Feedback and Questions

Glad to be of service!
Hope you're landing somewhere fun

There are numerous possible ways to solve any single puzzle and your choices are just as valid. As I wrote before, basic ones must come first but other than that, your choice. You have to step through each step but the solver will skip ahead to the first to si

No you dont need to try the lost in order except for basic strategies, whihc should be exhausted before moving to anything higher. Within basic strats these can be looked for en mass or in a different order but I think as listed then are in a natural order of complexity. There is a lot of overlap between more advanced strategies and it is difficult to order them in terms of complexity - so my ordering is my best subjective ordering. You can skip some by turning them off. I've wondered about allowing the user to re-order them but it's a complex UI problem, not a sudoku strategy problem.

Very sorry! Made a change yesterday and didn't check all probably browsers.
[CRTL]+[F5] to refresh the page now

I get where you are coming from. Searching for a *any*pattern in the blizzard of numbers is difficult just with pen, paper and eyeball. I started this project hoping there would be simpler strategies (but still not trial and error) and was surprised, that in the difficult puzzles those I've documented really are the best solutions. I'm certain that few newspaper puzzles rarely have the need for them and we compilers have to look very hard for hard puzzles (the distribution is weighted to easy by a long way). The site tries to record all valid approaches even ones that are very academic (that is ones that are best handled by a computer) but I do get some people claiming they solve with these methods by hand. I don't know, is the short answer. I can see that some better tools on the solver and in the app could help. But I'd need a very detailed description of how a person would use them as I don't solve individual puzzles in the same way. My work is generalizing from a strategy to find algorithms that work across many puzzles.
The other problem is space. The app doesn't really have the space to fiddle with candidates. I could try a zoom or make it tablet only. I'll take a look at candidate highlighting and links in the next version then.

I will attempt to do that next version

Welcome to the Internet. Thanks for the typos. Believe me some very eagle eyed people have been reading these documents for over a decade (the older ones) and sending me corrections, so it still amazes me there are typos to be found, so well spotted.

Yes, you are correct in identifying a bias. To me the puzzle is how we can crack Sudoku puzzles generally rather than specific instances. I've struggled to follow the discussions and strategies on the forums (back in the day, some of these forums are sadly lost now). Partly this is because some people, clearly very clever people, were nevertheless poor at describing their ideas and I was often reduced to studying a bare example to see if/how it could be generalised and picking through the esoteric terminology. The process of documenting it at a level I could understand was part of the way to understanding it for me and that helped make an algorithm to detect all cases. I try and illustrate the scope of a strategy with exemplars and get all the variations which must come from the algorithm. Sudoku is very deep in that there are still example puzzles with bottlenecks that will never be solved with simpler strategies, that's just it's nature. Some strategies are far from pen and paper solutions and are rather academic, I agree, but one of the aims of getting them out in public is to research ways to replace them with better, simpler strategies, if that is possible. Which is why it's such an exciting puzzle to ponder.

Thanks!

That’s a sound idea and you catch me in an idle moment. So refresh the solver. I've given new solved cells a blush. I can't see to do yellow since the shadow color seems to blend with the font color which is blue. Odd but otherwise ok I think