I was thinking that we could do polls on this site to try out various cardinal PR methods with real data. The first might be Sequentially Spent Score (with the highest-scored candidate for each round winning), but we could also branch out to other methods. Focusing on one method at a time is more likely to allow us to test various strategies specific to it, so we probably shouldn’t use the same set of votes to simulate various methods. We could solicit two sets of input - honest votes, and strategic votes, to see how the method performs under either and compare them. If this sounds like a good idea, what question(s) and answer(s) should be in the poll, and when/how should we do it? I was thinking maybe something slightly controversial, to solicit disagreeing opinions, so maybe favorite car brand, and then we ask voters to pretend they are hyperpartisan and want to game the system to maximize the number of cars of their preference in the final winner set. Some options might then be “Ford, Lexus, Tesla, etc.” There should probably be only about 3 to 5 winners max, because this site tends to only get up to ~10 votes on polls. What do you think?