I might have missed something but isn't the "solution" to the concerns about the completeness money pump equivalent to the agent becoming complete.

E.g. after the agent has chose B over A it now effectively has a preference of B over A-. 

I haven't worked this through e.g. the proof of VNM etc. but are we sure this weaker notion of completeness might end up being enough to still get the relevant conclusions?

(quite busy might have a bit more of a think about this later)

I'm really uncomfortable with this post. It implicitly supports immoral ends justifying the means actions (trapping people in their cars) if the result is good.

So personally I think there's a bit of a difference with that example given there were all sorts of laws preventing black people's political participation at the time. Also the fact MLK embraced non violence indicates that there are still lots of relevant side constraints on actions beyond just ends justifying the means.

Good to know about your second bit.

Sorry, editing to make this easier

I'm a bit unsettled by this post. One of the major concerns about utilitarian recently on the forum has been the means justify the ends thinking that would licence people to do bad things in the pursuit of good ends.

It's completely right that EA should distance himself from a sort of Fraud to Give scheme. But equally protests for a good cause that involve causing lots of harm to the public by blocking roads should also be rejected.

I guess I'm just a bit worried if lots of EAs are associated with the Extinction Rebellion/Just Stop Oil etc stuff.

"Please also note that my computer stubbornly refuses to calculate the true geometric mean of odds of the distribution by taking the 5000^th root of the results, so I’ve used an approximation. However, this approximation is close enough to the actual value that you can treat it as being correct for the purpose of discussion."

Just a thought about this one, you should be able to get better results here by summing the logarithms, dividing by 5000 then exponentiating. It's the same reason people maximise the log likelihood rather than the likelihood for parametric distribution parameter estimation, it just lets the computer work with much more regular scaled numbers.