User talk:Matijaskala/Probabilistic Approval Voting
Trying to understand this
- P: A → B means the function P maps from set A to set B
- C×C is the cartesian product of set C with itself, meaning that the domain of the function is all ordered pairs of candidates
- C = the set of candidates (not the set of complex numbers)
- [0, 1] is the set of real numbers between 0 and 1, inclusive
- W = the set of candidates, again?
- Since C×C includes (A, A), then the "expected number of elected candidates belonging to the same faction as A" includes A? And so is always at least 1?
- What is "expected average load"?
- What is V? Number of voters who approve of a given candidate?
- = (number of voters who approve of A and B) / (number who approve of A or B or both)?
- Should link "Hare quota" to an article that defines it
- What is "universal liking condition"?
- This system is deterministic, correct? Even though it says "probabilistic" in the name?
- It would really help to have a worked-out example
— Psephomancy (talk) 05:51, 19 February 2020 (UTC)
- User:Matijaskala and User:Dr. Edmonds, I was trying to understand this here before looking at the Python code. — Psephomancy (talk) 22:05, 25 February 2020 (UTC)
- W = the set of already elected winners
- "expected average load" is the expected number of elected allies divided by the number of voters. Perhaps I should call this "voter efficiency". — Matijaskala (talk) 06:03, 27 February 2020 (UTC)
- I think in general, it would be helpful to indicate, for any single-winner voting method, how it tends to work in the 3-candidate case, and for any PR method, how it works in the party list case. I did what I'd consider a decent example of this for Sequentially Shrinking Quota#Notes. BetterVotingAdvocacy (talk) 20:56, 3 April 2020 (UTC)