The New Sudoku Players' Forum

[Sudtyro2]

If one wants to understand AICs, best to stick with Myth Jellies original threads on the subject.

Thx to Arkietech, DAJ and DonM for their comments!

Myth's AIC is defined as an open chain that starts and ends with strong-inference links. It is bi-directional! There is NO need to assign specific parities (true or false) to the start/end candidates. The start/end candidates are actually linked by a "derived" (ronk's term) strong inference, meaning that they cannot both be false, which obviously means that at least one of them must be true. Hence, any candidate that can "see" (weakly link to) both ends of the chain can be eliminated. If one insists on a closed (discontinuous) loop, then just include the two weak links. But, why obfuscate the open chain?

I once asked Myth about open chains ending with two weak links, which "places" the candidate having a strong-inference link to each end. He responded that those cases are actually "subsumed" by the regular AIC. It took me awhile to realize that, but it's true. So, no need to have a second rule for "derived" weak inference.

As a manual solver, I prefer the KISS principle...keep it simple, stupid!

SteveC

[ronk]

The start/end candidates are actually linked by a "derived" (ronk's term) strong inference, meaning that they cannot both be false, which obviously means that at least one of them must be true.

The "derived strong/weak inference/link" terms were actually coined by Steve Kurzhals AFAIK. That said, it is perhaps true that he did not apply the term to an inference as simple as this.