An altruistic wager is a type of argument that seeks to establish that an agent should act as though a conclusion is true because, given what we know, that seems to be the right decision from an ethical perspective, even if the conclusion does not actually seem likely to be true. For example, one might argue that one should act as though animals are sentient and are moral patients when deciding whether to be a vegetarian, as long as that has any nontrivial chance of being true, because if it is true the ethical harms of eating meat might very much outweigh the ethical or self-interested benefits.
Altruistic wagers often assume an agent should behave according to expected value theory, but some such wagers may also hold under some alternatives to expected value theory. Some wagers may even hold more strongly under particular alternatives to expected value theory, such as a theory that incorporates risk-aversion.
Further reading
Baumann, Tobias (2020) An overview of wagers for reducing future suffering, Reducing Risks of Future Suffering, January 31.
Related entries
alternatives to expected value theory | decision theory | decision-theoretic uncertainty | expected value theory | fanaticism | moral uncertainty | risk aversion
In future, it’d probably be good to expand this entry to mention Pascal’s wager (the original form, not the mugging) as perhaps the most iconic “wager” and as having a similar structure, though not being altruistic.
My impression is that these are usually/often more about establishing that you should act as if the proposition is true, rather than that you should conclude the proposition genuinely is true? I could be wrong, though.
Ok, I googled this full term, and it seemed that this page coined the term. That suggests that how to define it is up for grabs. And the definition my comment suggests seems more useful and to better match how I’ve heard the term “wager” used in EA. So I’ve changed the first sentence and added more text fleshing out the idea.
But maybe the original definition was more useful for some reason—I’m open to people editing it back or debating it here.