Man Versus "No Knight's Move" Sudoku

[@meredithnavin1358]

This video is 3yrs old and I can see how much Simon has improved since this video was put out. I would say he'd be able to do this puzzle in probably 15-20 minutes now.

I love how he repeatedly goes out of his way to admire how good/beautiful/gorgeous the puzzle is.

[@bzzzb8]

I find it very interesting how no two people will have the same approach to a puzzle. I was doing this on paper along with him trying to learn this no Knights move and while he is much much better than I am I solved squares in a different method at times. I will admit it would have taken me a lot lot longer to do it without watching him solve the first 35 minutes LOL

[@stuartmcconnachie]

35:38 “Can’t be a 5. Err, why? It is a 5!

[@33gazer33]

At 4:30 when he talks about the one's I realised I was so far out of my league it was more interesting to watch the brilliance than take part

[@danieltaber4924]

Anyone else screaming at him that he solved the 5s extremely early, and didn't notice because he didn't clean up the pencil marks?

[@Wallopy_Joe]

"this can't be a five" - /removes the correct five/

[@aebbingeable]

At it now haha, it is interesting to see that he was solving one of those fives and then not continuing with it. Normally when I solve classic sudokus I always continue with the number i just put in because it's ruling out a lot.

[@kevgermany]

But we all do this. Miss the blindingly obvious chasing complex answers.

[@aebbingeable]

@@kevgermany i'm not judging him in any way (hope it didn't look that way, in that case: sorry) i just find it oddly fascinating how he sees it different than i do

[@kevgermany]

@@aebbingeable me too, part of the attraction of watching. But it does stop me from concentrating on the solve, I find myself thinking 'he must be blind, how can he not see it?'. Better entertainment than a scripted perfect solve, makes me feel less bad when I do the same miss. And, to be clear, this one is way beyond my ability.

[@Kriskiwi7]

I bet your next sudoku puzzle will be a "No Rook's Move".

[@tomekduresov706]

Lol, but the interesting question is, is it even possible to create no queen moves sudoku puzzle..

[@AznUzer]

@@tomekduresov706 No because diagonals are not usually considered for sudoku puzzles

[@tomekduresov706]

@@AznUzer yeah, but the question is, is it actually possible to even set it up..

[@AznUzer]

@@tomekduresov706 probably not

[@emorgan0085]

I think that such a puzzle could exist, but with even a very small number of givens would be incredibly easy to solve. It may just be impossible too, I'm not totally sure.

[@DrZaius3141]

34:14 You got an 89 pair that both interact via knight's move on the 3rd tile in the bottom row which is already limited at 189, fixing it at 1.

[@perry4754]

This was a fun and difficult puzzle. I'm glad I was able to work out most of the basic logic myself, but I did end up getting a decent amount of help from the video.

[@bogbel1]

That was a BEAST! Very much enjoyed it.

[@voiceoftheguns27]

As someone who enjoys playing chess and sudoku, this is a brilliant puzzle variation. Every time I think there can't be another variation of sudoku, I am proven wrong. So much fun to watch and solve with you for my first experience with this.

[@RedGryz]

Goodness, this puzzle is a monster. Even having adopted your techinques, it still took me more than an hour for my successful solution.

[@Mizziri]

The uniquenesses in this puzzle were absolutely beautiful, I think the first time I've spotted uniqueness so easily.

[@NelielSugiura]

At 21:45, you incorrectly placed a five in r8c5, so you took a while to see the combo to remove a five in r5c4 & r5c6 due to the two fives in r3c5 and r7c5, thus making r6c6 your five.

[@Umbra451]

25:03 Once you have the 1 and the 5 in the middle square, that locks in the 1 and the 5 in the middle column, doesn’t it? By the same logic discussed earlier in the video, where a number on the middle of a square’s side locks it into one of the corners of that square, if you know the four corners of the middle square, you also know the middle numbers on the adjacent sides.

[@robertu2927]

Great puzzle. But I dont get the seven (30:04) in box 9. How can it ruel out the 2-6 pair?

[@robertu2927]

OK! Got it!

[@CodyPyper]

I also don’t understand how that logic works. Can anyone explain?

[@dkalambokis78]

Be more specific ​@@CodyPyper

[@cz1822]

I have a sudoku puzzle book called “sudoku on the half shell.” It has sudoku puzzles that contain only half the grid. It contains the bottom left corner of the grid so that row 1 has one cell, row 2 has 2 cells, row 3 has 3 cells, etc. and column 1 has 9 cells, column 2 has 8 cells, column 3 has 7 cells, etc. The entire puzzle is a triangle. It solves like normal sudoku, but some of the smaller 3x3 grids are missing cells due to the puzzle’s triangle shape, so they contain only 6 of the numbers. Additionally there can be no repeated digits along diagonals.

[@jakobwachter5181]

This is a variant called Sujiken. Pretty fun!

[@tiemen596]

20:00 at this point I realized that the 4 numbers in r3c5, r5c3, r5c7, r7c5 must always be the same numbers as r4c4, r4c6, r6c4, r6c6 in a no-knights move sudoku. In this case it would solve the 1 in r3c5 early. But as it is generally applicable logic, it can be a powerful trick to remember

[@MyLolkaka]

At 37:20 you explain why the square can't be a 9 and in the same move you put an 8 in there to which the same rules apply :p

[@leestoddart7014]

For some reason this video from 3 years came up as a suggestion today. I really enjoyed this "old fashioned" software and Simon's approach but ... why oh why did he use uniqueness to solve the 267 in box 9? I was able to solve without that so surely Simon could have done it.

[@samdavis6722]

When you're determining that the square can't be a 7 at around the middle of the video, you can actually determine that square must be a 1 due to the powerful knight logic you are talking about. Each number in those positions that create the cross pattern must end up in a corner of the middle 3x3. Since one of the corners is already filled in as a 1, you know it must either go in the top 3x3 or bottom 3x3 since the side positions are already occupied by the 2 and 8. Since the bottom 3x3 already contains a 1, you can put the 1 in the position you are testing the 7 in.

[@6stones]

If there are N givens in a No Knight's Move" Sudoku variant, how many 'missing givens' would the possible squares ruled out by knight's distance (for a particular number) actually correspond, in relation to N, (since we clearly have more information available, not only givens)? Or is this a mathematical knightmare a la Alice from which we would like to wake up?

[@jakobwachter5181]

This shouldn't be too hard assuming you know how many times a square can be clued in by its neighbors. You can do the math, but I'll spare you the calculations; just know that: 4 squares can be clued in by 2 other squares; 8 squares can be clued in by 3; 24 squares can be clued in by 4; 20 squares can be clued in by 6; 25 squares can be clued in by 8. Assuming there are N givens on any board, the total number of ruled out values by knights' moves is 2 * (4/81) + 3 * (8/81) + 4 * (24/81) + 6 * (20/81) + 8 * (25/81) = 5.53 extra "ruled out" values per clue given (per N). Without really doing math (because there is no rigorous proof) but using the fact that each given in sudoku rules out 24 others, this gives a magic number of 408 (17 givens is the smallest possible number of givens that produce uniquely solvable sudokus: 24 * 17). Since the above calculation basically says that each extra given in NKM Sudoku gives you 5.53 extra ruled out values on average, that means that you are getting 29.53 fewer values to consider per clue. This implies that No Knight's Move sudoku can be uniquely solvable with as few as 14 givens: 14 * 29.53 is greater than 408. This also implies that on average, a 17 clue knights move sudoku is almost as easy as a 21-clue normal sudoku. Take with a grain of salt. :)

[@stefanholbek2449]

Simply fantastic! ;O)

[@thisisonlymyfew2863]

i liked that . thank you for sharing

[@benderakakingen8367]

what program are you using?

[@somniad]

12:40 I'm not super great at this stuff; what's your logic exactly for that being a 3?

[@msolec2000]

That's the only place it can go in that block. Column 7 is taken by the 7-9 pair, and row 6 already contains a 3, so that just leaves one possible square that can contain a 3.

[@somniad]

@@msolec2000 haha I see that now, that's very silly of me to miss.thanks!

[@thomaswilke6312]

This is such an interesting concept

[@Socrates...]

Surely you can take a hint number and do a knights trail across the whole board , mark off each of those blocks as being impossible in red, and a huge portion of the board is knocked out from the get go

[@StrategicGamesEtc]

Like a knightrider? Why would that be the case?

[@Socrates...]

@@StrategicGamesEtc if a knight move is invalid from the source of the provided numbers then surely you can do a 'knight's tour' from origin and discount many supposed options along the way. When the tour reaches a dead end, buffeted by the origin provided numbers then you go on to next number........I don't know if this is true I want to know either way

[@kristiandyhrberg9189]

@@Socrates... If you do a knight's move, the only thing you know about the square you land on, is that it cant hold the same value as the one you jumped from. Continuing the knight's tour from there, will not tell you anything about the next square you land on. If that makes sense.

[@Socrates...]

@@kristiandyhrberg9189 so the knight's move is only valid for a one move boundary around the discounted initial square? In other words if you knew 2 squares that are invalidated, you cannot use a knight's tour between those two arbitrary squares and invalidate a huge swathe of the board for that particular number?

[@thomassutton4670]

The knights rule isn't a standard strat I can use in regular sudoku, is it. The board has to be specifically made for that rule. Am I correct in my thinking?

[@icycloud6823]

Yes. It's unique to this puzzle and other puzzles that are made with the specifc rule.

[@Pawelo00992]

aaaaaah i get it now... never played chess in english hence the confusion

[@bixnash1]

Is the knights move an actual thing throughout all puzzles???

[@Nullref_Arcana]

32:58 where he marks a 5, why couldn't it be an 8?

[@mcmanhoso768]

It's the only place the 5 can go, if you put an 8 there then there is no place for the 5

[@Pawelo00992]

i dont understand whats the knights move, help

[@treksez]

In chess a knight moves in an L pattern, so 3 moves in a straight line, then 1 move off that. So in this variant numbers can't be a "knights move" away from each other

[@hazurl]

@@treksez 2 moves in a straight line*

[@AKHalex]

Isn't this easier than normal sudoku? It seems like it's exactly the same, but with more hints and ways to rule out candidates.

[@albertlyngpetersen8702]

32:41 Wow it annoyed me for like 10 minutes that the foves were pencil marked wrong. But beautiful puzzle though...

[@samholden5758]

You can place the 1 at row 3 column 5 right at the start, right? The one in the corner of the central box means there must be a 1 either there or r7 c5, r5 c3, or c5 c7. And all the other ones are either occupied or can't have a 1 due to there already being a 1 in their box. That would admittedly remove some of the fun logic later...

[@jamesahibbard]

I also saw this move. C5R3 C3R5 C7R5 C5R7 must all match C4R4 C6R4 C4R6 C6R6. That 1 in C5R3 was available very early in the solve.

[@joek511]

I put this puzzle into HoDoKu. It says multiple solutions.

[@Lens98052]

At one point you assumed that the puzzle has a unique solution to proceed. It drives me crazy when my wife uses that assumption to solve things, as you are assuming that the puzzle is well defined and the author has not made a mistake. OK to do in a solving contest, I guess. However, mistakes happen all the time in chess problems. Once a chess puzzle creator started showing his puzzles to the then world champion chess player Capablanca who, real time, started showing multiple solutions to several of them, making them bad puzzles. In this case, the puzzle was generated by a computer program so, failing bugs in the program (not a given), it will have a unique solution.

[@edgechill]

Very interesting but way too difficult for my tiny brain. Even though I play chess.

[@atis1976]

Some of the tricks not too difficult. At start c2r7 can not be a 5, cos it eliminates block 8. At 3:16 c4r1 can not be 5 cos it eliminates row 3. At 4:50 c5r5 can not be 5, cos it eliminates column 2.

[@IcyWasTakenAgain]

first but no one cares :P

[@AznUzer]

You’re right

[@kai-ju5fv]

I don't care that you're the first one or whatever, but I do care about your motive with this channel. Surely your brain capacity is insufficient to understand these puzzle as evidenced by your comment, so why bother watching? Don't get me wrong, it is a good thing that this channel is growing, just that I never thought it would get on the interest of a fellow ape. cheers.

[@MrJosephdrummond]

in my book, using a pencil is cheating. i have never seen someone use a computer to play sudoku.