 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
| HomeDiscussion GroupsEnglish UsageBritish EnglishESL Teaching Learnglish.com Contact UsLink To UsSearch & Site Map |
Discussion Groups / English Usage / January 2010Sodoku
Hello All! In view of the current popularity of Sodoku puzzles, it is interesting to see that the OED entry has not caught up. sodoku Path. [Jap.] The form of rat-bite fever caused by Spirillum minus. I admit I was surfing the dictionary after I found I had been wrong about "somewhen".  SignatureJames Silverton Potomac, Maryland Email, with obvious alterations: not.jim.silverton.at.verizon.not James Silverton: > In view of the current popularity of Sodoku puzzles, it is interesting > to see that the OED entry has not caught up. [quoted text clipped - 3 lines] > [Jap.] > The form of rat-bite fever caused by Spirillum minus. Well, then. > I admit I was surfing the dictionary after I found I had been wrong > about "somewhen". Did you surf as far as "sudoku", to see if they had an entry for the puzzles? SignatureMark Brader "I wasn't the one who misplaced the entire Toronto Deltivid asteroid belt!" msb@vex.net "Deja Q", ST:TNG, Richard Danus > Hello All! > [quoted text clipped - 10 lines] > I admit I was surfing the dictionary after I found I had been wrong > about "somewhen". The OED has cunningly hidden it under the correct spelling, "sudoku". SignatureJames James wrote on Sat, 23 Jan 2010 22:01:51 +0100: >> Hello All! >> [quoted text clipped - 13 lines] > The OED has cunningly hidden it under the correct spelling, > "sudoku". In hiding, and thanks for the correction. I tried a few sudoku puzzles but have not looked at them in a while since I found the process of trial and erasure to be rather a bore. I suppose a semi-automatic computer method showing possibilities could be set up and I did begin to program that before losing interest. SignatureJames Silverton Potomac, Maryland Email, with obvious alterations: not.jim.silverton.at.verizon.not > In hiding, and thanks for the correction. I tried a few sudoku puzzles > but have not looked at them in a while since I found the process of > trial and erasure to be rather a bore. I suppose a semi-automatic > computer method showing possibilities could be set up and I did begin to > program that before losing interest. You can find such a system on the websites of more than one newspaper or similar service, and of course freeware versions and shareware versions abound. I decided a year or two ago, rather belatedly, that I was going to figure out how to do sudoku puzzles. I can now do them, but haven't yet acheived the ability to do any but the simplest with paper and pencil. The erasing gets on my nerves, as does the scribbling of guesses followed by the attempts to read them. I might make another of my rare attempts on cryptic crosswords someday. Regular crosswords are fine, but those drive me crazy. SignatureCheryl >> In hiding, and thanks for the correction. I tried a few sudoku >> puzzles but have not looked at them in a while since I found the [quoted text clipped - 11 lines] > pencil. The erasing gets on my nerves, as does the scribbling of > guesses followed by the attempts to read them. Any sudoku puzzle that you're likely to find published will not require guessing. What it may require is keeping track of the remaining possibilities for a cell and erasing them or crossing them out as they become eliminated. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |It is error alone which needs the 1501 Page Mill Road, 1U, MS 1141 |support of government. Truth can Palo Alto, CA 94304 |stand by itself. | Thomas Jefferson kirshenbaum@hpl.hp.com (650)857-7572 http://www.kirshenbaum.net/ >>> In hiding, and thanks for the correction. I tried a few sudoku >>> puzzles but have not looked at them in a while since I found the [quoted text clipped - 15 lines] > remaining possibilities for a cell and erasing them or crossing them > out as they become eliminated. If I get sufficiently frustrated with it - especially with hand-scribbled notes on what might be where that I can't quite read and don't quite trust, I start guessing. I know I'm not supposed to. There are people who don't even write down their guesses - or rather, the possibilities they've identified - when they do the puzzles. That's astonishing. SignatureCheryl >>>> In hiding, and thanks for the correction. I tried a few sudoku >>>> puzzles but have not looked at them in a while since I found the [quoted text clipped - 23 lines] > the possibilities they've identified - when they do the puzzles. That's > astonishing. I can get quite a long way before needing to write anything down. I play Challenge Sudoku on Facebook: playing against the clock against very fast players is a real challenge. I have yet to complete a puzzle at "harder" level in 5 minutes, which is the top level you can achieve, but I keep trying. SignatureLaura (emulate St. George for email) >>>>> In hiding, and thanks for the correction. I tried a few sudoku >>>>> puzzles but have not looked at them in a while since I found the [quoted text clipped - 29 lines] > at "harder" level in 5 minutes, which is the top level you can > achieve, but I keep trying. I was given a book of the "Times"'s "difficult" ones for Christmas. The only writing down I do is when a row or column is down to two or three choices I sometimes write them at the end of the row - just to save me having to work them out each time I come round to check it. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk >>> There are people who don't even write down their guesses - or >>> rather, the possibilities they've identified - when they do the [quoted text clipped - 4 lines] >choices I sometimes write them at the end of the row - just to save me >having to work them out each time I come round to check it. The Times classifies the harder puzzles as difficult, fiendish or super-fiendish. For a long time super-fiendish meant you had to use a particular technique, but they've changed, and I think dumbed-down, all three classifications. I now reckon to do fiendish ones and occasionally super-fiendish ones without writing anything down (except the answers; I have not yet progressed to doing the whole thing without a pen, as some are said to do the crossword). I like the ones where you have to make the contents of boxes add up to the right thing better, though. This is partly because I haven't yet worked out how to write an efficient computer program to do those, though no doubt somebody somewhere has. I haven't programmed the standard one either, but that's just idleness; I can see how it could be done. I'm not sure I want to do them against the clock, though - they are meant to be a stress-buster not a stress-inducer. Katy ke10@cam.ac.uk: >The Times classifies the harder puzzles as difficult, fiendish or >super-fiendish. For a long time super-fiendish meant you had to use a [quoted text clipped - 13 lines] >meant to be >a stress-buster not a stress-inducer. Same here: I prefer the "killer" sudoku that also involves arithmetic, and I prefer not to write anything down other than the answer. Doing it against the clock seems to miss the point. I'm often baffled by classifications, finding the actual puzzle much easier or harder than suggested. The puzzles at killersudokuonline.com are classified as "easier", "easy", "moderate", "hard", "extreme", "outrageous", and "mind bending". Some "easy" are harder for me than some "mind bending". My mind must work differently. Regarding crosswords I do them without a pen only if there isn't one available. SignatureMike Barnes Cheshire, England >>>> There are people who don't even write down their guesses - or >>>> rather, the possibilities they've identified - when they do the [quoted text clipped - 9 lines] > particular technique, but they've changed, and I think dumbed-down, > all three classifications. Ah, I'm glad I'm not alone in thinking that. I now reckon to do fiendish ones and occasionally > super-fiendish ones without writing anything down (except the answers; > I have not yet progressed to doing the whole thing without a pen, > as some are said to do the crossword). > > I like the ones where you have to make the contents of boxes add up to the > right thing better, though. The killer ones? I like those too and have observed that they seem to appeal more to mathematicians. This is partly because I haven't yet worked out > how to write an efficient computer program to do those, though no doubt > somebody somewhere has. I haven't programmed the standard one either, but > that's just idleness; I can see how it could be done. > > I'm not sure I want to do them against the clock, though - they are meant to be > a stress-buster not a stress-inducer. I like trying to beat the regular solvers on Facebook, some of whom are incredibly fast. It hadn't occurred to me before now but I suppose they may cheat, as people seem to do on Scramble, althpugh I don't see the point. SignatureLaura (emulate St. George for email) > I like trying to beat the regular solvers on Facebook, some of whom are > incredibly fast. It hadn't occurred to me before now but I suppose they > may cheat, as people seem to do on Scramble, althpugh I don't see the point. A bit like lying in bed listening to the evening repeat of Brine of Britain or Round Britain Quiz. Impresses the wofe. > I now reckon to do fiendish ones and occasionally > super-fiendish ones without writing anything down (except the answers; > I have not yet progressed to doing the whole thing without a pen, > as some are said to do the crossword). That, to me, is the difference between sudoku and crosswords. I always do crosswords with a pen, and sudoku with a pencil. SignaturePeter Moylan, Newcastle, NSW, Australia. http://www.pmoylan.org For an e-mail address, see my web page. >> I now reckon to do fiendish ones and occasionally >> super-fiendish ones without writing anything down (except the answers; [quoted text clipped - 3 lines] >That, to me, is the difference between sudoku and crosswords. I always >do crosswords with a pen, and sudoku with a pencil. But it's also the difference between different types of crossword. With a cryptic crossword there's usually no possibility of doubt: when yuo've got it, you've got it. That's not always true of some of the US-style crosswords which are purely a matter of finding a word from a definition. You can only eliminate your incorrect alternatives by reference to other answers, which seems to me more like sudoku. SignatureKaty Jennison spamtrap: remove the first two letters after the @ >>> I now reckon to do fiendish ones and occasionally >>> super-fiendish ones without writing anything down (except the answers; [quoted text clipped - 10 lines] >definition. You can only eliminate your incorrect alternatives by >reference to other answers, which seems to me more like sudoku. Yes. Sometimes non-cryptic so-called "quick crosswords" can be more difficult than cryptic ones. I always use a pen for puzzles, but one with erasable ink. SignaturePeter Duncanson, UK (in alt.usage.english) LFS <laura@DRAGONspira.fsbusiness.co.uk>: >>>>> In hiding, and thanks for the correction. I tried a few sudoku >>>>> puzzles but have not looked at them in a while since I found the [quoted text clipped - 24 lines] > >I can get quite a long way before needing to write anything down. I only ever write anything down (other than the answers) as an absolute last resort - perhaps once every couple of months. At first that meant that the harder puzzles were impossible for me, but that was no great loss, and I soon got better. I actually prefer killer sudoku to ordinary sudoku, because it involves arithmetic as well as logic. SignatureMike Barnes Cheshire, England >>> In hiding, and thanks for the correction. I tried a few sudoku >>> puzzles but have not looked at them in a while since I found the [quoted text clipped - 11 lines] >> pencil. The erasing gets on my nerves, as does the scribbling of >> guesses followed by the attempts to read them. I find them to be a total waste of time. Write a poem, solve an equation, read a book, paint a picture -- almost anything is better exercise for the mind and, IMO, more enjoyable. >Any sudoku puzzle that you're likely to find published will not >require guessing. What it may require is keeping track of the >remaining possibilities for a cell and erasing them or crossing them >out as they become eliminated. Which is a fancy way of saying "scribbling guesses", as Cheryl said. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE Chuck Riggs skrev: > I find them to be a total waste of time. Write a poem, solve an > equation, A sudoku *is* an equation with 81 variables. SignatureBertel http://bertel.lundhansen.dk/ FIDUSO: http://fiduso.dk/ > Chuck Riggs skrev: > >> I find them to be a total waste of time. Write a poem, solve an >> equation, > > A sudoku *is* an equation with 81 variables. There are, at most, 64 degrees of freedom in a blank sudoku grid, because each row and column total is fixed at 45, and each digit from 1 to 9 has to be present in each row and column. So, when 8 digits have been found or are known in a specific row or specific column, the final digit in that same row or column is not free to vary. This reduces the degrees of freedom, and hence the number of equations needed to solve for the entire grid. Consequently, although your main point has some weight, it is 64 equations, not 81. As a much more fine point (some may wish to label it as being pedantic), I would rather state it as "A sudoku can be viewed as being an equation with 64 variables", because, often, part of the solving issue with problems is finding an appropriate abstraction and representation in order to solve it. If you don't like logic puzzles (of which sudoku can be thought of as one example), but do like solving equations, the higher problem of getting enough interest in them to try to solve them could be thought of as merely finding an appropriate form of representation to allow them then to fit in with your own preferences. I will now disappear through my own strange loop. SignatureZhang Dawei: Stoke-on-Trent, UK. Please use the Reply-To field for my email address, which is certain to remain valid for 2 weeks from the posting of this message. > > Chuck Riggs skrev: > [quoted text clipped - 10 lines] > reduces the degrees of freedom, and hence the number of equations > needed to solve for the entire grid. However, you can view that constraint as an equation. > Consequently, although your main point has some weight, it is 64 > equations, not 81. [quoted text clipped - 5 lines] > representation in order to solve it. If you don't like logic puzzles > (of which sudoku can be thought of as one example), I can't see what else someone would think of it as. > but do like > solving equations, the higher problem of getting enough interest in [quoted text clipped - 3 lines] > > I will now disappear through my own strange loop. Another iteration: A sudoku is some number of simultaneous equations in an equal or smaller number of unknowns, one for each blank cell. Though I'm having trouble seeing the constraints as equations. They're all non-equals statements. Maybe there's some ingenious way of turning it into a system of simultaneous equations, though. -- Jerry Friedman >> > Chuck Riggs skrev: >> [quoted text clipped - 12 lines] > > However, you can view that constraint as an equation. You can, but most people would not see it as one in need of "solving". >> Consequently, although your main point has some weight, it is 64 >> equations, not 81. [quoted text clipped - 7 lines] > > I can't see what else someone would think of it as. Neither can I, but I've been too closed-minded in the past, so I thought I'd be cautious at this point. >> but do like >> solving equations, the higher problem of getting enough interest in [quoted text clipped - 10 lines] > They're all non-equals statements. Maybe there's some ingenious way > of turning it into a system of simultaneous equations, though. "Simultaneous linear inequalities" might be a better term, do you think? SignatureZhang Dawei: Stoke-on-Trent, UK. Please use the Reply-To field for my email address, which is certain to remain valid for 2 weeks from the posting of this message. > Another iteration: A sudoku is some number of simultaneous equations > in an equal or smaller number of unknowns, one for each blank cell. > > Though I'm having trouble seeing the constraints as equations. > They're all non-equals statements. Maybe there's some ingenious way > of turning it into a system of simultaneous equations, though. There are 27 equations of the form "These nine variables add up to 45". There are also 810 inequalities of the form "This variable is not equal to that variable" and a general constraint that all variables are positive integers. And, of course, provided values for some variables. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |"Revolution" has many definitions. 1501 Page Mill Road, 1U, MS 1141 |From the looks of this, I'd say Palo Alto, CA 94304 |"going around in circles" comes |closest to applying... kirshenbaum@hpl.hp.com | Richard M. Hartman (650)857-7572 http://www.kirshenbaum.net/ >> Another iteration: A sudoku is some number of simultaneous equations >> in an equal or smaller number of unknowns, one for each blank cell. [quoted text clipped - 7 lines] > not equal to that variable" and a general constraint that all > variables are positive integers less than ten > . And, of course, provided values for > some variables. SignaturePeter Moylan, Newcastle, NSW, Australia. http://www.pmoylan.org For an e-mail address, see my web page. >>> Another iteration: A sudoku is some number of simultaneous >>> equations in an equal or smaller number of unknowns, one for each [quoted text clipped - 10 lines] > > less than ten If they are all positive, and each of the values in a (WLOG) row are different and have to add up to 45, you shouldn't need the additional constraint that they're less than ten. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |He who will not reason, is a bigot; 1501 Page Mill Road, 1U, MS 1141 |he who cannot is a fool; and he who Palo Alto, CA 94304 |dares not is a slave. | Sir William Drummond kirshenbaum@hpl.hp.com (650)857-7572 http://www.kirshenbaum.net/ >>>> Another iteration: A sudoku is some number of simultaneous >>>> equations in an equal or smaller number of unknowns, one for each [quoted text clipped - 12 lines] > different and have to add up to 45, you shouldn't need the additional > constraint that they're less than ten. Yes, but your formulation didn't contain the "all different" rule. That adds some extra inequalities. Oops. Now I see it. Those are your 810 extra rules. SignaturePeter Moylan, Newcastle, NSW, Australia. http://www.pmoylan.org For an e-mail address, see my web page. >Chuck Riggs skrev: > >> I find them to be a total waste of time. Write a poem, solve an >> equation, > >A sudoku *is* an equation with 81 variables. Only nine, I thought. Whatever, you learn nothing but patience by solving one, IMO. They are a waste of time and, at my age, I don't have time to waste. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE >>Any sudoku puzzle that you're likely to find published will not >>require guessing. What it may require is keeping track of the >>remaining possibilities for a cell and erasing them or crossing them >>out as they become eliminated. > > Which is a fancy way of saying "scribbling guesses", as Cheryl said. I disagree. A "guess" is to say "Let's say that this square held a '2'. Does that work?" Process of elimination is different. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |ActiveX is pretty harmless anyway. 1501 Page Mill Road, 1U, MS 1141 |It can't affect you unless you Palo Alto, CA 94304 |install Windows, and who would be |foolish enough to do that? kirshenbaum@hpl.hp.com | Peter Moylan (650)857-7572 http://www.kirshenbaum.net/ > >>Any sudoku puzzle that you're likely to find published will not > >>require guessing. What it may require is keeping track of the [quoted text clipped - 5 lines] > I disagree. A "guess" is to say "Let's say that this square held a > '2'. Does that work?" Process of elimination is different. I really dislike puzzles that reach an end position such that the only way to proceed is to make a guess, then follow all the eliminations until either the puzzle is solved or a contradiction is found. I discussed this with a constructor and he was aware that his method created puzzles of this sort. He even had a name for that solution method: "Ariadne's Thread". He took offense when I called this method "trial and error", and proceeded to give me a lecture on the meaning of "trial and error". (What he called "trial and error" was what I call "successive approximation".) SignatureJohn Varela Trade NEWlamps for OLDlamps for email >> I disagree. A "guess" is to say "Let's say that this square held a >> '2'. Does that work?" Process of elimination is different. [quoted text clipped - 8 lines] > the meaning of "trial and error". (What he called "trial and error" > was what I call "successive approximation".) I agree with you on "trial and error", but I wanted to note that in my son's elementary school math classes, they now call it "guess and check". I'd rather they had gone with "hypothesize and test", to bring out the notion that it's usually a good idea to have a reason behind your guess. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |First Law of Anthropology: 1501 Page Mill Road, 1U, MS 1141 | If they're doing something you Palo Alto, CA 94304 | don't understand, it's either an | isolated lunatic, a religious kirshenbaum@hpl.hp.com | ritual, or art. (650)857-7572 http://www.kirshenbaum.net/ John Varela skrev: > I really dislike puzzles that reach an end position such that the > only way to proceed is to make a guess, then follow all the > eliminations until either the puzzle is solved or a contradiction is > found. I discussed this with a constructor and he was aware that his > method created puzzles of this sort. That sort of sudokus belong to the difficult category. The program that I use, makes only the kind where logic can give you the answer, if you choose "medium" or "easy". > He even had a name for that > solution method: "Ariadne's Thread". He took offense when I called > this method "trial and error", and proceeded to give me a lecture on > the meaning of "trial and error". (What he called "trial and error" > was what I call "successive approximation".) I go for "trial and error" or for Evan Kirshenbaum's "hypothesize and test". I'd add as an argument that that is the way scientists work. SignatureBertel http://bertel.lundhansen.dk/ FIDUSO: http://fiduso.dk/ > That sort of sudokus belong to the difficult category. The > program that I use, makes only the kind where logic can give you > the answer, if you choose "medium" or "easy". I have a couple of books by Will Shortz, the puzzle editor of the New York Times, with sudoku puzzles rated up to "Beware! Very Challenging!", all of which can be solved with logic with no need to resort to trial and error. SignatureJohn Varela Trade NEWlamps for OLDlamps for email >John Varela skrev: > [quoted text clipped - 17 lines] >and test". I'd add as an argument that that is the way scientists >work. Not all scientists. Some of the greatest discoveries are the result of "Suck it and see" which is not quite the same, but closely related. (This methodology is frequently used when testing newly created software for the first time - or even by end-users of software who refuse to succumb to RTFM strategies.) Jitze > On Sun, 24 Jan 2010 23:49:08 +0100, Bertel Lund Hansen > [quoted text clipped - 27 lines] > software for the first time - or even by end-users of software who > refuse to succumb to RTFM strategies.) Hey, who was that? -- Jerry Friedman > > On Sun, 24 Jan 2010 23:49:08 +0100, Bertel Lund Hansen > > [quoted text clipped - 29 lines] > > Hey, who was that? Yes, who was that masked man? Should we have a welcome-home party? SignatureJohn Varela Trade NEWlamps for OLDlamps for email Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >>>Any sudoku puzzle that you're likely to find published will not >>>require guessing. What it may require is keeping track of the [quoted text clipped - 5 lines] >I disagree. A "guess" is to say "Let's say that this square held a >'2'. Does that work?" Process of elimination is different. I think you're seeking to make a distinction that lies elsewhere, if it exists at all. A "remaining possibility" is simply a guess that you haven't yet discarded, isn't it? I don't think it's possible for a person to solve sudoku without guesswork of some sort. Some guesses are so obviously wrong that most people don't even think of them as guesses. Some guesses are so hard to evaluate that most people don't have the mental capacity to do it without memory aids, or at all. But they're all guesses. To me the essential feature of a satisfactory sudoku puzzle is that the guesses occupy the middle ground. SignatureMike Barnes Cheshire, England > Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> [quoted text clipped - 20 lines] > the essential feature of a satisfactory sudoku puzzle is that the > guesses occupy the middle ground. I would consider a sudoku puzzle completely unsatisfactory if it couldn't be solved without writing something down that couldn't be justified as "Given the rest of the board, that absolutely has to be correct", where "that" could be "It's a 7", "It can't be a 3", "It has to be a 1, 5, or 8", or the like. I don't see any reason to call such things "guesses". Is there a distinction between "There are eight other numbers filled in in this row, so this cell has to be 7" and "There are six other numbers filled in in this row, so this cell has to be 1, 5, or 8, but the other two empty cells are in columns that have the 5, so this one must be the 5"? When does it become a "guess"? SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |Voting in the House of 1501 Page Mill Road, 1U, MS 1141 |Representatives is done by means of a Palo Alto, CA 94304 |little plastic card with a magnetic |strip on the back--like a VISA card, kirshenbaum@hpl.hp.com |but with no, that is, absolutely (650)857-7572 |*no*, spending limit. | P.J. O'Rourke http://www.kirshenbaum.net/ >> Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >>> [quoted text clipped - 29 lines] > have the 5, so this one must be the 5"? When does it become a > "guess"? Designers of the best-selling sudoku books eg Paul Stephens, Frank Longo are emphatic that guesswork is not needed for any of their puzzles. The chains of inference in their "extreme" or "absolutely nasty" puzzles can be very complex but they are all logical. SignatureLaura (emulate St. George for email) LFS <laura@DRAGONspira.fsbusiness.co.uk>: >Designers of the best-selling sudoku books eg Paul Stephens, Frank >Longo are emphatic that guesswork is not needed for any of their >puzzles. That might be true but it's also true that in practice people *use* guesswork a great deal of the time. Given the simplest sudoku puzzle in the world (80 numbers already filled in, one left for the user), how many of us would guess answers until finding one that fits, and how many would use a non-guessing method such as arithmetic? >The chains of inference in their "extreme" or "absolutely nasty" >puzzles can be very complex but they are all logical. "Logical" doesn't preclude guesswork, because logic is used to evaluate guesses. SignatureMike Barnes Cheshire, England Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >Is there a distinction between "There are eight >other numbers filled in in this row, so this cell has to be 7" and >"There are six other numbers filled in in this row, so this cell has >to be 1, 5, or 8, but the other two empty cells are in columns that >have the 5, so this one must be the 5"? When does it become a >"guess"? In the first case, I arrive at "7" by trying numbers until I find one that isn't a duplicate. That's guesswork[1][2][3]. In the second case I arrive at three numbers by the same method, then I use logic to locate one of them. [1] I am of course making the assumption that others solve this problem the same way. If anyone reading this would use a non-guessing method, such as "subtract the sum of the eight other numbers from 45" (which would work for the first example but not the second), I'd be interested to hear about it. [2] Relevantly to the original point ("scribbling guesses"), the first example is guesswork where writing things down actually wouldn't help. In the second example I might have written (when less practised than I am now) "1 5 8" in the margin, but those aren't guesses, they're certainties. [3] But after a year or two of doing sudoku I find that the process of evaluating each guess is *much* quicker than it used to be. I think somewhere in the eye/brain processing I've got better at finding numbers based on their overall shape, and eliminating them in the most efficient order. "1" is definitely quickest and therefore guessed first. "4" probably comes second, not much ahead of "7". I think straight strokes are easier to spot than curved ones. SignatureMike Barnes Cheshire, England > Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> Is there a distinction between "There are eight [quoted text clipped - 6 lines] > In the first case, I arrive at "7" by trying numbers until I find one > that isn't a duplicate. I can look at the set of numbers from 1 to 9 and identify the missing one almost instantly. I wouldn't describe that as guessing: that to me is a certainty - there is no 7. Guessing implies a slower process of trying something to see if it fits a gap - placing jigsaw puzzle pieces for example. That's guesswork[1][2][3]. In the second case I > arrive at three numbers by the same method, then I use logic to locate > one of them. [quoted text clipped - 18 lines] > probably comes second, not much ahead of "7". I think straight strokes > are easier to spot than curved ones. SignatureLaura (emulate St. George for email) LFS <laura@DRAGONspira.fsbusiness.co.uk>: >> Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >>> Is there a distinction between "There are eight [quoted text clipped - 9 lines] >I can look at the set of numbers from 1 to 9 and identify the missing >one almost instantly. Same here, but I recognise that what I'm doing is eliminating wrong guesses very quickly. >I wouldn't describe that as guessing: that to me is a certainty - there >is no 7. Once I've eliminated the wrong guesses, certainty is what I'm left with. >Guessing implies a slower process of trying something to see if it fits >a gap - placing jigsaw puzzle pieces for example. We'll have to disagree on that. To me, speed is not the essence of guesswork at all. Consider the computer program that guesses passwords millions of times a second. SignatureMike Barnes Cheshire, England Mon, 25 Jan 2010 11:28:15 +0000 from Mike Barnes <mikebarnes@bluebottle.com>: > In the first case, I arrive at "7" by trying numbers until I find one > that isn't a duplicate. That's guesswork[1][2][3]. In the second case I [quoted text clipped - 6 lines] > would work for the first example but not the second), I'd be interested > to hear about it. Bring this back to the English language, I don't agree with your meaning for guesswork. "Process of elimination", maybe, though even that doesn't seem to fit well. I ask myself, "which number is missing from that row? I don't guess, "Is it a 1? Is it a 2?" I know that 1 through 9 must be there, I see which numbers are there, and that tells me inexorably which number belongs in the empty square. When there's only one answer, and the available facts are enough to show what it is, I don't call that guesswork. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... Stan Brown <the_stan_brown@fastmail.fm>: >Mon, 25 Jan 2010 11:28:15 +0000 from Mike Barnes ><mikebarnes@bluebottle.com>: [quoted text clipped - 16 lines] >and that tells me inexorably which number belongs in the empty >square. I'm coming to the conclusion that my brain works a different way from other people's here. I do guess "Is it a 1?" (etc). It takes just a fraction of a second for the complete evaluation[1] and I'm generally not conscious of all the guesses, but I know full well that I'm working that way. I don't know of any quicker way of doing it. [1] 1 is much quicker than 9. I'm pretty sure that I try the digits in order of shape recognition speed. SignatureMike Barnes Cheshire, England Sun, 24 Jan 2010 23:45:01 -0800 from Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: > I would consider a sudoku puzzle completely unsatisfactory if it > couldn't be solved without writing something down that couldn't be [quoted text clipped - 7 lines] > have the 5, so this one must be the 5"? When does it become a > "guess"? Very well put! That's what I meant by a deterministic puzzle -- not routine, but one where there's only one correct answer and you can find it logically. I use the sort of logic you described. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... > Sun, 24 Jan 2010 23:45:01 -0800 from Evan Kirshenbaum > <kirshenbaum@hpl.hp.com>: [quoted text clipped - 13 lines] > routine, but one where there's only one correct answer and you can > find it logically. I use the sort of logic you described. In difficult puzzles I sometimes get to a point where logic doesn't help, or I don't see enough information to proceed. By that stage there is always some cell where the number of possible answers has been reduced to two. Then I switch from ballpoint to pencil and try one of the two possibilities. continuing until it's proved either right or wrong. Some people might call that guessing but I prefer to call it testing. Others call it a waste of time, and I admit there is something to be said for that judgement. SignatureJames Mon, 25 Jan 2010 15:51:56 +0100 from James Hogg <Jas.Hogg@gOUTmail.com>: > > Very well put! That's what I meant by a deterministic puzzle -- not > > routine, but one where there's only one correct answer and you can [quoted text clipped - 6 lines] > the two possibilities. continuing until it's proved either right or > wrong. Some people might call that guessing but I prefer to call it testing. While the number of possible answers in *that* cell, considered in isolation, may be two, I'll bet that when other cells are also considered there is only one possibility. I think what makes a more difficult sudoku more difficult is that, to come up with the one possible solution, one must consider a greater number of cells at a time. I agree that "testing" is a better word for what you do than "guessing". > Others call it a waste of time, and I admit there is something to > be said for that judgement. :-) There you have me. I do find myself turning to Sudoku and other games when I could be wasting time -- oops, "improving my mind" -- on Usenet. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... > In difficult puzzles I sometimes get to a point where logic doesn't > help, or I don't see enough information to proceed. Actually, it means that you've reached the limit of your elimination techniques. There are other more sophisticated ones. I have not become proficient in these higher-order techniques either, but I have some bookmarks for study someday. Brian SignatureDay 357 of the "no grouchy usenet posts" project >> In difficult puzzles I sometimes get to a point where logic doesn't >> help, or I don't see enough information to proceed. [quoted text clipped - 3 lines] > proficient in these higher-order techniques either, but I have some > bookmarks for study someday. And at that stage you really can start arguing about whether you are guessing or not. What many of these techniques are doing is teaching you that when you have numbers in a particular pattern, then only one of the possible numbers in one cell is the right one. You could achieve exactly the same result by picking one and seeing if it lead to a contradiction or not. Instead you say "ah, it's a .... pattern" and pick the right one, or eliminate the impossible one. That's not guessing, but it's based on a very similar approach. The "colouring" technique that I often use on really complicated web-based puzzles is very much like that. You're not guessing, because you start with a square with 2 candidates, and follow both options simultaneously until one gives you a contradiction. So not guessing - chiefly because there's no back-tracking - but very very similar at heart. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk > >> In difficult puzzles I sometimes get to a point where logic doesn't > >> help, or I don't see enough information to proceed. [quoted text clipped - 6 lines] > And at that stage you really can start arguing about whether you are > guessing or not. I don't think so. Once I learn the techniques well enough, then they will be more logic tools. It won't really be any different than ones like "hidden pairs" or other eliminations. > What many of these techniques are doing is teaching you that when you > have numbers in a particular pattern, then only one of the possible > numbers in one cell is the right one. You could achieve exactly the > same result by picking one and seeing if it lead to a contradiction or > not. You CAN, but that's not what you are doing. Many roads to Dublin and all that. Just because you get the same result with two methods doesn't mean that they are the same. > The "colouring" technique that I often use on really complicated > web-based puzzles is very much like that. You're not guessing, > because you start with a square with 2 candidates, and follow both > options simultaneously until one gives you a contradiction. So not > guessing - chiefly because there's no back-tracking - but very very > similar at heart. I'll agree that what you describe is essentially guessing, but that's not the sort of technique I mean. I'm talking things like "X-wings" and "swordfish" and other logical rules. Brian SignatureDay 357 of the "no grouchy usenet posts" project >> The "colouring" technique that I often use on really complicated >> web-based puzzles is very much like that. You're not guessing, [quoted text clipped - 6 lines] > not the sort of technique I mean. I'm talking things like "X-wings" and > "swordfish" and other logical rules. Yes, but X-wing is just a particular pattern that colouring will detect. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk [...] > In difficult puzzles I sometimes get to a point where logic doesn't > help, or I don't see enough information to proceed. By that stage [quoted text clipped - 7 lines] > Others call it a waste of time, and I admit there is something to be > said for that judgement. Thank God the unimpressed party has received a mention! I did a few when I first noticed them, and enjoyed it well enough, but the game soon palled. I mean, when you've completed the thing, you're left with something entirely meaningless. At least a crossword puzzle is made of words. SignatureMike. >> Sun, 24 Jan 2010 23:45:01 -0800 from Evan Kirshenbaum >> <kirshenbaum@hpl.hp.com>: [quoted text clipped - 20 lines] >the two possibilities. continuing until it's proved either right or >wrong. Some people might call that guessing but I prefer to call it testing. At that point of difficulty, where the puzzle is complete enough with potential 'forces' that there is somewhere to go, I do something that avoids messy erasing. Because I don't like pencils. I have a small pad of paper handy, with pages of 3" by 5". I draw a square, about 2x2 inches, with additional lines to make a 3x3 set of boxes. Then I position my new starting point ("hypothesis") relative to the edges, and so on with the logical follow-ons. I make eye-ball references to the original, and superimpose the status/images, even though I don't have the eidetic imaging that some artists possess. I was surprised, a bit, when I first tried this, and discovered that I could, indeed, trace out paths of 30 or 40 or more, and not get confused. (Unless the ballgame that's on gets too interesting.) In my original puzzles, I do annotate the margins of boxes with tiny numbers, and I can't do much of that for the supplement. But it usually isn't needed. >Others call it a waste of time, and I admit there is something to be >said for that judgement. Something, but not entirely a waste. I've learned better to ignore distractions, or to successfully split my attention - At first, I had to sit quietly at home to solve much of anything. (That is - avoid making some fatal error.) Now I can solve while waiting for food in a diner, or work on one intermittantly while watching TV. SignatureRich Ulrich >Sun, 24 Jan 2010 23:45:01 -0800 from Evan Kirshenbaum ><kirshenbaum@hpl.hp.com>: [quoted text clipped - 13 lines] >routine, but one where there's only one correct answer and you can >find it logically. I use the sort of logic you described. I haven't yet come across a puzzle that has two or more correct answers, but I haven't been doing them long. I bought a book of easy to hard puzzles, with an introduction that goes through techniques for their solution (looking for open and hidden doublets, triplets etc.) and worked my way through them. There's really no guessing, if you can determine the correct three squares for hidden triplets, so the book's author maintains. SignatureRobin (BrE) Herts, England > I haven't yet come across a puzzle that has two or more correct > answers, but I haven't been doing them long. I bought a book of easy [quoted text clipped - 3 lines] > can determine the correct three squares for hidden triplets, so the > book's author maintains. I have, but by mistake. That is, I sometimes put online puzzles into one of my programs (I ended up getting a couple because neither was totally satisfactory), and if I make a mistake, the program will warn me that there's more than one solution. But I haven't run into a commercially-prepared one (or one from one of the pieces of software) that has more than one solution. Although, really, without the software telling me, how would I know? SignatureCheryl >> I haven't yet come across a puzzle that has two or more correct >> answers, but I haven't been doing them long. I bought a book of easy [quoted text clipped - 11 lines] > that has more than one solution. Although, really, without the > software telling me, how would I know? It's generally taken that the puzzle can only have one correct solution. There are some techniques ("unique rectangles", "BUG removal" that work on the principle that the puzzle isn't ambiguous. For example, suppose you end up with every square that could have a 1 or 2 in it as having just those two as options, except for one that could be 1,2 or 3. You "know" that the final one must be 3, because otherwise you'd have a set of squares that could all be swapped 1 <-> 2 without affecting any others. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk >> I have, but by mistake. That is, I sometimes put online puzzles into >> one of my programs (I ended up getting a couple because neither was [quoted text clipped - 7 lines] > solution. There are some techniques ("unique rectangles", "BUG > removal" that work on the principle that the puzzle isn't ambiguous. That depends on who sets the puzzles. I once picked up a cheap book of them from an airport news stand, and about half the puzzles had more than one solution. I'd never seen that before. SignatureRegards John for mail: my initials plus a u e at tpg dot com dot au Sun, 24 Jan 2010 23:44:29 +0000 from Mike Barnes <mikebarnes@bluebottle.com>: > I don't think it's possible for a person to solve sudoku without > guesswork of some sort. I disagree, strongly. Every single one I've done -- and the figure is embarrassingly high -- has been completely deterministic, no guesswork involved. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: > Sun, 24 Jan 2010 23:44:29 +0000 from Mike Barnes > <mikebarnes@bluebottle.com>: [quoted text clipped - 4 lines] > figure is embarrassingly high -- has been completely > deterministic, no guesswork involved. In that case, I suppose someone could write a computer program to do the solving. However, I assume that you distinguish between "trial and error" and guesswork. In my opinion, arbitrarily placing a number and then seeing if it leads to contradictions after a few steps is guesswork. SignatureJames Silverton Potomac, Maryland Email, with obvious alterations: not.jim.silverton.at.verizon.not Mon, 25 Jan 2010 09:42:52 -0500 from James Silverton <not.jim.silverton@verizon.net>: > Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: > [quoted text clipped - 9 lines] > In that case, I suppose someone could write a computer program to do the > solving. Indeed yes. Many someones have done so. > However, I assume that you distinguish between "trial and > error" and guesswork. In my opinion, arbitrarily placing a number and > then seeing if it leads to contradictions after a few steps is > guesswork. Of course, and I don't do that. Where would be the fun? SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... Stan Brown <the_stan_brown@fastmail.fm>: >Mon, 25 Jan 2010 09:42:52 -0500 from James Silverton ><not.jim.silverton@verizon.net>: [quoted text clipped - 3 lines] > >Of course, and I don't do that. Not even mentally? SignatureMike Barnes Cheshire, England Mon, 25 Jan 2010 18:41:00 +0000 from Mike Barnes <mikebarnes@bluebottle.com>: > Stan Brown <the_stan_brown@fastmail.fm>: > >Mon, 25 Jan 2010 09:42:52 -0500 from James Silverton [quoted text clipped - 6 lines] > > Not even mentally? No. I prefer working forward ("I need a 3, 5, 6, and 9 in this row; which one(s) can go in this cell?") to working backward ("suppose I put a 3 in this cell; will that get me into trouble?"). I don't say that one method is superior; it's just my preference. Similarly, in math I tend to look for a direct proof before a proof by contradiction. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... > Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: > [quoted text clipped - 12 lines] > number and then seeing if it leads to contradictions after a few > steps is guesswork. Yeah, that's what I've never seen a published (or computer-generated) puzzle require. Mike has said that he regards determining which three numbers in a row are missing as "guesswork", and that's the only definition that would fit. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |It is one thing to be mistaken; it is 1501 Page Mill Road, 1U, MS 1141 |quite another to be willfully Palo Alto, CA 94304 |ignorant | Cecil Adams kirshenbaum@hpl.hp.com (650)857-7572 http://www.kirshenbaum.net/ > > Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: > [quoted text clipped - 17 lines] > numbers in a row are missing as "guesswork", and that's the only > definition that would fit. I think it depends on what techniques you know. As I said, the "very hard" and "super hard" puzzles usually require /me/ to guess, but that's because I don't recognize certain patterns. So maybe you can do them without guessing. I don't know whether one can say in general that if the solver can recognize some complete set of patterns, guessing is never required. -- Jerry Friedman Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: >> [quoted text clipped - 17 lines] >numbers in a row are missing as "guesswork", and that's the only >definition that would fit. I don't quite understand that, but since my name is mentioned, I'll not leave it unremarked-on. Elsewhere I've revised my opinion to say that that determination is guesswork *the way I think most people do it*. That's not to say that guesswork is the only way - for instance it would be possible to (mentally or physically) start with a list of numbers 1-9, strike out the ones that are already present, and use what's left as the result. No guesswork there. Another way is to try the digits 1-9 and see which ones aren't already present. That's guesswork. Both are valid techniques that lead to a definite result. I'm less definite than I was at the start of this thread about my (deliberately provocative) assertion that guesswork is essential. But I'll only be convinced by a realistic worked example that doesn't use it. SignatureMike Barnes Cheshire, England > Elsewhere I've revised my opinion to say that that determination is > guesswork *the way I think most people do it*. That's not to say that [quoted text clipped - 4 lines] > which ones aren't already present. That's guesswork. Both are valid > techniques that lead to a definite result. I disagree that it's guesswork. It's elimination, where known rules are applied with the current data, removing all other choices from contention. At that point, it isn't a guess as there is no alternative. That, of course, includes the case where there's only one choice left in a block, row, or column. Now, people might often take an intuitive leap, but the reason you can put in a number in that case is because you can eliminate the other eight as contenders (having already been used). Brian SignatureDay 357 of the "no grouchy usenet posts" project Default User <defaultuserbr@yahoo.com>: >> Elsewhere I've revised my opinion to say that that determination is >> guesswork *the way I think most people do it*. That's not to say that [quoted text clipped - 8 lines] >applied with the current data, removing all other choices from >contention. I think I see what you're getting at. In your view a guess isn't a guess if it can be discarded simply by applying the rules of the game. But if it's a legal (but possibly wrong) number, that's a guess. I can understand that distinction though it's not one that I observe myself. My original turn of phrase "guesswork of some sort" was intended to convey a broad understanding of the word, but that broadening seems to have been lost somewhere along the line. I'd still like to see a realistic worked example of a hard puzzle that can be solved without guesswork, even using the narrow definition of guesswork that you use. Until I see it, I remain unconvinced that it's possible (never mind practical). >At that point, it isn't a guess as there is no alternative. I think it's important to distinguish between "guess" and "guesswork". I use guesswork (a procedure that includes guessing) to arrive at an answer that is not a guess. >That, of course, includes the case where there's only one choice left >in a block, row, or column. Now, people might often take an intuitive >leap, but the reason you can put in a number in that case is because >you can eliminate the other eight as contenders (having already been >used). [sound of can of worms opening] "intuitive leap"? SignatureMike Barnes Cheshire, England > Default User <defaultuserbr@yahoo.com>: > > I disagree that it's guesswork. It's elimination, where known rules > > are applied with the current data, removing all other choices from [quoted text clipped - 3 lines] > guess if it can be discarded simply by applying the rules of the > game. But if it's a legal (but possibly wrong) number, that's a guess. Every open square has nine possibilities. You then go through a process of eliminating those possibilities until the square has only one possible number. The other way, which I resort to on some tough puzzles is to go through a process like, "Ok, I have a couple 9/2 pairs here. Let's assume the first one is the 9, that makes the other the 2, then this becomes a 7, over here we get an 8, oops that's two 8s in a block. That original one must be a 2." To me, that's guessing. Saying that it couldn't be 9 because there was a pair with a 9 elsewhere in the row isn't guessing. > I can understand that distinction though it's not one that I observe > myself. My original turn of phrase "guesswork of some sort" was > intended to convey a broad understanding of the word, but that > broadening seems to have been lost somewhere along the line. Perhaps I'm not understanding what you mean. Some examples might help. > I'd still like to see a realistic worked example of a hard puzzle that > can be solved without guesswork, even using the narrow definition of > guesswork that you use. Until I see it, I remain unconvinced that it's > possible (never mind practical). I believe that some advanced technique or the other can solve most if not all of the puzzles put forth. > > At that point, it isn't a guess as there is no alternative. > > I think it's important to distinguish between "guess" and > "guesswork". I use guesswork (a procedure that includes guessing) to > arrive at an answer that is not a guess. Again, I'm not totally clear on that. > > That, of course, includes the case where there's only one choice > > left in a block, row, or column. Now, people might often take an [quoted text clipped - 5 lines] > > "intuitive leap"? Probably a bit loose with my terminology. I mean glancing at the row, column, or block and noticing, "Hey, there's no 8." Intution: direct perception of truth, fact, etc., independent of any reasoning process; immediate apprehension. Scrub the term if you prefer. Brian SignatureDay 357 of the "no grouchy usenet posts" project Default User <defaultuserbr@yahoo.com>: >> Default User <defaultuserbr@yahoo.com>: > [quoted text clipped - 18 lines] >To me, that's guessing. Saying that it couldn't be 9 because there was >a pair with a 9 elsewhere in the row isn't guessing. OK. In both cases you've answered the question "Is it a 9?". The difference to you being that it's "guessing" if the 9 is legal but wrong, and "not guessing" if it's illegal? >> I can understand that distinction though it's not one that I observe >> myself. My original turn of phrase "guesswork of some sort" was >> intended to convey a broad understanding of the word, but that >> broadening seems to have been lost somewhere along the line. > >Perhaps I'm not understanding what you mean. Some examples might help. Many "logical" techniques use guesswork under the surface. For instance, do you ever look at a column with (say) two voids and ask yourself what the two numbers are? How would you go about that? Would you (1) try digits 1-9 and find while you're doing it that 2 and 9 are missing, or (2) start with a mental list 1-9, scan the column striking out the numbers that are already there, and find that 2 and 9 are left? Method (1) uses guesswork - trying an arbitrary number and seeing if it creates a problem. That wouldn't qualify as guesswork under your definition, but I hope you'll agree that it qualifies as "guesswork of some sort" when compared with method (2). Or - and on reflection perhaps this is more likely - you *never* look at a column with two voids and ask yourself what the two numbers are. It's becoming apparent to me through this debate that there is a fundamental difference between the way I and some others approach these puzzles (entirely mentally) and the way most people do (recording the possibilities on paper or on the screen). The techniques used are quite different in each case. So the reason I can't proceed with my argument is that I've realised that I'm not familiar enough with the procedures used by most people. >> I'd still like to see a realistic worked example of a hard puzzle that >> can be solved without guesswork, even using the narrow definition of [quoted text clipped - 3 lines] >I believe that some advanced technique or the other can solve most if >not all of the puzzles put forth. Sure, but my assertion (it's a guess, really) is that guesswork underlies some of those advanced techniques. >> > At that point, it isn't a guess as there is no alternative. >> [quoted text clipped - 3 lines] > >Again, I'm not totally clear on that. I hope it's clear by now, but just in case, let's revisit your example of "guesswork" with the 9s and 2s. You made a guess that the first one was a 9. That led to an impossibility so you "inked in" a 2. That 2 was arrived at through guesswork, but it isn't a guess. You know it's correct. And you arrived at it through a logical process. SignatureMike Barnes Cheshire, England > Default User <defaultuserbr@yahoo.com>: > > The other way, which I resort to on some tough puzzles is to go > > through a process like, "Ok, I have a couple 9/2 pairs here. Let's [quoted text clipped - 8 lines] > difference to you being that it's "guessing" if the 9 is legal but > wrong, and "not guessing" if it's illegal? No, I consider it guessing regardless of which you come up with as the correct one. You're trying to make fine distinctions that I don't think are appropriate. There methods to solving the puzzles that don't require a guess at any point. > > Perhaps I'm not understanding what you mean. Some examples might > > help. [quoted text clipped - 12 lines] > definition, but I hope you'll agree that it qualifies as "guesswork of > some sort" when compared with method (2). I disagree. You're trying to make "guess" so general as to be meaningless. Either way you think of the process in your example is actually the same process of elimination. In fact, the first method is to me an unlikely way to do things. I look at the row/column/block that I'm analyzing and subtract the ones that exist from the set of all possibles, leaving the remaining set. > Or - and on reflection perhaps this is more likely - you never look at > a column with two voids and ask yourself what the two numbers are. [quoted text clipped - 3 lines] > (recording the possibilities on paper or on the screen). The > techniques used are quite different in each case. That could be. > So the reason I can't proceed with my argument is that I've realised > that I'm not familiar enough with the procedures used by most people. My method generally consists of a first pass, looking for the same number within rows or columns within aligned blocks, following those into the remaining block, and looking at the cross row or column to see if I can pick off an easy number to fill. After that, I determine my "possibles" for open cells, and begin to analyzed the results. That consists of looking for hidden singles, pairs, or triples, plus looking for cases where a number is confined to one row or column in a block, allowing it to be removed from the remainder of the row or column. This goes on iteratively until the puzzle is solved or I get stuck. Then if it's not too bad, I'll look at guessing. > > I believe that some advanced technique or the other can solve most > > if not all of the puzzles put forth. > > Sure, but my assertion (it's a guess, really) is that guesswork > underlies some of those advanced techniques. I don't think so, but I confess that I haven't gone through them well enough. As I understand, they are mathematically-based. > I hope it's clear by now, but just in case, let's revisit your example > of "guesswork" with the 9s and 2s. You made a guess that the first one > was a 9. That led to an impossibility so you "inked in" a 2. That 2 > was arrived at through guesswork, but it isn't a guess. You know it's > correct. And you arrived at it through a logical process. If it's result of guesswork, then the puzzle was solved by guessing. Even if later logic could be applied to whatever was discovered by the guess. Brian SignatureDay 358 of the "no grouchy usenet posts" project Default User <defaultuserbr@yahoo.com>: >> Default User <defaultuserbr@yahoo.com>: > [quoted text clipped - 15 lines] >are appropriate. There methods to solving the puzzles that don't >require a guess at any point. Actually I thought it was you that was making the fine distinctions about what was a guess and what wasn't. I'm just trying to understand what that distinction is. I thought we were reaching an understanding but we seem to be drifting apart again. Let me try to understand you better. In the first case you describe, it couldn't be a 9 because if you put a 9 in there you get two 8s in the same block. In the second case it couldn't be a 9 because if you put a 9 in there you get two 9s in the same row. You regard one as a guess and the other as not a guess (that's the fine distinction I'm trying to understand). What's the essential difference that makes one a guess and one not? I advanced a theory in my previous message, but you said "no". I confess that I don't really understand what you mean by "a couple 9/2 pairs here", but I don't want to burden you with providing an explanation if you don't feel like it, and so far it hasn't seemed necessary. >> > Perhaps I'm not understanding what you mean. Some examples might >> > help. [quoted text clipped - 19 lines] >I'm analyzing and subtract the ones that exist from the set of all >possibles, leaving the remaining set. Could you be more specific about what you disagree with about my ad hoc definition of a guess, "trying an arbitrary number and seeing if it creates a problem", or why you think it doesn't apply to the procedure (1) above? >>[...] >> I hope it's clear by now, but just in case, let's revisit your example [quoted text clipped - 6 lines] >Even if later logic could be applied to whatever was discovered by the >guess. Over-snipping has resulted in the context of my explanation being lost, and I can't be bothered to restore it. I'm happy if you understand that the result of guesswork isn't necessarily a "guess", even though a guess was used to arrive at it. SignatureMike Barnes Cheshire, England > Default User <defaultuserbr@yahoo.com>: > > No, I consider it guessing regardless of which you come up with as > > the correct one. You're trying to make fine distinctions that I > > don't think are appropriate. There methods to solving the puzzles > > that don't require a guess at any point. > I thought we were reaching an understanding but we seem to be drifting > apart again. Let me try to understand you better. In the first case > you describe, it couldn't be a 9 because if you put a 9 in there you > get two 8s in the same block. You can think of it that way. You can also say, "If I put an 8 in a row that already has one, then it's two 8s. But the key is, you don't NEED to put in a number to test, you work with the list of possibles. I reduce the set of possible numbers for each cell until one remains, then that number goes in. To me, that's different from trying each possible and seeing if it leads to a contradiction elsewhere, then undoing all that. > I confess that I don't really understand what you mean by "a couple > 9/2 pairs here", but I don't want to burden you with providing an > explanation if you don't feel like it, and so far it hasn't seemed > necessary. Say you have a block: 1 3 _ 5 6 _ 8 7 4 The only possibles for the two open cells are 9 and 2. That's helpful in earlier stages, because that means that nowhere else in that column can there be a 9 or a 2, so you can strike those from the list of possibles in all other cells of the column. > Could you be more specific about what you disagree with about my ad > hoc definition of a guess, "trying an arbitrary number and seeing if > it creates a problem", or why you think it doesn't apply to the > procedure (1) above? Because it's not done that way. You don't try a number, you eliminate candidates until one remains. > > If it's result of guesswork, then the puzzle was solved by guessing. > > Even if later logic could be applied to whatever was discovered by [quoted text clipped - 4 lines] > understand that the result of guesswork isn't necessarily a "guess", > even though a guess was used to arrive at it. Sorry if I lost the meaning. As to the latter, puzzles are either solved purely through logic or they're not. If guessing is required, then it wasn't a "pure" solution. Brian SignatureDay 359 of the "no grouchy usenet posts" project Default User <defaultuserbr@yahoo.com>: >> Default User <defaultuserbr@yahoo.com>: > [quoted text clipped - 15 lines] >possible and seeing if it leads to a contradiction elsewhere, then >undoing all that. Yet again I find myself wanting to direct your attention to material that you've snipped. The problem I have with this discussion is that you seem to be reacting to my individual messages in isolation, rather than relating them to what went before. You can see above that I made explicit reference to "the first case you describe", but you snipped your description of that case, and (as far as I can see) responded without giving it any consideration. You could go back and look at the first case that you posted earlier ("Ok, I have a couple 9/2 pairs here"). I think you'll find that your latest response doesn't actually relate to it. Or we could simply drop the subject. SignatureMike Barnes Cheshire, England > Or we could simply drop the subject. Probably best. I'm not sure we're getting anywhere. Sorry if my efforts to trim muddied the waters. Brian SignatureDay 359 of the "no grouchy usenet posts" project Mike Barnes <mikebarnes@bluebottle.com>: >Default User <defaultuserbr@yahoo.com>: >> [quoted text clipped - 21 lines] >that you've snipped. >[...] Sorry if that sounds a bit tetchy. It wasn't meant unkindly. SignatureMike Barnes Cheshire, England > Mike Barnes <mikebarnes@bluebottle.com>: > > Yet again I find myself wanting to direct your attention to material > > that you've snipped. > > [...] > > Sorry if that sounds a bit tetchy. It wasn't meant unkindly. No offense was taken. It wasn't my intention to confuse things through over-snipping, but apparently I did. Brian SignatureDay 359 of the "no grouchy usenet posts" project > Default User <defaultuserbr@yahoo.com>: >> [quoted text clipped - 16 lines] > game. But if it's a legal (but possibly wrong) number, that's a > guess. Pretty much. It's the difference between asking "Do I already have any evidence that rules out this being a 5?" and "If I assume that this is a 5, will I acquire new evidence that shows that the assumption is wrong?" > I can understand that distinction though it's not one that I observe > myself. My original turn of phrase "guesswork of some sort" was [quoted text clipped - 5 lines] > definition of guesswork that you use. Until I see it, I remain > unconvinced that it's possible (never mind practical). I think I can safely say that it's the method I always use. Indeed, when I do them on my iPhone and accidentally forget whether it's set to write "guesses" (as in "I think it's this number; let me know if I'm wrong") or pencil marks (of what possibilities I think are still left) and get told that I was right, I erase the cell and pencil in the possibilities (and forget the answer) until I can prove it. The basic algorithm I use says that if my current knowledge says that n cells in a row/column/section can only have (some subset of) the same n values, then nothing else in that r/c/s can. Initially, of course, n is only 1, for the squares that you're given, but it will typically get up to 3 or 4. And if all possibilities for a number within a (WLOG) section are within a row, then that number doesn't appear elsewhere in that row. For harder problems, I pick a (WLOG) row and enumerate the numbers 1-9, noting which ones are missing. I then go through the empty spaces in the row, checking if there's evidence from the column and section to rule out any of the missing numbers, writing down the ones (or one) that remain. Whenever a set of possibilities is written down for a cell, I check to see if this new evidence has ruled out possibilities for other cells in the cell's row, column, and section. If so, I write down a smaller set for those cells (or, more accurately, cross off little marks). I have yet to find a puzzle that that isn't sufficient for. The challenge, not being a computer that can keep track of a huge context stack, is in doing the full propagation, so typically most of the time is spent searching the grid for evidence that I have that I haven't used yet. For easier problems, there's another move of noticing that a particular digit is know in two of the three sections in a section row or column and checking to see whether the (typically n=1) evidence only allows that digit to be in one cell in the third section. SignatureEvan Kirshenbaum +------------------------------------ HP Laboratories |This gubblick contains many 1501 Page Mill Road, 1U, MS 1141 |nonsklarkish English flutzpahs, but Palo Alto, CA 94304 |the overall pluggandisp can be |glorked from context. kirshenbaum@hpl.hp.com | (650)857-7572 | David Moser http://www.kirshenbaum.net/ Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> I'd still like to see a realistic worked example of a hard puzzle >> that can be solved without guesswork, even using the narrow >> definition of guesswork that you use. Until I see it, I remain >> unconvinced that it's possible (never mind practical). [My thinking has changed elsewhere in this thread since I wrote that: but your response is still interesting.] >I think I can safely say that it's the method I always use. Indeed, >when I do them on my iPhone and accidentally forget whether it's set >to write "guesses" (as in "I think it's this number; let me know if >I'm wrong") or pencil marks (of what possibilities I think are still >left) and get told that I was right, I erase the cell and pencil in >the possibilities (and forget the answer) until I can prove it. Sound. But I wasn't intending to limit my remark to guesswork that's written down. >The basic algorithm I use says that if my current knowledge says that >n cells in a row/column/section can only have (some subset of) the [quoted text clipped - 13 lines] >If so, I write down a smaller set for those cells (or, more >accurately, cross off little marks). Nice explanation. >I have yet to find a puzzle that that isn't sufficient for. If you'd like a challenge, try your hand at this one: http://www.sudoku-help.com/SHPuzNo71.gif The solution and a method are here: http://www.sudoku-help.com/Worked-Example.htm I note that the given method includes trying each of the two possible numbers in a square, propagating the situation, and comparing the results. To me that's getting pretty close to guesswork, though the result is the elimination of a value elsewhere rather than one of the candidates. And of course there might be a simpler way to solve the puzzle. >The >challenge, not being a computer that can keep track of a huge context >stack, is in doing the full propagation, so typically most of the time >is spent searching the grid for evidence that I have that I haven't >used yet. I think the repetitiveness of that task is what put me off doing regular sudoku, along with the need to record the possibilities in the squares in order to tackle the more difficult puzzles (I would need to use a pencil and eraser: yuk). I prefer killer sudoku. SignatureMike Barnes Cheshire, England > If you'd like a challenge, try your hand at this one: > [quoted text clipped - 10 lines] > candidates. And of course there might be a simpler way to solve the > puzzle. That's a stinker. I tried it on paper, but find it almost impossible to find the patterns when I've got pencilled numbers in. Using a tool which just shows me "squares with this number" it was easy to spot the X wing and the Swordfish. I find XY patterns very hard to find, although this was easier than most. That's at the tough end of BrainBashers "Very hard" level. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk > If you'd like a challenge, try your hand at this one: > [quoted text clipped - 10 lines] > rather than one of the candidates. And of course there might be a > simpler way to solve the puzzle. I played with that one for about 15 minutes, and I think I can say that the methods I know can't take me any further. If the result truly involves picking something at random (even from a reduced set) and seeing where it leads, I'd say that it violates what I take to be the implicit contract with these puzzles. >>The challenge, not being a computer that can keep track of a huge >>context stack, is in doing the full propagation, so typically most [quoted text clipped - 5 lines] > the squares in order to tackle the more difficult puzzles (I would > need to use a pencil and eraser: yuk). Nah, you just need to cross off the little numbers (or dots) you put in the cells. I prefer to do them in pen. Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> If you'd like a challenge, try your hand at this one: >> [quoted text clipped - 16 lines] >and seeing where it leads, I'd say that it violates what I take to be >the implicit contract with these puzzles. Agreed. During this thread I've become aware of advanced techniques known picturesquely as X-Wing (which I'd found for myself but didn't know had a name) and Swordfish. They are more complex than the methods you described in your previous message, but are not guesswork. AIUI in X- Wing you look for two rows and two columns (think tic-tac-toe) whose four intersections contain the same possible number. If the rows contain no other possible locations for that number, the columns cannot either - and vice-versa. Swordfish is similar to X-Wing with three rows, three columns, and six intersections. I found out about those techniques by googling (WTQ) "x-wing swordfish". The first hit was a page "Solving Sudoku" at <http://www.angusj.com/sudoku/hints.php>. I see that after X-Wing and Swordfish, but before the even more advanced XY-Wing, it describes a technique called "Solving with Colors". That technique is actually pure guesswork - you arbitrarily pick one of two possibilities for a square and see whether it leads to an illegal situation. The page makes no claim to rigour and doesn't state or even imply that the guessing technique is ever necessary, but it is described as a "recognised strategy" on what is clearly a popular page on advanced techniques. The published method for the puzzle quoted at the top of this message uses X-Wing, Swordfish, and the even more advanced XY-Wing. But not "Solving with Colors". >>>The challenge, not being a computer that can keep track of a huge >>>context stack, is in doing the full propagation, so typically most [quoted text clipped - 8 lines] >Nah, you just need to cross off the little numbers (or dots) you put >in the cells. I prefer to do them in pen. Crossings-off would be possible (if somewhat impractical with small squares, rough newspaper, and the nibbed pen that I like to use). But I rather like the satisfaction of completing a puzzle without mechanical assistance, and the elegance of a completed puzzle with no crossings out or other superfluous markings. To each his own, I suppose - a lot of people like to use a computer, and can no doubt progress further by so doing, but for some reason that doesn't appeal to me at all. Stop press: At the bottom of the page "Solving Sudoku" (URL above) it says "There are some puzzles which can't be solved using simple logic and the only way to solve them is by resorting to trial-and-error." Why "simple" logic, I wonder? Anyway, that's an interesting viewpoint. Every other statement I've seen on that subject says the opposite. SignatureMike Barnes Cheshire, England > Evan Kirshenbaum <kirshenbaum@hpl.hp.com>: >> [quoted text clipped - 29 lines] > and vice-versa. Swordfish is similar to X-Wing with three rows, three > columns, and six intersections. [..] For anyone interested in advanced solving techniques, I recommend the Mensa Guide to Solving Sudoku by Peter Gordon which contains lots of puzzles by Frank Longo. The book guides you through examples at all levels and includes Gordon's solving technique Gordonian Rectangles which he claims he discovered when challenged by Longo with a puzzle that could only be solved by guesswork. I haven't got very far in working through the book but it seems far better in terms of explanation than anything I've found on line so far. SignatureLaura, once a finalist in the Independent Sudoku Championship (emulate St. George for email) James Silverton <not.jim.silverton@verizon.net>: >In that case, I suppose someone could write a computer program to do >the solving. However, I assume that you distinguish between "trial and >error" and guesswork. In my opinion, arbitrarily placing a number and >then seeing if it leads to contradictions after a few steps is >guesswork. Just to be clear, when you say "arbitrarily placing a number", are you arbitrarily choosing a square and placing a random number in it, or are you selecting a square with a small number of legal choices and placing one of those? And when you say "placing", does that imply writing it in its square, or is one allowed to simply imagine it being there? And how many is "a few"? Lastly (sorry for all the questions), what's the difference between that and "trial and error"? SignatureMike Barnes Cheshire, England > Stan wrote on Mon, 25 Jan 2010 09:30:24 -0500: > [quoted text clipped - 12 lines] >then seeing if it leads to contradictions after a few steps is >guesswork. Yes, that's what I would describe as guesswork, and I agree with Stan that it should not be needed. However, I can see that there could be disagreement about where to draw the line; I can look at a puzzle and say "that can't be a two because if it were there'd be two twos in the same row", or I can say "that can't be a two because if it were that other would also be a two, and then there'd be nowhere to put the three in that other column.....". If I don't write the latter argument down, then I don't think of it as guesswork, but if I were stuck and wrote it down as a trial and then discovered it led to an impossibility, that would be guesswork in Stan's definition and mine. And yes, of course you can write a computer program to do the solving; that doesn't distinguish guesswork from deterministic techniques, because a computer program can perfectly well do guesswork; they're rather good at it, in fact. Katy Stan Brown <the_stan_brown@fastmail.fm>: >Sun, 24 Jan 2010 23:44:29 +0000 from Mike Barnes ><mikebarnes@bluebottle.com>: [quoted text clipped - 4 lines] >is embarrassingly high -- has been completely deterministic, no >guesswork involved. Not being a mathematician I'm not certain of the meaning of "deterministic" and whether that's necessarily incompatible with guesswork being required. Additionally I wonder whether substituting "trial and error" for "guesswork of some sort" would make any difference to your reaction. Perhaps someone could point to a worked example that we could discuss. SignatureMike Barnes Cheshire, England > On Sat, 23 Jan 2010 14:49:21 -0800, Evan Kirshenbaum > [quoted text clipped - 19 lines] > equation, read a book, paint a picture -- almost anything is better > exercise for the mind and, IMO, more enjoyable. Funny you should mention that. I've been spending some free time doing sudokus and writing a poem lately. > >Any sudoku puzzle that you're likely to find published will not > >require guessing. What it may require is keeping track of the > >remaining possibilities for a cell and erasing them or crossing them > >out as they become eliminated. Published in a newspaper? I can't do the "very hard" or "super hard" ones at http://www.brainbashers.com/sudoku.asp without guessing (once per puzzle). I often find the "super hard" ones easier than the "very hard" ones. Speculating on the reasons for that would probably be of little interest. Also, I don't think erasing possibilities is ever _required_. You can always check the possibilities for a cell and find there's only one. Writing in possibilities and erasing them speeds that process up a lot, though. > Which is a fancy way of saying "scribbling guesses", as Cheryl said. "Scribbling guesses" could also mean narrowing the possibilities for a cell down to two, arbitrarily picking one, and then seeing whether that leads to a solution. If it doesn't, the other possibility is right. In an extremely hard sudoku, you might have to pick from more than two possibilities or go through this process more than once. Anyway, that's what I took Evan to have taken Cheryl to mean. -- Jerry Friedman > Published in a newspaper? I can't do the "very hard" or "super hard" > ones at [quoted text clipped - 4 lines] > ones easier than the "very hard" ones. Speculating on the reasons for > that would probably be of little interest. I can always do the "very hard" brainbashers ones, but sometimes (1 in 4 or so) get stuck on the "super hard". That's with the "auto pencil marks" to help - I'd hate to do that sort on paper. If you really care, I'd be quite happy to give you hints on any of the that you get stuck with - despite all appearances, this comes from a valid email address. SignatureOnline waterways route planner | http://canalplan.eu Plan trips, see photos, check facilities | http://canalplan.org.uk > > Published in a newspaper? I can't do the "very hard" or "super hard" > > ones at [quoted text clipped - 7 lines] > I can always do the "very hard" brainbashers ones, but sometimes (1 in 4 > or so) get stuck on the "super hard". We must have different methods. > That's with the "auto pencil > marks" to help - I'd hate to do that sort on paper. I like to put in a lot of the "pencil marks" myself so the "auto pencil marks" won't give me any answers, though today I didn't quite manage that. > If you really care, I'd be quite happy to give you hints on any of the > that you get stuck with - despite all appearances, this comes from a > valid email address. Thanks, but I don't enjoy it unless I do it without hints. -- Jerry Friedman >>> Published in a newspaper? I can't do the "very hard" or "super hard" >>> ones at [quoted text clipped - 22 lines] > -- > Jerry Friedman In trying out different programs, I discovered that they all seem to have different methods of deciding the level of difficulty of a puzzle. All the programs seems to decide that all of the ones on the CBC website, from the one-star (very easy) to the five-star (most difficult) are pretty easy. SignatureCheryl Cheryl filted: >I decided a year or two ago, rather belatedly, that I was going to >figure out how to do sudoku puzzles. I can now do them, but haven't yet >acheived the ability to do any but the simplest with paper and pencil. >The erasing gets on my nerves, as does the scribbling of guesses >followed by the attempts to read them. I decided from the moment I found out about them that I was going to avoid learning anything more than the general concept...it's intuitively obvious that they're the purest known form of OCD-bait....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? >Cheryl filted: >> [quoted text clipped - 7 lines] >learning anything more than the general concept...it's intuitively obvious that >they're the purest known form of OCD-bait....r Is OCD Japanese for nerd? SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE Chuck Riggs filted: >>I decided from the moment I found out about them that I was going to avoid >>learning anything more than the general concept...it's intuitively obvious that >>they're the purest known form of OCD-bait....r > >Is OCD Japanese for nerd? No, that's "otaku"....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? >Chuck Riggs filted: >> [quoted text clipped - 5 lines] > >No, that's "otaku"....r Dank u. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > Cheryl filted: > [quoted text clipped - 12 lines] > An optometrist asks whether you see the glass > more full like this?...or like this? Do you know I thought exactly the same whence I first read of these puzzles. I solved one and then thought well that's that,solve one and you've solved them all. R. PS. I've g**gled the term and find that they are "related" to magic number squares that have been in existence for quite a while. BUT, and I have had argued this point in many a tavern discussion that these things in their modern form HAVE to be computer generated. I have no objections to being disabused of this stance if wrong. R H Draney in <hjg2i3022k6@drn.newsguy.com>: >Cheryl filted: >> [quoted text clipped - 7 lines] >learning anything more than the general concept...it's intuitively obvious that >they're the purest known form of OCD-bait....r IBTD: I was confronted with a similar riddle of numbers in the verbal math exam for my Abitur (like college). I'm afraid this was three decades ago and I forgot the title of the task. After 30 mins of preparation I presented it to the examiners. Everything went fine in an adequate analytic and mathematic way - up to a certain point. I wrote down the following number stating "At this place, it is a 5" and was startled by the examiner's question: "How did you know?" I had to admit "I guessed it" and to my great pleasure he confirmed "Guessing was the correct way to solve at this point!" ObTopic: I like Sudoku, especially online versions where non-binding notes are impossible. You enter a number to a field, and if it proves wrong later you may delete it. Alas, you'll rather have to cancel the whole thing because you can't distinguish the previous correct entries from the following wrong ones. OCD? I prefer "brain jogging"! Ciao, Paul Paul Schmitz-Josten filted: >R H Draney in <hjg2i3022k6@drn.newsguy.com>: >> [quoted text clipped - 9 lines] > >OCD? I prefer "brain jogging"! It stops being fun when you can't look at anything with crossing vertical and horizontal lines without trying to stuff numbers into it...I had that with chess back in high school and I still can't enjoy waffles....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? >Paul Schmitz-Josten filted: >> [quoted text clipped - 15 lines] >horizontal lines without trying to stuff numbers into it...I had that with chess >back in high school and I still can't enjoy waffles....r Since Sudoku puzzles are essentially all the same, don't you prefer the challenge of solving word puzzles of, for example, the Hofstadter or Martin Gardner variety? SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > Since Sudoku puzzles are essentially all the same Not so. They differ widely in difficulty, and some demand intricate logic for solution. SignatureJohn Varela Trade NEWlamps for OLDlamps for email >> Since Sudoku puzzles are essentially all the same > >Not so. They differ widely in difficulty, and some demand intricate >logic for solution. Intricate logic means guesswork, in the case of solving Sudoku puzzles, does it not? Guesswork as in, try for a solution and if it doesn't work, try for another combination of numbers, and so on until you have the final answer. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE Chuck wrote on Tue, 26 Jan 2010 12:33:52 +0000: >>> Since Sudoku puzzles are essentially all the same >> >> Not so. They differ widely in difficulty, and some demand >> intricate logic for solution. > Intricate logic means guesswork, in the case of solving Sudoku > puzzles, does it not? Guesswork as in, try for a solution and > if it doesn't work, try for another combination of numbers, > and so on until you have the final answer. I suppose "trial and error" Sudoku could be considered equivalent, in a trivial sense, to mathematical proofs that use computer listings to show that various possibilities lead to contradictions. I forget how many pages of computer listings the recent proof of the four color theorem used. I understand that it is regarded as inelegant but correct. SignatureJames Silverton Potomac, Maryland Email, with obvious alterations: not.jim.silverton.at.verizon.not > Chuck wrote on Tue, 26 Jan 2010 12:33:52 +0000: > [quoted text clipped - 13 lines] >pages of computer listings the recent proof of the four color theorem >used. I understand that it is regarded as inelegant but correct. At least I don't have to waste any more paper napkins on the thing. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > >> Since Sudoku puzzles are essentially all the same > > [quoted text clipped - 5 lines] > doesn't work, try for another combination of numbers, and so on until > you have the final answer. Logic as in: there are only two sevens in this row, and there are only two sevens in another row, and all four sevens fall into the same two columns. Therefore, there can be no other sevens in either of those columns. No guesswork, pure logic. That particular piece of mildly advanced logic is called a "swordfish" for reasons that completely escape me. SignatureJohn Varela Trade NEWlamps for OLDlamps for email >> >> Since Sudoku puzzles are essentially all the same >> > [quoted text clipped - 12 lines] >That particular piece of mildly advanced logic is called a >"swordfish" for reasons that completely escape me. To attack a Sudoku problem at the start, you guess a particular number for a particular square. After cranking in successive guesses for the various squares, arriving at a solution that satisfies the Sudoku conditions amounts, IMO, to mind-numbing drudgery, work for the beast in us. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > To attack a Sudoku problem at the start, you guess a particular number > for a particular square. No, that's not how they are solved for the most part. I can't believe that you could have read the other messages in the thread and looked at the referenced sites and still held this belief. > After cranking in successive guesses for the > various squares, arriving at a solution that satisfies the Sudoku > conditions amounts, IMO, to mind-numbing drudgery, work for the beast > in us. Your conclusion comes from a false premise. Brian SignatureDay 359 of the "no grouchy usenet posts" project > To attack a Sudoku problem at the start, you guess a particular number > for a particular square. After cranking in successive guesses for the > various squares, arriving at a solution that satisfies the Sudoku > conditions amounts, IMO, to mind-numbing drudgery, work for the beast > in us. I've read that some people solve Sudoku puzzles by guessing, but I've never understood how that can work. If you guess, you have to be prepared to back up and undo everything you've done since the guess. How many of us could keep track of what was there before the guess was made, and what is new? It's easy to do this with a computer program, but very difficult for a human. And in practice you'll end up with a nested sequence of guesses, making things even more complicated. Guessing is out unless you want to make things very difficult for yourself. SignaturePeter Moylan, Newcastle, NSW, Australia. http://www.pmoylan.org For an e-mail address, see my web page. Peter Moylan filted: >> To attack a Sudoku problem at the start, you guess a particular number >> for a particular square. After cranking in successive guesses for the [quoted text clipped - 9 lines] >difficult for a human. And in practice you'll end up with a nested >sequence of guesses, making things even more complicated. I think this is the point of most of the arguing going on in this thread right now...programs can try a partial solution and continue until the puzzle is either completely solved or no more progress can be made...if the latter, the program backs up to the last "guess", changes it to some other partial solution and proceeds again...if all possibilities have been exhausted at that juncture, the program backs up to the *previous* "guess" and does the same thing there.... People can do this too, but most of us have much shorter stack capacity.... The brute-force, depth-first solver places a backtrack mark at *every* step of the solution; the skilled solver appears to be jumping over many steps at once, but is actually collapsing them into a single operation by doing the intermediate steps intuitively...practice improves the intuition by burning common patterns into the mind, which is exactly the kind of mental damage I hope to avoid by staying clear of the puzzles....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? >Peter Moylan filted: >> [quoted text clipped - 27 lines] >common patterns into the mind, which is exactly the kind of mental damage I hope >to avoid by staying clear of the puzzles....r Good man. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE ... > I've read that some people solve Sudoku puzzles by guessing, but I've > never understood how that can work. If you guess, you have to be [quoted text clipped - 5 lines] > > Guessing is out unless you want to make things very difficult for yourself. Say I'm doing a paper sudoku, which I hardly do any more. and I've done everything I can without guessing (not counting Mike Barnes's sense). In all the unknown cells I've written in a small "font" every number that I can't eliminate for that cell. Now I pick a cell with only two possible numbers, say 35. I guess the 3, so I draw a square around the 3. Say there's another square in the same row with the pencil marks 36. I put a circle around the 6. I go on circling numbers till I've solved the puzzle or reached a contradiction. If the latter, I know the cell with the 35 was the culprit by the square instead of a circle. So I put a full-size 5 in there and solve the puzzle based on that. This isn't hard to adapt to on-line sudokus such as the ones I do at Brain Bashers, but the paper version is easier to explain. Brain Bashers and the paper sudokus I've done hardly ever require nested guessing. I enjoy that much less. -- Jerry Friedman >... > [quoted text clipped - 26 lines] >Brain Bashers and the paper sudokus I've done hardly ever require >nested guessing. I enjoy that much less. You and I, Jerry, follow the same routine, more or less. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > ... > [quoted text clipped - 26 lines] > Brain Bashers and the paper sudokus I've done hardly ever require > nested guessing. I enjoy that much less. An alternative technique is to draw a diagonal cross in the square in which you put your guess. Place the guess in one of the four triangular compartments. Continuing drawing crosses and entering the result in subsequent squares. If you eventually get to a contradiction and need to guess again, go back to the original square and use a different compartment. I don't really regard this as guessing, just an aid to conducting a reductio ad absurdum of arbitrary length. -- Mike Page >> ... >> [quoted text clipped - 34 lines] >compartment. I don't really regard this as guessing, just an aid to >conducting a reductio ad absurdum of arbitrary length. There is no reason to be ashamed of sophisticated guessing. Many differential equations are most easily solved that way. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE >> To attack a Sudoku problem at the start, you guess a particular number >> for a particular square. After cranking in successive guesses for the [quoted text clipped - 11 lines] > >Guessing is out unless you want to make things very difficult for yourself. Back when I used to mess with the ornery things, I put my initial guesses in small numbers in the corners of the squares. When I found that a particular guess was wrong, I'd draw a line through it and I'd put my better guess in another corner. When I knew for certain a number was the right one, I put it in the centre. If all of that sounded messy, it was simpler and quicker in the execution than in the explanation. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE Peter Moylan <gro.nalyomp@retep.?.invalid>: >> To attack a Sudoku problem at the start, you guess a particular number >> for a particular square. After cranking in successive guesses for the >> various squares, arriving at a solution that satisfies the Sudoku >> conditions amounts, IMO, to mind-numbing drudgery, work for the beast >> in us. I don't think Chuck's description realistically depicts the process for most people, though a child or someone with reduced mental faculties might proceed that way (and find it rewarding). >I've read that some people solve Sudoku puzzles by guessing, but I've >never understood how that can work. If you guess, you have to be [quoted text clipped - 3 lines] >difficult for a human. And in practice you'll end up with a nested >sequence of guesses, making things even more complicated. You seem to be assuming a high degree of parallelism. That isn't implied by Chuck's description, or usually required. SignatureMike Barnes Cheshire, England >Peter Moylan <gro.nalyomp@retep.?.invalid>: >> [quoted text clipped - 7 lines] >most people, though a child or someone with reduced mental faculties >might proceed that way (and find it rewarding). Piss off, Mike. SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE Sat, 23 Jan 2010 16:15:38 -0500 from James Silverton <not.jim.silverton@verizon.net>: > In hiding, and thanks for the correction. I tried a few sudoku puzzles > but have not looked at them in a while since I found the process of > trial and erasure to be rather a bore. I suppose a semi-automatic > computer method showing possibilities could be set up and I did begin to > program that before losing interest. I recommend http://www.websudoku.com/ It works quite well on line, though the $20 downloadable version has a few extra features. SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... >> Hello All! >> [quoted text clipped - 12 lines] > >The OED has cunningly hidden it under the correct spelling, "sudoku". I've heard some British people pronounce it persistently as "soduko". SignaturePeter Duncanson, UK (in alt.usage.english) >>> Hello All! >>> [quoted text clipped - 14 lines] > > I've heard some British people pronounce it persistently as "soduko". I can't help myself from saying a different wrong form, "suduko". It doesn't help that I know how to spell it. SignatureJames James Hogg filted: >>> The OED has cunningly hidden it under the correct spelling, >>> "sudoku". [quoted text clipped - 3 lines] >I can't help myself from saying a different wrong form, "suduko". It >doesn't help that I know how to spell it. I reckon it's too late to do anything about "karaoke"....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? >>>> Hello All! >>>> [quoted text clipped - 17 lines] >I can't help myself from saying a different wrong form, "suduko". It >doesn't help that I know how to spell it. For its correct spelling, see http://www.sudoku.name/ SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE On Jan 23, 2:22 pm, "Peter Duncanson (BrE)" <m...@peterduncanson.net> wrote: > >> Hello All! > [quoted text clipped - 14 lines] > > I've heard some British people pronounce it persistently as "soduko". First syllable as in "Sod this for a lark"? -- Jerry Friedman >On Jan 23, 2:22 pm, "Peter Duncanson (BrE)" <m...@peterduncanson.net> >wrote: [quoted text clipped - 19 lines] > >First syllable as in "Sod this for a lark"? Unfortunately not. The "o"s are as in "so". SignaturePeter Duncanson, UK (in alt.usage.english) Peter Duncanson (BrE) skrev: > I've heard some British people pronounce it persistently as "soduko". That is not a British problem. It is international - for German-Roman languages at least. I have to force myself to pronounce it properly. The word is being discussed in the German language group as well. Googlehits on the net: sudoku 39'400'000 sudoko 251'000 suduko 235'000 sodoku 229'000 soduku 132'000 soduko 115'000 sodoko 92'300 suduku 57'900 In all: 40'512'200 Errors: 1'112'200 - prct: 2.8 % I would have thought that it was worse. SignatureBertel http://bertel.lundhansen.dk/ FIDUSO: http://fiduso.dk/ > Peter Duncanson (BrE) skrev: > [quoted text clipped - 19 lines] > > I would have thought that it was worse. Even "sodutu" gets some hits. By the way, "prct." is not in my dictionary of abbreviations. SignatureJames Bertel Lund Hansen filted: >Peter Duncanson (BrE) skrev: > [quoted text clipped - 14 lines] >sodoko 92'300 >suduku 57'900 I wonder if the top two followers are experiencing interference from a couple of Indonesian political figures....r SignatureA pessimist sees the glass as half empty. An optometrist asks whether you see the glass more full like this?...or like this? R H Draney skrev: > >Googlehits on the net: > >sudoku 39'400'000 [quoted text clipped - 5 lines] > >sodoko 92'300 > >suduku 57'900 > I wonder if the top two followers are experiencing interference from a couple of > Indonesian political figures....r It did cross my mind that I might have hit upon a normal word in another language or a transcription thereof, but I do not have the expertise (nor the inclination) to check it. SignatureBertel http://bertel.lundhansen.dk/ FIDUSO: http://fiduso.dk/ On Jan 24, 2:52 pm, Bertel Lund Hansen <splitteminebrams...@lundhansen.dk> wrote: > Peter Duncanson (BrE) skrev: > [quoted text clipped - 22 lines] > -- > Bertelhttp://bertel.lundhansen.dk/ FIDUSO:http://fiduso.dk/ Hmm. Eight spelling variations. Find another one and we could replace the numerals 1 to 9 with these alternative forms. But what should we call such a puzzle? R. > On Jan 24, 2:52 pm, Bertel Lund Hansen > <splitteminebrams...@lundhansen.dk> wrote: [quoted text clipped - 28 lines] > > But what should we call such a puzzle? I have a friend who constructs a variant of sudoku with 16 x 16 grids. You might be thinking he uses hexadecimal numbers, but no, he uses the 16 runes of the younger futhark. The result is a fiendlishly difficult time-waster. He calls them by the Old Norse-inspired name "stavstad", a variant of "number place" meaning "(runic) stave place". SignatureJames > I have a friend who constructs a variant of sudoku with 16 x 16 grids. > You might be thinking he uses hexadecimal numbers, but no, he uses the > 16 runes of the younger futhark. The result is a fiendlishly difficult > time-waster. The Washington Post has started printing 16x16 hex sudokus in the Sunday paper. SignatureJohn Varela Trade NEWlamps for OLDlamps for email Redshade in <dc574caa-3e20-42bd-bd7c-5466ae472d4e@e11g2000yqe.googlegroups.com>: >> sudoku 39'400'000 >> sudoko 251'000 [quoted text clipped - 4 lines] >> sodoko 92'300 >> suduku 57'900 >Hmm. Eight spelling variations. Find another one and we could replace >the numerals 1 to 9 with these alternative forms. > >But what should we call such a puzzle? "Harry P." or "Miracle" ? These are permutations of two vowels at three positions. We get 2 times 3 results, making up 8. . . Rataplan... . . Abracadabra... . . 9 ! Wizard's regards, Paul >> Hello All! >> [quoted text clipped - 12 lines] > >The OED has cunningly hidden it under the correct spelling, "sudoku". The more correct spelling is found here: http://www.sudoku.name/ SignatureRegards, Chuck Riggs, An American who lives near Dublin, Ireland and usually spells in BrE > The OED has cunningly hidden it under the correct spelling, "sudoku". Does it explain what the name means? Is it one of those things like "basubaru" for baseball? >> The OED has cunningly hidden it under the correct spelling, "sudoku". > >Does it explain what the name means? Is it one of those things like >"basubaru" for baseball? OED: Japanese S{umac}doku (1984 or earlier) < s{umac}- (in s{umac}ji number) + -doku (in dokushin single status), after S{umac}ji wa dokushin ni kagiru, lit. ‘the numbers are restricted to single status’, former name of the puzzle. Thank goodness we got the abbreviated form of the name. SignaturePeter Duncanson, UK (in alt.usage.english) > Hello All! > [quoted text clipped - 10 lines] > I admit I was surfing the dictionary after I found I had been wrong > about "somewhen". You didn't get to "su..."? SignatureI don't know what they have to say It makes no difference anyway; Whatever it is, I'm against it! [Prof. Wagstaff] James Silverton set the following eddies spiralling through the space-time continuum: > I admit I was surfing the dictionary after I found I had been wrong > about "somewhen". It gets occasional use along with elsewhen and everywhen, all these are usually combined or contrasted with the corresponding -where forms. Signatureξ:) Proud to be curly Interchange the alphabetic letter groups to reply Sat, 23 Jan 2010 15:54:33 -0500 from James Silverton <not.jim.silverton@verizon.net>: > In view of the current popularity of Sodoku puzzles, it is interesting > to see that the OED entry has not caught up. Perhaps if you tried the spelling "sudoku"? I have never seen "sodoku". SignatureStan Brown, Oak Road Systems, Tompkins County, New York, USA http://OakRoadSystems.com Shikata ga nai... |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2012 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.