Lummox'

Constraint Subsets look useful. It’s easy to see that one of the big fish is in fact one of his constraint subset configurations. In this case, the "A" subset consists of columns 5, 7, and 8 and subset "B" consists of box 3 and rows 5 and 9.

- Code: Select all
`. . . | . . . | # # * `

. . . | . . . | # # *

. . . | . . . | # # *

------+-------+-|-|--

. . . | . . . | | | .

* * * | * X * | X X *

. . . | . | . | | | .

------+---|---+-|-|--

. . . | . | . | | | .

. . . | . | . | | | .

* * * | * X * | X X *

Similarly for the "carnivorous" jellyfish with the "A" subset composed of columns 1, 4, 6, and 9 and the "B" subset of box 5 and rows 3, 5, and 7, which is okay since set "B" is allowed to have intersections.

- Code: Select all
`| . . | | . | | . . | `

| . . | | . | | . . |

X * * | X * X | * * X

|-----+-|---|-+-----|

| . . | X * X | . . |

X * * | * * * | * * X

| . . | X * X | . . |

|-----+-|---|-+-----|

X * * | X * X | * * X

| . . | | . | | . . |

| . . | | . | | . . |

Unfortunately I don't see how to apply the theory to the other forms of big fish. Can it be used or is POM or some other technique the correct approach? Also do Constraint Subsets stand on their own as a solving technique or are they consumed in other approaches. For example the

hidden pattern is a nice loop consisting of grouped strong links.

Havard, if it makes you feel any better, your jellyfish is not really carnivorous. It consists of two finned swordfish (columns 1, 4, and 9 and columns 1, 6, and 9). The first eliminates r5c56; the second, r5c45. The jellyfish, as near as I can determine, then eliminates candidates in the other 16 cells.