Hi PIsaacson,

Ronk wrote:I posted a similar deduction recently, which had only one cannibalistic elimination. After Allan Barker pointed out via PM a smaller pattern which made the same elimination non-cannibalisticly, I quickly deleted the post. If there's a similar smaller pattern for your three 8s, I can't find it.

I'm not qualified to comment on rules or common practice, as I often don't know myself. However, there are "sub-solutions" in your solution that will eliminate the 8s non-cannibalistically. It takes 3 slightly different solutions to eliminate them one at a time. One example is, first in terms of "cell sets":

Sets: A = {r123c4} = {2589}

B = {r7c4,r9c456} = {156789}

Then in terms of base/cover sets:

PIXA 22 Nodes, Raw Rank = 1 (linksets - sets)

7 Sets = {12379N4 9N5 9N6}

8 Links = {589c4 2b2 1678b8}

--> (8c4*8b8) => r8c4<>8,

In terms of base/cover sets, there are 7 base and 8 cover sets so any two cover sets that overlap can cause an elimination, in this case column set 8c4 overlaps box set 8b8 thus => r8c4<>8, one of your (now exorcized) cannibalized victims.

Curiously, there is a 13th elimination, r9c2, due to a particular cover set group. This is eliminated by row set 8r9, which is part of this cover set group.

PIXB 27 Nodes, Raw Rank = 0 (linksets - sets)

8 Sets = {123789N4 9N5 9N6}

8 Links = {8r9 589c4 2b2 167b8}

--> r5c4<>5, r5c4<>9, r7c5<>7,

r9c2<>8], r9c4<>8,

Maybe the cannibalized candidates should be considered collateral eliminations?