A Brand New Trick For Very Hard Sudoku Puzzles

Thanks to Derek Neal for this extremely interesting new method he calls "The Slot Machine". You can try the puzzle shown in the video here:

This puzzle is taken from the Brainium Sudoku app and you can find out more about the app at these links:

App Store

Google Play

▶SUPPORT US + ACCESS EXCLUSIVE "CRACKING THE CRYPTIC" PUZZLES/VIDEOS◀

▶SEND US PUZZLES TO SOLVE/CONTACT US◀

▶FOLLOW US◀
Twitter: #crypticcracking
@crypticcracking

▶SOFTWARE◀
Or our own software, which you can access by clicking on the links under the video.

▶ABOUT US◀
Hi! We're Simon Anthony and Mark Goodliffe, two of the UK's most enthusiastic puzzle solvers. We have both represented the UK at the World Sudoku Championships and the World Puzzle Championships. We're also "cryptic crossword" aficionados. Mark is the eleven-time winner of The Times championship and Simon is the former record holder for most consecutive correct solutions to The Listener crossword. We hope we can help your puzzle solving while also introducing you to some of the world's best puzzles.

Thank you for watching!
Simon and Mark

[kristofersokk]

This is genius, it makes total sense. I've had the intuition myself that following one number and its possibilities can give results but I haven't put much attention to trying it out myself.

[msclrhd]

If you look at the logic for the X-Cycle from the solver at 1:49, that is the same as the logic with the 5s at 5:44. This technique looks like an interesting way of easily identifying a subset of X-Cycles.

[hotonis]

So I do a ton of Sudoku puzzles on an app that generates. And this happened in the hardest difficulty after watching the video. It gave me an immediate break on the puzzle and was able to solve it in less then 10 minutes for the first time. I love this new method and thank you for sharing!

[SudokuSwami]

Well, I realize that you guys like to solve puzzles without using automated and complete Candidate Lists. So this is a very interesting technique. It strikes me as a Trial & Error method. Right? Similar to Forcing Chains. I especially like the idea that if you can construct 2 chains with no contradictions, based on opposing premises, you can then eliminate the "elbow" Candidates, as you call them. But I must say, after you filled in all the Candidate 5's at 4:41, it was immediately obvious that you had a Sashimi Swordfish in Rows 3, 6 & 9, allowing for the elimination of the 5 in R4C7, which then leads to the solution of the same three 5's that you eventually found, but with a lot less rig-a-ma-role. And then at that point, if you are looking at ALL the Candidates, the Puzzle just falls apart, and is nothing but Singles, except for a set of Locked Candidates on 2 in Column 5, about 2/3 of the way through. So really, just the one move, at the very beginning, solves the whole puzzle. But, to each his own, I guess! Ha-ha. There was also an obvious Standard Swordfish on the 1's when you filled them in, (again in Rows 3, 6 & 9), but after solving those three 5's, you didn't even need it. :-))

[draconicdusk]

This is genius and so much FUN to use. not only that, it becomes so much more potent the more other numbers you have filled.

[shadowwalker]

That's an incredible technique. I tried staring at the grid for 5-10 minutes with just a few pencil marks to show for it. I used the technique discussed and solved the whole puzzle from a fresh timer in 16 1/2 minutes

[jdoe]

Super cool! I actually do this technique myself, but only as desperation, because I never realized having the set of 3 numbers like that are what enables it. Knowing that makes it so much simpler to intelligently incorporate into my solves!

I tried this myself before watching, and did find the "slotmachine" on 1s, but only after much head scratching over all the things I couldn't do. Hopefully next time I spot it sooner!

[Malygon]

Oh wow, I've been using that technique for years but I didn't know that this wasn't well-known and I also didn't know what the ideal conditions were for this technique to work, as in having exactly three numbers in positions so that each square is affected by 2 numbers. I just used it whenever I saw numbers forming loops like that. Now I can look for it more actively!

[-siberian-]

I tried this technique on this sudoku. Now I imagine this sudoku was made for this technique, as this completely solves itself once you go at it. I will have to try this in some of the sudoku's I could never even come close to solving with the more traditional techniques I know. But honestly love this technique already!

[Qazqi]

I'd be interested in hearing Mark's thoughts on this. It has that guessing element that he's spent so much time mastering in order to quickly solve his puzzles, and I'd like to know if there's any overlap for him or if he has anything to add.

[SonnySandberg]

Okay, first time using that method and I cracked the puzzle you provided in 27 minutes.
I was able to use the method on first 5, 1 and 8. Gave me a little each time. Then by doing normal sudoku and usual methods, no x-wings and all the other stuff, just the easy methods always shown, it was cracked open.
I do not regard myself as a fast solver. I am very impressed by this method.

[petemagnuson]

comparing the 1 and 5 examples with the steps the solver was taking, it looks like the slot machine is a more human-friendly way to find identify certain X wings and X cycles.

[olemification]

I believe this technique works well on any loop with at max 3 candidates within each 3x3. The 3 numbers in different 3x3 column/row is not the limiting factor, but it certainly makes it easier to find candidate loops to solve. You could also achieve similar loops where ever you have locked the candidates of the necessarry squares. I've sort of been doing this technique, but have been struggling to identify when the technique might be fruitful and when it's not. Derryl's insights definately will help me improve. Great work, thanks Derryl!

[Kyrelel]

I am not a Sudoku master - far from it. I am self-taught, sadly, by being forced to wait for my partner to turn up - sometimes 2 hours late!!! - and decided to do puzzles in the mean time.
I came across this channel a few weeks ago as it appeared in my suggestions list, and I have watched 3 or 4 of the videos.
Imagine my surprise to find that this "new" technique is one I have been using for over 10 YEARS !!!

[williamwester]

Going back in time to solve some classics that pre-date my finding the channel. Like others, I feel I use a variation of this technique without calling it as much. The presence of 3 numbers in separate "3x3 rows" and "3x3 columns" is something I look for when I'm stuck. I use a color to highlight possible cells. It helps me spot swordfishes. If no swordfish, then I look for a 3x3 with only two cells where choosing either option "knocks out" two possibilities in another 3x3. Using different colors for the two "chains" allows me to eliminate cells where both colors point (i.e. elbow cells).

[isaacbooth]

Wow. This crazy. My grandpa was the first one to teach me sudoku and I noticed this. He said it didn’t work so I agreed and I use it now whenever I can. It kinda made sense to me but this makes it click and it’s easy to explain. Thanks

[emorgan]

The "closed loops" always satisfy at least one of the following properties: a) the loop can be broken down into multiple sub-loops, or b) you can eliminate enough possibilities from each box so that at most only 2 locations remain.

[CaptainHandsome]

One minor observation I have is that if you look at the grid with the 'elbow' cells highlighted, if you imagine placing a 1 in any of the elbow cells it eliminates both the other elbow cells. The same also seems true with the 5s, and placing the correct 5 as shown in the video seems like the only possible move that doesn't eliminate 5 from any of the elbow cells. I could be totally off base here, but maybe it's something worth considering.

[someguy-kh]

Like many others on this comment board, I too have been doing the same thing for years. It's called guessing. It's highly educated guessing followed by execution of it-then-else logic to find a contradiction. It's much more powerful when you involve two numbers that are bound together, because then you discover two number chains at once, but this can also be very effective.


If this is what you need to get a doctorate in Sudoku theory, then I think half of use deserve one too.

[ctsirkass]

I have been using this technique as long as I can remember, but never from the start like this, mainly because the puzzles I solve are not diabolical. It was always a sort of last resort for me in order to eliminate "elbows" but it rarely revealed whole sequences of numbers. Interesting to see it being formulated as a system in its own right.
It's also interesting to note that this technique works best for people like me that like solving puzzles without pencil marks.