Utility: Difference between revisions
Content deleted Content added
Psephomancy (talk | contribs) a note from talking with Effective Altruists yesterday |
No edit summary |
||
Line 1: | Line 1: | ||
Utility broadly means satisfaction. |
Utility broadly means satisfaction. |
||
Though often used in voting theory to refer to cardinal utility i.e. [[Rated method|rated method]] utility, it can also be used for discussing ordinal utility i.e. ranked-preference utility. |
Though often used in voting theory to refer to cardinal utility i.e. [[Rated method|rated method]] utility, it can also be used for discussing ordinal utility i.e. ranked-preference utility. (The [[Borda count]] uses ranked ballots to find what this article would describe as rated utilities). |
||
An example would be (using a [[preference-approval]]): |
An example would be (using a [[preference-approval]]): |
||
Line 26: | Line 26: | ||
There are two ways to derive ranked ballots using ordinal utility. The first is for a voter to ask themselves "who are the candidates I would want to win if I could choose the winner myself?" This is equivalent to asking who you would honestly vote for in [[FPTP]], and it shows who your 1st choice(s) are. If you then remove them from consideration and repeat the question, you find your 2nd choices, etc. The second way is for a voter to ask themselves, for every possible [[head-to-head matchup]], who they'd prefer. The [[Copeland]] ranking shows the voter's ranking of the candidates. This is arguably one way to justify [[Smith-efficient]] [[Condorcet methods]]: if, for an individual voter, the best candidate(s) are the ones from the smallest group that win all head-to-head matchups against all other candidates based only on that voter's judgment, then why not for society? Similar reasoning shows why [[Score voting]] can be justified using rated utilities in head-to-head matchups to quantify harm or benefit done to the voter. |
There are two ways to derive ranked ballots using ordinal utility. The first is for a voter to ask themselves "who are the candidates I would want to win if I could choose the winner myself?" This is equivalent to asking who you would honestly vote for in [[FPTP]], and it shows who your 1st choice(s) are. If you then remove them from consideration and repeat the question, you find your 2nd choices, etc. The second way is for a voter to ask themselves, for every possible [[head-to-head matchup]], who they'd prefer. The [[Copeland]] ranking shows the voter's ranking of the candidates. This is arguably one way to justify [[Smith-efficient]] [[Condorcet methods]]: if, for an individual voter, the best candidate(s) are the ones from the smallest group that win all head-to-head matchups against all other candidates based only on that voter's judgment, then why not for society? Similar reasoning shows why [[Score voting]] can be justified using rated utilities in head-to-head matchups to quantify harm or benefit done to the voter. |
||
One notable contrast between ordinal and cardinal utility is that with ordinal, one voter can shift their preference to make a good majority-preferred candidate become a bad minority-preferred candidate, whereas with cardinal utility, there is a degree of damage i.e. it is not too bad a thing to elect a candidate with slightly less utility than the utilitarian winner. |
|||
== Utility vs utility == |
== Utility vs utility == |