I'm working with this grid as part of a new ALS search strategy to be posted soon. But, now I'm totally Para-noid!
Please tell me....is there a fish in here?
For there to be a stand-alone fish, you have to be able to have candidates seeing each other in a loop through the unsolved boxes. With this setup, the only cyclic path through any of the unsolved boxes would be b1-b2-b5-b6-b9-b7-b1.... Since the candidates in b1 and b7 cannot directly see each other, there is no cyclic box path so there won't be a useful stand-alone fish or coloring reduction (outside of possibly a line-box or box-box locked candidates deduction).
One quick follow-up question on your "trick"...
Does “stand-alone” mean naked only, or can a successful cyclic path also indicate a possible hidden fish?
I am not sure what a hidden fish is. Unlike locked sets where a naked set is coupled with a hidden set, a fish group is always coupled with just another fish group.
By stand-alone, I meant a fish deduction that could not be represented more simply, like with a hidden single or locked candidates deduction.
To show what I mean, in your example, pretend that the 7 in r4c4 does not exist. Then you would have an r135/c246 swordfish acting to eliminate r5c7, but you would also have the simpler locked candidates in box 5 row 5 making the same deduction.