14 ms·
This set of simulations ( explanations: https://docs.google.com/document/d/1dzt-q6Gb8PHAgkLBv2j7UhiijbWGSaQeoeYy16YOt4Y/edit https://docs.google.com/document/d/
by jpfed 10y ago
This set of simulations ( explanations: https://docs.google.com/document/d/1dzt-q6Gb8PHAgkLBv2j7UhiijbWGSaQeoeYy16YOt4Y/edit https://docs.google.com/document/d/1dzt-q6Gb8PHAgkLBv2j7Uhii... , graphs: http://rpubs.com/Jameson-Quinn/vse3 http://rpubs.com/Jameson-Quinn/vse3) shows that basically anything is better than FPTP.
(It also shows that IRV is only marginally better; while "worse is better" sets my teeth on edge, if IRV is the doable option I'll take it rather than wait for people to come around on score methods).
- dragonwriter 10y agoScore methods are worse for a reason that doesn't show up in most simulations, because mapping of expected utility to an an externalized score isn't consistent between individuals, but instead highly variable (both among individuals within groups and, particularly, along ethnic and other cultural lines), which has been shown in many studies of score-based rating systems.in many domains. Most simulations assume consistent mapping from preferences to ballot markings, which is reasonable for ranked, but not score, voting.
- karmelapple 10y agoHave links to any of those studies? I am unfamiliar with the terminology but would like to learn more.
- iainmerrick 10y agoExactly, score voting is just as vulnerable to tactical voting as any other method for that reason. For example, if the scoring range is 1-10, it's better for me to use the full range even if none of the options on offer is really a 1 or 10. Using the full range gives my opinion more weight in the result. I strongly suspect score voting tends towards ranked voting when people vote tactically, which they always will, just with extra noise and randomness. (I don't know of any studies on this, though.)