Why Play Sudoku - At ALL!

Advanced methods and approaches for solving Sudoku puzzles

Why Play Sudoku - At ALL!

Postby RichardGoodrich » Wed Jun 18, 2025 3:58 pm

I can't remember when I first started playing this infernal game! Perhaps the Devil made me do it. I bought Denis Berthier's DB Hidden Logic of Sudoku 2nd Edition HLS2 and spent over a year trying to read the damn thing! I ended up joing this forum and started learning Python. Even then I was toward the end of my career as a digital hardware designer who did a lot of r&D. What was I doing programming? I am now 72 going on 73. The only programming class I had at Texas A&M was Fortran WhatFor. The program had to be submitted on those dreaded IBM punch cards by bowing before those guardians of the Main Frame! So I asked myself What IV am I even doing this kind of nonsense - and stuck with digital hardware design. As a Junior I head about these crazy kids, Bill Gates and Steve Jobs - a few years younger and doing wild things with little computers ...

Anyway, I programmed little processors with machine language, discovered assembly, later Basic and finally toward "the end" Python - which is the one I really liked - fits my personality! I digress ... I had just enough math to be dangerous - enough to deal with 1's and 0's anyway! Then this crazy French Prof DB got me irritated with all his pedantic stuff about MS-FOL. I decided to just get to the crux of it - the 4 Grids - rcn, rnc, ncr, bns. I learned some Python in the process. Used a bit of it in my career. All of a sudden: The "naked" in the 4 grids "exposed" the "hidden"! Clarity at last! Those naked subsets in the non rcn grids often revealed fishy patterns in the rcn. Wow!

As an r&D guy I was always excited about the "new", the "exotic" making hard things easier. I developed a "systematic" way of thinking in my career and personality. I was discovering all these ad hoc approaches and said to myself: -Self, there has to be a better way! Along the way I found about Knuth's DLX and eventually found Ali Asaf (now a google researcher - working on his PhD then). An article "Dancing Links in 30 Lines" showed up. I got permission and probably his help to add that to my Python SDK (call it biGSDK). Now I could easily with this "exact cover" algorithm solve any puzzle. Also found Peter Norvig trying to cure his wife of her "Sudoku Addiction" with his well publicized code (had a major error in the begining which he of course fixed - must be 8 of the 9 values present in the puzzle) So why keep playing?

Where was the fun in all that! As time passed: 4 back surgeries, the great fun of getting older and sicker meant I needed something! I regretted NOT doing more math! Recently, I heard about this ED stuff! Now I was just focused on about 5.4 billion grids. It got interesting again! "1to9only" actually programmed a minlex algorithm for me in Python. And all these math types on this forum. Glen Stephen Fowler became a "biggy". Just seems AT&T did him and his old colleague, David A. Korn a "dirty" Suddenly Bell Labs was NOT interested in free Open Source software any more and let the ole' "grey beards go" (Experienced a bit of that in my carreer too.)

biGSK was written to be the "perfect player" that I could not be! John Welch of Systematic Sudoku was also an influence. But only on the Systematic idea. I still wanted to use a PC. One of my ideas about programming is to simply use the PC as a less error prone and extemely patient human! It just "cracks" me up when some say Sudoku does not use math - just logic Where does "logic" come from - math / philosphy? Anyway, the 4 grid implementation in my mind was just an extension of pencil marks [PM]. Any human can do this and understand the process! That made it Interesting and thus FUN to me. I was born an awkward klutz both mentally and physically. Using a program is a human extension reducing the simple mechanical bookkeeping to almost nil! Win-Win for me.

Of course there is all this CHAIN theory. Interesting! DB elevated it to such a HIGH PEDANTIC level. Got his own special notation which I can understand, but why? His chains, lassoes, whips, braids just boil down to chains and collections of chains (with certain useful restrictions). The big problem is using these advanced techinques and chains! HOW do you find them! Chains can be so intellectually satisfying!

To expand on chains: All patterns can be expressed as chains. Subsets in other grids discover patterns thus compacting what could also be solved with chains. Then the big Bug-A-Boo. Forcing Chains and Nets subsume most if NOT all chains. Easy to do in a systematic way. But, they are not DICk! The "purists" want that!
They just work and they work easily. HoDoKu is a great platorm to use to find them!

Then there is all the T&E debate! Oh NO: this is not scientific. Math guys do not like "reasoning by cases"! Where is the closed form solution / the symmetry? Satan himself is causing you to CHEAT. Some of us love to cheat: Creatively of course! The trick is to find a systematic way of doing it! DB has his plan, I have mine, and others have theirs. Who cares? Wonder how many chains have actually been found using Forcing Methods. I am now using HoDoKu by restricting the solver to subset and forcing methods. I use apriori knowledge of the solution to see what I can see. E.G. On Arto Inkala's 2012 hardest puzzle. I can "cheat" by asserting the solution to one cell. The rest of the puzzle just falls apart with basic methods (sets, intersections, and basic forces)

Or I can with "apriori" knowledge assert each unsolved cell and count in the solution path the longest strings of singles. Then by asserting two such cells solve the rest of the puzzle with mostly singles and two intersection patterns (claiming & pointing) ! A BACKDOOR for sure! Is this Cheating? Of course it is and I love it! So how did you cheat with your discoveries! Found those exotic patterns and chains because they were "obvious to the casual observer" did you!

Just previewed DB's PBCS3 work. Yeah just more of the exotic. All beautiful theory. Just no systematic way of finding how to apply it! DB even admits that much longer and simple applications of T&E methods is not just easier, but perhaps more satisfying to the player in the long run!

So what now! Well there is the ED. The minlex canon. Digging into the respositories of 1to9only. The gangs of 44 and 416. You get the idea! Now, I want to learn C programming again (run the gcc compiler). Can I make my biGSDK Python 4 grid thing work again (using PyCharm probably). Maybe, I will learn the Rust language. Could I learn combinatorial / graph theory for the first time? Could a database of puzzles and solutions be created? What will be some of the many-to-one and one-to-many mappings? Should we reverse the idea of simplest first? Would that help find puzzles that serve as examples for the larger subsets or fishy patterns without having to do eliminations first? If I live long enough and want to keep the brain alive!

So why do I play? Same reason mountains are climbed - because it is there! Sudoku may be more about the journey than the destination. Why do you play?
Big1952
RichardGoodrich
 
Posts: 119
Joined: 12 December 2012
Location: Josephine, TX (north of Dallas, TX)

Return to Advanced solving techniques