The New Sudoku Players' Forum

[Homer]

Hi – The Samurai puzzles are flawed! A couple of The Times Saturday 30th July’s constituent SuDoku puzzles were not unique (“B” and the central one). Therefore depending on the order that you took to solving them, there was a 50% chance of not solving the whole puzzle correctly.

I initially solved “B” with r5c1 = 9 & r5c3 = 5 and r9c1 = 5 & r7c3 = 9, giving a valid solution to that Sudoku. This permutation though does not allow a valid central Sudoku solution. However, interchanging the 5’s & 9’s does give a correct solution to the whole Samurai. Solving this problem was of course made worse by the fact that the central Sudoku did not have a unique solution either! R9c4 could be a 1 or 7 with r9c5 being a 7 or 1 correspondingly and two other pairs cycled round as well for 1s & 7s (r1c5 with r3c4 and r1c8 withr3c8).

I feel this made solving this particular Samurai a case of pot luck and therefore against the whole spirit of logically solving these types of puzzles. What do you reckon? I wonder how many others fell into the same trap? Does Pappocom need to offer an apology? Comments please.

[george-no1]

Homer,

Thank you for clearly expressing what I have been feeling since Saturday morning. I fell into the same trap, and since I have never failed to complete a Su Doku puzzle before, I decided that the puzzle was invalid and stopped immediately. I hoped someone would post what you did, as I was too angry to do it myself.

However, its is not Pappcom's fault; Pappocom does not supply the Samurai puzzles to The Times, ergo he does not owe us an apology. It certainly is against the spirit of Su Doku in my opinion and The Times should be notified of this.

George

[Homer]

Ooooops my fault. Who sets the Samurai puzzles in the Saturday Times? I thought it was Wayne Gould / Pappocom?

If its not you, Wayne (or your Pappocom company) - sorry, no slight ment against you or your excellent software.

[PaulIQ164]

I have to say I disagree. I solved the Samurai using entirely logical moves - this implies there is indeed a unique solution. From your post, it sounds as if you're trying to tackle the constituent sudoku puzzles in isolation, which will never work. The idea of the puzzle is that you solve it as a whole - the five grids cannot be considered separate until the overlapping boxes have been completed.

[Homer]

Paul, I do initially try solving the Samurai as a "whole" taking into account the other Sudoku parts. My point is that if you then concentrate on one area, then even applying the logical Samurai rules, because two of the parts had more than one solution, it is possible to end up with a conflict. I can forward on all the details (a bit tedious granted - but I have dubbed them into the Pappocom software) to anyone who is interested in checking that two of the constituent Sudokus are not unique.

I have previously solved Samurai Sudoku puzzles relatively easily; indeed I was one of the winners of the Samurai puzzle on July 2nd (anyone want to make me an offer for my prize of 2 (unused) Times Sudoku books?

[george-no1]

Paul,

It seems as if you solved the grids in the *right* order. That doesn't mean that everyone did, and in fact two unlucky people who didn't are Homer and myself.

The very fact that *there is a right order* goes against the spirit, and, in my opinion, the whole point of the puzzle.

George

[PaulIQ164]

I'll have a crack at solving it in your order, then.

But first, let me try to figure this out. You say you found a valid solution to grid B, but that it didn't let you solve the central grid. You then changed your grid B to another valid solution (by swapping the 5 and 9, I think you said). Well then you didn't solve grid B at all did you? There might very well be multiple solutions to grid B if you try and finish it before you can, because you haven't done the central grid. It's not that I stumbled upon the right order, the way it would seem is that I simply did it in the right (indeed, the only logical) order, and you did it wrong.

Edit: to clarify, because I realise that that last paragraph was a tad garbled: When you were solving grid B, if you found one of two pissible solutions, you must have made a guess somewhere along the way (there's no way to arrive at one of several solutions without guessing at some point). That means you're effectively doing trial and error. However, if you do trial and error on a Samurai, you have to do it over the whole five grids. So when you found the central grid to be unsolvable, that was the error, so your guess way back in grid B was wrong. Incidentally, the puzzle could be solved without recourse to trial and error.

I apologise profusely if I've misunderstood what you said.

[Enigma]

I'm with Paul on this one. I solved most of grid B first, with about 3 or 4 pairs of 5 and 9 needing assistance from the central grid and beyond to be correctly placed. Grid B couldn't be solved in isolation - this is the beauty of the Samurais to me. Just because you can find a "valid" solution to one of the grids doesn't mean it's actually valid in the broader context of the whole puzzle. I've never had to guess yet on the Samurais - each has been solvable IMHO by using straight logic (and not very fiendish at that).

Paul

[Homer]

Enigma, I didn't try to solve Grid B in isolation - but after getting some cells completed correctly from the central grid in the overlapping box. Grid B does have 2 solutions as I described in an earlier posting. However after the central grid is completely solved correctly, then yes Grid B comes down to a single solution. My point is that if, as I and others did, you tackle Grid B without fully completing the central grid, then you have a 50% chance of getting a solution, where the overlapping 3x3 box is then wrong for completing the central grid. I don't think you are supposed to solve Samurai / Sudoku in a particular sequence to get them right!

Also if you first solved the central Sudoku grid (with minimal help from the other grids), then again that has two solutions. If you have the Pappocom software, I can send you the saved puzzles to verify for yourself these points. Please do send a private message to me with your e-mail address & will e-mail you the saved puzzles.

[PaulIQ164]

Of course you have to solve them in a particular sequence! The crucial point here is that when you tackled grid B, you by your own admission put in one of two possible solutions. You must therefore have made a guess (as evidenced by you saying you had a 50% chance). If you're going to guess, of course you might get the wrong solution. Instead of making a guess at all, you should have moved back to the other grids, and tried them a bit.

[Anette]

I did the same mistake as Homer the first time I tried Samurai Sudokus.

But what you need to keep in mind is that when you enter a number in any of the shared squares, you must make elimination of candidates in both the center part and the corner parts. It is can happen very easily that I miss making all eliminations, and I find Samurai Sudokus take up too much time for this mechanic elimination compared to normal Sudokus.