Well [-l + come]; one of your comments was erroneous, as you said yourself (the one you retracted), another comment reads like a restatement of a popular comment predating yours by a over year (which you acknowledged yourself), and the third makes a pretty sweeping claim about superdeterminism not being Turing computable. Unfortunately, the proof you provide seems flawed on a couple of counts.* However, even if the proof did turn out to stand, people frown upon comments which do not give more explanations and context to sweeping statements that seemingly come out of thin air (even if they did turn out to be correct). FYI, I didn't read (until now) or vote on any of your comments.

That makes 3 plausible downvote explanations for 3 comments, two of which you mentioned yourself. I'm surprised about your surprise.

* (Superdeterminism doesn't require that part of the overall program can be perfectly predicted by a much smaller program in advance, nor that the outcome of the smaller program can then be used to change the overall outcome. At least two reasons: 1) Not being able to verify complete correspondence (except by fiat), given all hidden variables and their potentially unknowable context (unknowable from within the program, and the context may encompass the entire universe); 2) superdeterminism can in principle be saved simply by saying that the agent isn't able to show a contradiction; i.o.w. in a superdeterminist universe, a perfect prediction-machine conditional on which a contradiction can be derived cannot exist, by definition of what "superdeterminism" means. Your thought experiment would be inapplicable in a superdeterminist universe, strange as it sounds. In that light, your proof reads similar to the one that shows that a Halting problem decider cannot exist. Alternatively, the agent would be unable to use the result to show a contradiction. While such an inability would indeed seem strange, from the universe's point of view, every facet of that inability would have been predetermined anyways.)