17/09-11/23/09

Today was a really bad day for TBLR Sudoku. I had the incredibly fast time from last Tuesday staring over my shoulder, and I ended up imagining the incorrect number in a square, which lead to the inevitable TBLR fail. I got another copy of the Variety section that was laying around, and started doing the puzzle the regular way. I imagined something that wasn’t there then too, and I failed on a 2-star, which is pretty rare for me. I copied the puzzle onto a piece of paper and tried again, finally succeeding. Sounds like Swamp Castle (“Ah, but the fourth castle, it stayed up!”). I am therefore 1 out of 2 in solving 2-star puzzles TBLR in 30 minutes.

Wednesday, 11/18/09: 3-star. I did the puzzle, but didn’t record what I’ve usually been recording. I therefore count myself as 0 for 2 in solving 3-stars (by “solving” I mean “solving TBLR in 30 minutes”)
Thursday, 11/19/09: 4-star, 56 open squares, 0 filled, 31 identified in 30 minutes. That’s a new personal best on the number of squares identified in a 4-star. I filled in the squares I identified, and took an additional 3:03 to finish the puzzle. 0 for 2 in solving 4-stars.
Friday, 11/20/09: 5-star. I didn’t count the squares because I failed in both the TBLR solving and the solving of any kind. 0 for 3 in solving 5-stars
Saturday, 11/21/09: 6-star. I did the puzzle, but didn’t try to do it TBLR.
Sunday, 11/22/09: Didn’t do the puzzle.
Monday, 11/23/09: 1-star, 56 open squares, 56 filled in 18:13. New personal best, making up for the lousiness of earlier in the week. 1 for 2 in solving 1-stars.

A Slightly Armenian Sudoku

I realized when I did the last Armenian Sudoku, I left out a symbol, so I had to try again. In the process of trying again, I had a thought that it might be a good idea to gradually introduce the alternate symbols, and that it might be good to select which symbols get replaced by alternates at random. I did that today. I asked for a random number between 1 and 9, and got 1, which was exactly the random number I didn’t want, but I’m going to go with it anyway. So this puzzle has 1 Armenian character in it.

I would very much like to know your reactions to this puzzle.

Timing TBLR Sudoku

What is TBLR Sudoku, you ask? Top-to-bottom, left-to-right Sudoku, of course! It’s a Sudoku variation where the empty squares have to be filled in starting at the top row, going left to right, and continuing in that order until you get to the rightmost empty square on the bottom row.

I find it an amusing variation, because I have to recall numbers I’ve found in order to fill in the numbers in the correct order. The downside is that sometimes it takes a while, especially with the harder puzzles. Therefore I have created a compromise. I’ll spend 30 minutes trying to fill in the grid TBLR. If I don’t succeed in 30 minutes, I’ll fill in the empty squares I’m remembering, and try for another 30 minutes.

It would be really stuuupid to make a separate post for each day, but it might not be so bad to make one post per week, with the times and results for each week. Today was a particularly good day for me, as I solved a 2-star puzzle TBLR in 22 minutes, 48 seconds. That may be a personal best for me, but I haven’t done a very good job of keeping statistics. It is my plan in the coming days to keep track of my statistics in this space, and hopefully improve my times for each level of puzzle difficulty. I’ve done a 4-star puzzle TBLR in under 30 minutes, but I don’t have the exact time.

Tuesday, 11/10/09 : 2-star, 22:48
Wesnesday, 11/11/09: 3-star, 53 open squares, 5 filled in (all in the top row), 26 squares identified (that means I figured out what they were, but I hadn’t filled them in yet), 20 squares unidentified in 30 minutes. I filled in the 26 squares I identified, and took an additional 5:15 to fill in the rest.
Thursday, 11/12/09: 4-star, 53 open squares, 1 filled in, 28 squares identified, 24 unknown in 30 minutes. I filled in the 28 squares I identified, and took an additional 5:09 to fill in the rest.
Friday, 11/13/09: 5-star, 54 open squares, 0 filled in, a measly 8 squares identified, 44 unknown in 30 minutes. I filled in the 8 squares I identified, made a guess between two numbers as to what should go in a particular square, which turned out to be the wrong guess. Sad, sorry, miserable fail, but what should I expect on Friday the 13th?
Saturday, 11/14/09: 6-star, 52 open squares, 0 filled in, 19 identified. I filled in the 19 squares I identified, made a guess between two numbers, and it was the wrong guess. Fail. I’m 0 for 2 as far as guessing is concerned.
Sunday, 11/15/09: 5-star, 50 open squares, 0 filled in, 11 identified, 39 unknown in 30 minutes (An upstat of 3). I filled in the 11 squares I identified, and took an additional 28:47 to complete the puzzle (another upstat over my Friday fail)
Monday, 11/16/09: 1-star, 51 open squares, 45 filled in, 5 identified, 1 unknown in 30 minutes. I filled in the 5 squares I identified, and took and additional 0:04 to complete the puzzle. I had it figured out, and was racing to finish, but didn’t quite make it.

Since the last time I wrote about Sudoku, there’s been a couple of changes in my solving strategy. Now when I find that a number must go in a certain square, I look to see if that causes any other numbers to go in any other squares. Also, whenever I get six squares filled in either a row, a column or a 3 x 3 box, I check what’s left, as the chances are good that the remaining squares might already be determined.

I’ve noticed that after doing Sudoku puzzles TBLR, doing Sudoku puzzles the normal way is significantly easier.