If you bypass the whole "attack only the points of the analogy which do not relate to the argument" and focus on the points of the "analogy" that make it a suitably demonstrative specialization of the more general argument (and thus a formally valid argument, which in general an analogy is not, and thus a neither is a strawman counterargument), then:
General Premises:
1. Agent A enforces Action X onto Agent B
2. Action X hurts C with some probability. Not Action X hurts B with high probability.
3. For all persons D under A, the probability of Not Action X is vanishingly small.
Reasonable Argument:
1. If B performs Action X bad things happen to C. This makes them partly responsible. However, under premise 3, B's counterfactual contribution is vanishingly small, whereas A's contribution is close to 100%.
Unreasonable argument:
1. As the final link in this particular chain of events, B is wholly responsible for the harm done to C, because had they suffered harm with high probability, they would have ensured the reduction of the probability of non-harm to C by a vanishingly small amount.
General Premises: 1. Agent A enforces Action X onto Agent B 2. Action X hurts C with some probability. Not Action X hurts B with high probability. 3. For all persons D under A, the probability of Not Action X is vanishingly small.
Reasonable Argument: 1. If B performs Action X bad things happen to C. This makes them partly responsible. However, under premise 3, B's counterfactual contribution is vanishingly small, whereas A's contribution is close to 100%.
Unreasonable argument: 1. As the final link in this particular chain of events, B is wholly responsible for the harm done to C, because had they suffered harm with high probability, they would have ensured the reduction of the probability of non-harm to C by a vanishingly small amount.