The New Sudoku Players' Forum

[Pappocom]

I would dearly love to see an English-language version of Nikoli's quarterly Japanese-language puzzle magazine.

In fact, just doing a translation of it would be good enough for a start (with rights from Nikoli, of course). After a while, the English-speaking countries should acquire a fan-base of puzzle enthusiasts of their own who would keep the magazine supplied with new material.

The Nikoli magazine treats its readers as members of a Club. I can get only a broad idea of its flavor, because I don't read Japanese. But it's clear there's a respectful relationship between publisher and solvers. And certainly there's no dumbing down of the puzzles or the editorial content.

Is there some publisher out there who would be willing to start a new venture?

- Wayne

[lunababy_moonchild]

Puzzler Media publish Nikoli's puzzles in books and in magazines in this country here.

Not Nikoli's Magazine itself I'll grant you, but there nonetheless.

Luna

[motris]

I second Wayne's vision - I would love Nikoli's main magazine to be translated into English. I mean, some of the "secret" fun of Nikoli's "test" puzzle pages for me, someone who reads no Japanese, is to try to figure out what the rules of the puzzle are, but that most always requires looking at the solutions and back-solving what the rules were that got to that solution. But overall, the mix of logic puzzles in Nikoli is amazing and even with the language barrier I consistently get Nikoli and their annual Giants magazine. However, I feel I still completely miss the flavor of Nikoli's publications with my never possibly understanding their comics and little touches like that. When a magazine makes you want to learn a foreign language, that is a powerful sign of how much respect a puzzle solver has for it (not that other Japanese magazines aren't great - I've also experienced some excellent nanpure magazines with fun nanpure variants - ok, I'll call them sudoku variants even though those publications do not - that are much better than Nikoli's sudoku selection, but it is much harder to order them in the US).

I am not familiar with all of Puzzler's magazines, not having been in the UK in about 3 years and not having bought any of the magazines when I had visited, but the one Puzzler publication (I believe) that comes to the US using Nikoli puzzles is mainly regular sudoku each month with about 3-4 slitherlinks and 3-4 bridges (hashiwokakero) and maybe 1-2 killer sudoku or kakuro. A start, but not a publication I bother to buy as it pales to Nikoli's in terms of overall mix and evenness. I mean, I like seeing sudoku, but I also like seeing slitherlink and nurikabe and lights up and masyu and kakuro and heyawake and lineconnects and fillominos and ... (ok, maybe not the hitori - i've never found them intellectually satisfying but it may just be me).

Nikoli may eventually go to Wayne's dream scenario though. At least, the improvements over the last couple of years in the English order site for their books as well as the recent publication of Penpa Mix 2 with (brief) English instructions for all of the included puzzles gives me some hope that they are well aware of how their puzzles would sell in a captive, puzzle-starved western market. Or, more likely, they have such a respect for the beautiful puzzles they have made and for the solvers who enjoy them that they want to make it possible for us in the West to see their craftwork.

[garethm]

There is a new UK puzzle magazine with a wide range of 'Japanese' style puzzles in - issue 1 of Total Konzai from Future Publishing (a very large magazine publisher) is available now. It costs 2 pounds, or less by subscription.

The puzzles in it include Hanjie, Hashi, Hitori, Kakuro, Killer Sudoku, Masyu, Nurikabe, Slitherlink, Battleships and Sudoku, plus variants such as Killer Sudoku X, Sudoku X, Sudoku 6x6, Sudoku 16x16, Samurai Sudoku (2 grids) and so on.

I think it's a great mix, and there are roughly even numbers of each puzzle type. It was briefly discussed in this thread: http://forum.enjoysudoku.com/viewtopic.php?t=3439

If it does well then it will prove a general market for this sort of magazine in the UK, which might well encourage other similar titles. I should also declare an interest in that I supply the puzzles for it. In terms of difficulty it's aimed at a general audience, but even so some of the puzzles are pretty challenging.


Gareth

[lunababy_moonchild]

I do know that parts of Nikoli's website are in English, so they are capable of it.

Luna

[Pappocom]

When a magazine makes you want to learn a foreign language, that is a powerful sign of how much respect...

Touché! I have wanted to learn Japanese, for the same reason. In 2004, I nearly enrolled for one month in a Japanese language school, but decided to press on and finish my Sudoku software - which was probably a good decision! But I might get back to that idea in 2007.

Garethm, I noticed your eariler reference to Total Konzai, but (by the sound of it) it's not the same as the Nikoli quarterly magazine. Nikoli goes much further than presenting a static selection of puzzles. It's the interactive nature of the Nikoli magazine that appeals to me. There's an active and lively correspondence with solvers, about puzzles and future puzzles. There are solving leagues, and "Letters to the Editor", and regular columns on puzzle-related matters, book and puzzle reviews (including games and sometimes mechanical puzzles), a register of events, etc..

Of course, in a way, these forums are supplying a similar outlet.

I am on good terms with Nikoli. At least, I think I am - they may have a different idea! I call and see them when I am in Japan, provided I have anything useful to say at the time. I will be in Tokyo again at the end of May.

- Wayne

[garethm]

(by the sound of it) it's not the same as the Nikoli quarterly magazine. Nikoli goes much further than presenting a static selection of puzzles.

Yes, I can imagine that this makes a real difference, and you're right that Total Konzai's not like that (although I would personally prefer it if it was!). As just a small subset of what you describe above, I much prefer the puzzle magazines I buy to have plenty of editorial text added alongside the puzzles, for example to point out interesting features or comment on that issue's puzzles. Not only does it provide a nice counterpoint to actually solving the puzzles but it also makes it seem that the people who create it actually care about what they're doing, and you get the impression that they actually solve the puzzles in their own magazine! Certainly not all puzzle magazines give this impression and some just look like someone pushed 'print' on a computer program!

Taking that a step further to provide a real community feel in a magazine would be really great, although perhaps - and I might be wrong here! - this is something that would maybe work better in Japan than in many other countries, just because of the different mentality over how seriously it's taken by the majority. I think many UK magazines now struggle to receive anything in the way of feedback, even when they go out of their way to ask for it, perhaps because many people who would traditionally have written in to try and get their opinions across now just post them somewhere online. Also, apparently the vast majority of puzzle magazine sales (here at least) are to casual buyers, mainly at travel places such as stations, airports and so on. For most titles available in stores only a tiny percentage buy them regularly.

I would now go on to say how a good puzzle website with a wide mix of content and good community features could be a solution, but the fact is that no matter how good a web interface is I somehow always find that a nicely-printed magazine or book is a much more tactile and rewarding experience than tapping away at a web interface - and in any case most of them are frustrating when compared to the flexibility of an infinite variety of physical pencil marks! Plus, of course, less portable and harder to put down and pick up. (Although that said, some puzzles are more pleasant to solve online - for example I find Nikoli's Hitori interface much better than pencil and paper)

Another point is that creating lots of puzzle variations is a lot more work when your website works by having an interactive player, which somewhat discourages variety and creativity - you need to write new code for each version, whereas in a printed magazine you just describe it and print it! ;) Of course you can provide PDFs or similar to print out, but then you're basically back to a paper magazine with a web interface for sending in feedback... Although that does suggest an alternative of a web-only puzzle magazine that you could subscribe to and print yourself.

Anyway, after reading your posting last night I ordered a stack of stuff from Nikoli, so I'm looking forward to seeing what they do for myself! ;)


Gareth

[RW]

Earlier on this thread there was a discussion on how to have a fair playoff.

My ideal playoff in this kind of format would be a sudoku relay. Imagine 5-7 medium sudoku all set up so that solving the first one reveals the given numbers for the second and so on forward.

A variation of this has been used in the Nordic Championships last December. The national champions from Sweden, Finland, Norway and Denmark where invited to the finals. They where presented with 8 puzzles and had to complete as many as possible within 1 hour. All normal sudokus, except that two of them contained letters instead of numbers. The puzzles can be found here: 1-4, 5-8. The puzzles can be solved in any order. Does this seem like a fair model to you?

Under no pressure in the comfort of my home, these puzzles didn't seem so hard, I completed 5 in 1 hour. However, I know that the pressure in a competition does affect at least my solving. Didn't make it to this Nordic final, because in the Finnish finals I couldn't complete any of ten given puzzles within the 1 hour limit...

RW

[motris]

The WPC (which seems to always come up with a slightly different playoff system each year) has done something like this already and I think it works alright. Players get some number of puzzles and a total time of something usually about 40 minutes to an hour to solve them in. The person who gets farthest in the time limit wins, although typically the time is set so that some people finish and it is the best time that matters. The additional wrinkle the WPC uses is that, as the finishers have scores from the qualifying round, the start times are staggered based on the results of the day before. There may be an X minute staggered start on the playoff round, with player 1 starting first, the last qualifying player starting after X minutes, and everyone imbetween getting a time difference set linearly based on the poisition of their score from the two scores that define 0 and X minutes). This system could work well at the WSC. It certainly is much better than using 7 discrete rounds with no results carrying forward from the day before or from puzzle to puzzle during the playoff itself.


Another question that is worth discussing, particularly in any sudoku competition like this, is how to score partially completed solutions. While this is not possible for many logic puzzles and is never/rarely done at a WPC, it is very possible with sudoku. During the main day's competion, points were only earned for complete puzzles (either 0 or X points, X the value of the puzzle). During the playoff, points were rewarded for correct numbers minus incorrect numbers with the givens being added to your score only if you got them all right. Neither of these is fully satisfying.

With the former, I, for example, used a bad test taking strategy and solved all the irregular/jigsaw puzzles in the irregular round before tackling the 12x12, but had only 11 cells (1 more minute of work) left on it when time ran out. This earned me 0 of the 90 total points for this biggest puzzle in Lucca and allowed other competitors to catch up, but not pass me, in the overall standings. As another example, oftentimes you would see solutions where, at the end of a puzzle, the competitor had quickly miswritten 2 numbers in the grid to lead to clear contradictions but these 2 squares were the only things wrong (fortunately, I never did this in the first day and all my completed rounds were indeed clean for full credit). To me, having an obvious error in 2 cells you would instantly see if you had the time to check is much different than a competitor who spent no time on a puzzle and had it blank. These both were scored 0.

The latter system seems too punitive for wrong numbers. The point of subtracting mistakes is, perhaps, to keep people from guessing numbers. As this may happen anyway (people would likely mark their guesses in such a way so they can erase them if they prove wrong) I think this is not necessary. Removing "guessing" from the equation, there is a big difference between getting 40 numbers into a puzzle, then making a mistake, and solving forward to an end where you reach a contradiction but likely have 20 or more errors, and someone who gets 25 numbers in and is still stuck with the rest blank when time runs out. 40-20 would lose to 25-0 in the scoring system used at Lucca. In an extreme case, imagine I got the first 6 numbers in the toroidal playoff puzzle you can find on another thread here correct, and then errored moving forwards at that point exactly. I would have gotten a negative score on the puzzle and been eliminated just for getting past the sticking point after number 1 and continuing to try to solve the puzzle past where everyone else stalled.


My scoring system would be something like this - to reward completed puzzles, all puzzles would be graded out of 100 points/percent and solving a puzzle would get you credit for the given numbers in the puzzle (typically around 20) and an extra 19 points to account for the difference between 9x9 and 100. Solving puzzles would therefore account for about 40% of the total score for a puzzle, which seems about right. Those with incomplete/incorrect solutions get simply the number of correctly entered squares as their percentage score on a puzzle. If this is 50 squares, then they get a score of 50 out of 100, if this is 20 squares, they get 20 percent. If they haven't attempted the puzzle, it is 0. It seems there should be some accounting for correct work towards a solution, but a much bigger bonus for actually completing the puzzle clean without any mistakes, and that this scoring should be used in all rounds, not just the playoff.

Anyway, these are all things worth consideration and I'll be interested to see what the organizers of the next WSC put together for their playoff. I can understand why individual newspaper competitions may want single puzzle playoffs for both time and drama. The World Championships should be set up to be a better test of overall solving skill than that and find a good way to test skill over multiple puzzles - as the first day with 40+ sudoku puzzles in multiple rounds did to get the playoff qualifiers - to crown the world champion.

Thomas Snyder
motris.livejournal.com

[RW]

Another question that is worth discussing, particularly in any sudoku competition like this, is how to score partially completed solutions.

That is indeed a good question. The model you suggest sounds like a reasonably good one, but it also provides a very good loophole for the guessers: When there is only one minute left to solve, just guess as much as possible - there's nothing to lose. Then again, your example on what could have happened in the torroidal shows that penalizing wrong numbers isn't a good solution either. In an ideal world it would be possible to arrange a competition where no guessing was allowed, only logical solutions. But as you said:

The impossible to implement way to judge the top solver is to make the puzzle solution be presented more as a logical proof than as just a filled grid.

A variant on this could be that the solver on request must be able to account for all his moves afterwards. You would need supervisors that could follow your solving with programs that automatically reduce all candidates that can be reduced with basic techniques (pairs, triplets, quads, locked candidates, x-wing, swordfish, uniqueness, etc.) and display both naked and hidden singles as singles. If the competitor enters a number into a cell that doesn't hold a single on their screen, they could suspect a guess and ask him/her after the solving is done what that move was based on. If the solver found a forcing chain, used nishio or whatever strange logical technique, then he can explain the logic behind his move. If it was a guess, he can't.

There is however a downside to this too: When I solve, I use techniques like that quite often and almost as often I cannot exactly remember what I did when I've finnished the puzzle. Once I solved a puzzle in 15 minutes and when I immediately tried to recreate my solution it took me 45 minutes to find what I had actually done... So in a competition like I just suggested, I would probably have been disqualified.

I guess it is quite impossible to come up with a solution that is fair towards everybody, but I think this is a very important topic to discuss in order to improve future competitions.

RW

[tso]

I received this reply to an email to the WSC regarding the puzzles in the recent Sudoku championship:

"The WSC organization will publish an official book with all the
WSC's sudoku.
The Book may be avaible online at the end of April.
If you are interesting you can reply at this email and I'll send an
alert email when book will be avaible at our shop.

Best regards

WSC Organization"

http://www.wsc2006.com/eng/inquires.php

wsc@nonzero.it


=============

I too find that solving puzzles in Nikoli magazines adds two layers to the solving.

ONE is infering exactly what the rules are. This can be especially difficult when there are only two or three examples of the puzzle. If you like this, try Puzzle Cyclopedia which has hundreds of different types of puzzles all previously published in the experimental section of the quarterly Nikoli magazine. Unrelated to Nikoli is The Puzzle Laboratory, which also has several hundred logic puzzles (and growing), explanations only in Japanese.

TWO is finding the tactics for solving the puzzles. Admit it -- Sudoku was much more fun the first time when you didn't know what you were doing. "Solving" a puzzle to me is figuring out how solve them in general. Finding tactics is more satisfying that applying previously learned tactics. I don't want to see any hints on how-to at first -- and if there are any hints in Nikoli, I woudn't know.

[Pat]

I received this reply to an email to the WSC regarding the puzzles in the recent Sudoku championship:

The WSC organization will publish an official book with all the WSC's sudoku.
The Book may be avaible online at the end of April --

Will Shortz Presents The First World Sudoku Championship: 100 Wordless Crossword Puzzles (Paperback)

Publisher: St. Martin's Griffin (October 3, 2006)
ISBN-10: 0312363702
ISBN-13: 978-0312363703

[Pat]

World Sudoku Championship 2

2007.Mar.28-Apr.1

Prague, Czech Republic

[Pat]

Thomas Snyder -- the winner

[udosuk]

Congrats Thomas (motris)...

Also well done to Japan for winning the team competition. Beating the almighty team U-S-of-A in any international competition is always an impressive achievement!

For some reason Australia decided not to participate... I'm sure we have enough talent to earn a place in the top 8 at least...