Thanks for the feedback.

On top of that, it actually ends in the optimal outcome.

Just to clarify, no it doesn't. It's implied that the 20 deaths is worse than 5 for the consequentialist protagonist.

The analogy is problematic ... bizarre and unrealistic example.

Thanks for the feedback. It seems people have been split on this. Others have also found the analogy problematic. On the other hand an analogy doesn't usually attempt to provide proof, but illustrate the structure of an argument in an understandable way. I don't think it's bizarre if you think of thought-experiments like the Chinese Room or Descartes' invisible demon.

. The moral thing to do for a drowning person is save them. But if you saved these people, you'd all die anyways

I think this is a more substantial attack on my writing. I probably need to improve the scenario so that it is clear that the people in the river could have been saved in the later part of the parable. Thanks.

At best, it proves "there is an possible circumstance where a fairly poorly thought-out instrumentally rational belief is inferior to a true one

Well that's roughly right but a bit of limited interpretation. I'd say "truth-belief has a significant instrumental value beyond what we can derive from our immediate circumstances and goals" and additionally "non-truth can have cumulitive negative utility far beyond what immediate circumstances suggest". What I'm trying to suggest is that the uncertainty with which we go through life means that it is rational to assign truth-belief a very significant utlity beyond what can be explicitly identified with our present knowledge. In other words, be sloppy with the truth and you'll never even know you failed. I write this because "rationlity is winning" and the raw math of most people's approach to instrumental rationlity neglects this generalised component entirely, only focusing on how the truth helps us in the here-and-now. I hope I at least partially communicated this in the parable.