You're reading too much into what RomeoStevens wrote - at no point did he explicitly mention the one-shot Prisonner's Dilemma.
The only two contexts I know where the expression "defection risk" has meaning is prisoner's dilemma and Cold War spy/counterintelligence games.
he's just using "defection risk" as a shorthand
I think it's a wrong expression here both connotationally and denotationally.
You may object that the model is not good enough
I'm not saying it's not good enough, I'm saying it's wrong. To repeat myself, one-shot PD is quite rare is real life. Trying to shoehorn "many morally-relevant situations" into this mold is not the right approach.
Yes, Prisoner's Dilemma, but not only one-shot. You added that detail in with no attempt at justification, and it is not justifiable.
(also, games similar to Prisoner's Dilemma with explicitly cooperative and defecting strategies, but details details)
I was reading reviews of HPMOR on Goodreads and I noticed that the people who didn't like the book were essentially "put off by the rationality". They thought Harry was arrogant and condescending.
Then I was thinking, a lot of people are "put off by rationality" for similar reasons. What a shame. There's a lot of value in spreading rationality, and this seems to be a big obstacle in doing so.
Any thoughts on how to make people less "put off by rationality"? I think the core issues are: