GLOSSARY FOR VOTING ENTHUSIASTS

Have edits or additions for this document? Input your suggestions on the google doc

 

Approval Voting: A voting system where voters check a box for as many candidates as you approve. The candidate with the most approval votes wins.

 

Arrow’s theorem: Arrow’s theorem states that it’s impossible for any voting system to pass all desirable criteria because some criteria are mutually exclusive.

 

Bayesian Regret (BR): A system used to measure the accuracy (utility) of various voting systems.

Bayesian Regret measures the amount of “avoidable human unhappiness" at the end of an election. This is the inverse of Voter Satisfaction Efficiency. Bayesian Regret is measured quantitatively by computer simulations and it can be used to measure election accuracy regardless of voters strategies or honesty. Bayesian Regret and Voter Satisfaction Efficiency are considered by many election experts to be the gold standard for measuring accuracy (utility) of various voting systems, though some people prefer to use the Condorcet Winner as a measure of accuracy. Condorcet and BE/VSE usually agree on the best winner for a given election but in some close elections Condorcet favors the majority's 1st choice and BE/VSE prefers the candidate that will make the most people as happy as possible, i.e. the best compromise.

 

Borda Count: a voting system in which voters rank options or candidates in order of preference. These rankings are then counted as scores. The scores are totaled and the candidate with the lowest score wins. Variations which use other formulas to turn rankings into scores exist as well. Borda count is often described as a consensus-based voting system rather than a majoritarian one.

 

Bullet Voting: a voting strategy that can be used in more expressive voting systems where you give only one candidate a maximum and the rest a minimum vote. This is a common strategy for less informed or more hurried voters. Ranking only your first choice and leaving others blank. Bullet Voting when you honestly either love, or hate, or have no opinion on the candidates is an honest strategy that is not harmful. Bullet Voting when you have a nuanced opinion or like multiple candidates is a harmful strategy that can lead to less accurate results. An election where everyone Bullet Votes would be Plurality.

 

Compromise Acceptance Criteria: aka Favorite Betrayal Criteria. A voting system criteria that states that a voting system should pick the candidate that will make the most voters as happy as possible. Compromise Acceptance Criteria means that a voting system should take into account a voters second choice and other preferences even if doing so means that their favorite is less likely to win.

Compromise Acceptance Criteria is the opposite of Later-No-Harm Criteria, (aka Compromise Refusal Criteria.) The two criteria are both desirable but are mutually exclusive. It's debatable which is preferable and it's fine to agree to disagree on this point. This paradoxical criteria is the subject of much debate and the conclusion is a matter of personal preference.

The fundamental question is this:
"Do you prefer to work together to find what's best overall, or do you prefer to stick with what is best for you?"

 

Condorcet Criteria: A criteria which states that the Condorcet Winner should always win.

 

Condorcet Methods: Condorcet Methods are any voting system where the Condorcet Winner always wins. More common examples are Ranked Pairs and Schultz Method but there are many more that have different methods for resolving ties.

 

Condorcet Winner: The Condorcet Winner is the candidate that would beat any other candidate in a head-to-head election. Majority rules. The Condorcet Winner is often used to show when a spoiler effect occurred and elected a candidate that was not preferred by the majority. Some voting systems are Condorcet Methods, meaning that the Condorcet Winner always wins. It is important to note that in some elections a majority winner doesn't exist because there may be a 3 way tie, (like in rock, paper, scissors) there may be no candidate that the majority supports, or there may be multiple candidates with support from different majorities (picture a venn diagram).

Though some people prefer to use the Condorcet Winner as a measure of accuracy, many elections experts prefer Bayesian Regret and Voter Satisfaction Efficiency. Condorcet and BE/VSE usually agree on the best winner for a given election but in some close elections Condorcet favors the majority's 1st choice and BE/VSE prefers the candidate that will make the most people as happy as possible, i.e. the best compromise.

A Condorcet Winner is only as accurate as the information on voters ballots so it cannot take into account dishonest or strategic voting. A Condorcet Winner derived from a more expressive and more honest ballots will give a more accurate result. A score ballot derived Condorcet Winner would be quite accurate.

 

Favorite Betrayal: is a dishonest strategy where in order to prevent a spoiler effect a savvy voter realizes that they are better off ranking their favorite candidate as a number 2 or lower in IRV or just not voting for them at all in Plurality. In score voting Favorite Betrayal would be any time you give your favorite a score that is less than you gave to others. This is the most harmful strategy in voting.

 

Favorite Betrayal Criteria: is a criteria used to judge voting systems. In order to pass, a voting system must never create a situation where this is a good strategy for a voter, or where it helps them get a better outcome overall.

  • Plurality strongly incentivizes Favorite Betrayal for all voters whose favorite isn't polling in the top 2.  
  • IRV incentivizes Favorite Betrayal for voters whose favorite is pretty strong but is likely to be eliminated in a later round, i.e. those whose favorite will probably come in 2nd or 3rd place but not 1st.
  • STAR doesn't incentivize Favorite Betrayal though it could work in extremely rare scenarios. Doing it effectively would require knowing who was ahead in a dead tie between the 2nd and 3rd highest scoring candidates.  
  • Score Voting Passes Honest Favorite Criteria criteria but is more vulnerable to other strategies.    

Because the term Favorite Betrayal Criteria incorrectly makes it sound like we want people to betray their favorite, some people prefer the term “Honest Favorite Criteria” or “Compromise Acceptance Criteria.” When talking with a technical audience Favorite Betrayal may be the most well understood term, but for a general audience we want terms to be as self explanatory as possible in order to avoid confusion. All criteria would ideally be positive and describe the desired effect.

 

Favorite Betrayal Effect: is an effect in elections where people practice the Favorite Betrayal Strategy. This can lead to the actual favorite of the majority losing and the people having no idea that the true favorite even had the needed level of support. (Favorite Betrayal caused Spoiler Effect). Favorite Betrayal Effect can also cause viable parties to seem unviable and other negative consequences, even when the candidates in question didn't have the support required to actually win.

 

Favorite Betrayal Criteria: is a criteria used to judge voting systems. In order to pass, a voting system must never create a situation where this is a good strategy for a voter, or where it helps them get a better outcome overall.

 

First-Past-The-Post: The current system in the USA for voting. Otherwise known as Plurality Voting. Widely believed to be the least accurate voting system, but it is one of the oldest and most simple.

 

Honest Favorite Criteria: is the same as Favorite Betrayal Criteria. This criteria was renamed because Favorite Betrayal Criteria incorrectly makes it sound like we want people to betray their favorite. When talking with a technical audience Favorite Betrayal may be the most well understood term, but for a general audience we want terms to be as self explanatory as possible in order to avoid confusion. All criteria names would ideally be positive and describe the desired effect.

 

Instant Runoff Voting (IRV): the most commonly used type of Ranked Choice Voting. A voting system in which voters rank candidates are ranked on the ballot in order of 1st, 2nd, 3rd and sometimes 4th or more choices. It's fine to leave candidates blank. If a candidate has a majority of 1st choice votes, that candidate wins. Otherwise, the candidate with the fewest 1st choice votes is eliminated. If your 1st choice is eliminated, your vote goes to your next choice (if any or if they haven't already been eliminated in an earlier round) This process repeats until one candidate has a majority.

 

Later-No-Harm Criteria: a voting system criteria that states that you can never hurt your favorite by ranking or rating your other candidates honestly. This is obviously desirable, but unfortunately systems that pass this criterion are unable to help find a compromise candidate that would be preferred by the majority if that candidate isn't the majority's first choice. This paradoxical criteria is the subject of much debate and the conclusion is a matter of personal preference.

Later-No-Harm can also be called Compromise Refusal. The opposite is Compromise Acceptance Criteria. It's debatable which is preferable and it's fine to agree to disagree on this point.

The fundamental question is this:
"Do you prefer to work together to find what's best overall, or do you prefer to stick with what is best for you?"

 

Majority Criteria: a controversial criteria that states that a slim majority that agrees on its first choice should be able to overpower another larger majority that may not agree on their first choice but have come to a compromise. This is the inverse of Tyranny-Of-The-Majority Criteria. Majority Criteria and Tyranny-Of-The-Majority Criteria are mutually exclusive.

 

Nader Effect: an effect where a majority coalition can be defeated by a much smaller minority because the coalition's votes are split between two similar candidates, usually from the same side of the political spectrum or from the same party. This is also known as vote splitting. In 2000, Green Candidate Ralph Nader ran in the general election alongside Al Gore. George Bush Sr. was elected and Nader was blamed for being a spoiler though this is actually debatable for a few reasons. Regardless, the experience had a lasting effect on voters who still remain very reluctant to vote 3rd party for fear of the vote being split.

 

Plurality Voting: The current system in the USA for voting. Otherwise known as First-Past-The-Post. Widely believed to be the least accurate voting system, but it is one of the oldest and most simple.

 

Proportional Representation: The idea that representation in government should be in proportion to the demographics of the population. This term usually relates to the makeup of political bodies and councils. Proportional Representation is also a family of voting systems designed to elect multiple winners that represent both majority and minority interests. Versions include STV, RRV, Party List Systems and MMP among others.

 

Range Voting: a voting system for single-seat elections, in which voters give each candidate a score, the scores are added (or averaged), and the candidate with the highest total is elected. Otherwise known as Score Voting.

 

Ranked Choice Voting (RCV): a family of voting systems that use rank ordering of candidates to determine a winner. Pure Ranked Choice Voting Systems include Instant Runoff Voting, Ranked Pairs and other systems that use a ranked ballot and also use the ranked order to determine the winner with varying algorithms. Hybrid RCV systems include STAR and Smith Score, which use a score ballot but then uses the derived rankings from the ballot to determine the winner, and systems like Borda Count, Bucklin, or 3-2-1 Voting that use a ranked ballot but determine the winner by other means than ranking.

Some people, especially in the USA use RCV to mean Instant Runoff Voting (IRV), the most commonly used type of RCV.

 

Ranked Pairs: A Condorcet Voting System which starts with the strongest defeats and uses as much information as it can without creating ambiguity. Also known as the Tideman method. 

 

Reweighted Range Voting (RRV): a proportional representation voting system for electing multiple candidates at once. RRV is based on Score Voting and uses a series of algorithms to reweight votes and allow voters to show support for minority candidates. The goal of RRV is to have the winners match the voter demographics.

 

Schulze method: A Condorcet Voting System which repeatedly removes the weakest defeat until ambiguity is removed.

 

Score Runoff Voting (SRV): A voting method invented in 2014 which topping the charts for accuracy as measured by Voter Satisfaction Efficiency. SRV is a hybrid of Score and Instant Runoff Voting. Voters give a score to each candidate on a scale from 0-5 (other scales can be used as well) The two highest scoring candidates advance to an instant runoff. In the runoff, your full vote goes to the one you scored higher. Score Runoff Voting is the same as STAR Voting if a 0-5 scale is used. 

 

Score Runoff Voting- Proportional Representation (SRV-PR): a proportional representation voting system for electing multiple candidates at once. SRV-PR is a proposal based on Reweighted Range Voting and Score Runoff Voting. SRV-PR uses a series of algorithms to re-weight votes and allow voters to show support for minority candidates. Each candidate is selected in two rounds like in SRV. 

 

Score voting: a voting method for single-seat elections, in which voters give each candidate a score, the scores are added (or averaged), and the candidate with the highest total is elected. Otherwise known as Range Voting

 

Single Transferable Vote (STV): a proportional representation voting system used in Australia for electing multiple candidates at once. STV is based on Instant Runoff Voting but uses a complex series of algorithms to reweight votes and allow voters to show support for minority candidates. The goal of STV is to have the winners match the voter demographics.

Social Utility Efficiency: Social utility efficiency is a measurement of the utilitarian performance of voting methods—how likely they are to elect the candidate who best represents the voters' preferences. It is also known as utilitarian efficiency, voter satisfaction index or voter satisfaction efficiency

 

Spoiler Effect: an effect where a majority coalition can be defeated by a much smaller minority because the coalition's votes are split between two similar candidates, usually from the same side of the political spectrum or from the same party. This is well known as the "Nader Effect" from when Green Candidate Ralph Nader ran in the general election alongside Al Gore. George Bush Sr. was elected and Nader was blamed for being a spoiler though this is actually debatable for a few reasons in Nader's case. Regardless, the experience had a lasting effect on voters who still remain very reluctant to vote 3rd party for fear of the vote being split.

 

STAR Voting (Score Then Automatic Runoff):  A voting method invented in 2014 which tops the charts for accuracy as measured by Voter Satisfaction Efficiency. STAR is a hybrid of Score and Instant Runoff Voting. Voters give a score to each candidate on a scale from 0-5. The two highest scoring candidates advance to an instant runoff. In the runoff, your vote goes to the finalist you scored higher. STAR is the same as Score Runoff VotingRV if a 0-5 scale is used. 

 

Strategic Voting: There are a number of strategies that can be used by voters to try and get the best results possible. There are 2 fundamental types of strategic voting, honest strategy and dishonest strategy. Note that in different voting systems honest and dishonest voting might be either helpful or harmful to the accuracy of the final results or for the individual voter. For example Plurality Voting strongly incentivizes dishonest strategy because it's best possible results are obtained when most voters are dishonest and strategic. A given strategy is deemed to be incentivized if the strategy will help the individual voter more often than it will backfire and harm them. It is impossible to eliminate strategic voting but some voting systems do a good job of making sure that strategic voting isn't incentivized, isn't beneficial, or that it isn't clear how or when to use a given strategy.

Dishonest strategy includes tactics like Favorite Betrayal, ranking or rating candidates in a different order than you actually prefer them, (Up Ranking or Down Ranking Strategy) Bullet Voting when you have a nuanced opinion, as well as multi voter strategic schemes like vote trading.

Honest strategies include voting your conscience as well as other strategies that would be considered honest voting in less expressive systems but require more thought in expressive voting systems. Examples of honest voting strategies include Bullet Voting when you honestly love or hate multiple candidates, and giving candidates higher or lower scores on a score ballot as long as you rank them in honest order (Up Scoring or Down Scoring.)

 

Supporting the Weak Opponent Strategy: A strategy where you help out some weak candidate you don't like in order to disadvantage a candidate you think can beat your favorite. Some voting systems can incentivize voters to support a weak opponent, in order that their preferred candidate win the general election. In an open partisan primary election, for example, this is a viable strategy. If your party candidate is a shoe-in to the general election, you may cast a vote for the weaker opponent candidate so that your candidate has an easier time winning.

 

Straw Man Argument: a deceptive debate tactic for when you can't effectively refute your opponents point so you instead refute a different point that nobody was making. This can make it look as though you won, when in fact you are off on a tangent.

 

Tactical Maximization is a strategy where you increase support for candidates other than your favorite because you think your favorite is weak or you want to hedge your bet. This strategy is the inverse of Tactical Minimization. In many elections both strategies could be used by different voters so the effects could cancel each other out. Vulnerability to Tactical Maximization and Tactical Minimization are the leading arguments against Score Voting though at worst this strategy would make Score Voting as or more accurate than Instant Runoff Voting or Approval Voting. This is otherwise known as Up Voting Strategy. Bullet voting is the extreme of this strategy.

 

Tactical Minimization is where you decrease your support for candidates other than your favorite. This strategy is the inverse of Tactical Maximization. In many elections both strategies could be used by different voters so the effects could cancel each other out. Vulnerability to Tactical Maximization and Tactical Minimization are the leading arguments against Score Voting though at worst this strategy would make Score Voting as or more accurate than Instant Runoff Voting or Approval Voting. This is otherwise known as Down Voting Strategy. Bullet voting is the extreme of this strategy.

 

Tyranny-Of-The-Majority Criteria: a controversial criteria that states that a slim majority that agrees on it's first choice should not be able to overpower another larger majority that may not agree on their first choice but have come to a compromise. This is the inverse of Majority Criteria and the two are mutually exclusive.

Tyranny-Of-The-Majority Criteria is the same as Compromise Acceptance Criteria. Ideally all criteria should positive and describe the desired effect to help so we prefer the term Compromise Acceptance Criteria when advocating this concept. Tyranny-Of-The-Majority Criteria is more well know in academic contexts and is effective when used to compare it with its inverse Majority Criteria.

 

Up Ranking or Down Ranking Strategy: this is a strategy that can be used in ranked voting systems where you give a candidate who is not your favorite or your least favorite a higher or lower ranking in order to help or hurt them based on what you think other voters are going to do. This is a dishonest and destructive strategy to the whole, even if it helps the individual voter.

 

Up Scoring or Down Scoring Strategy: this is a strategy that can be used in score voting systems where you give a candidate who is not your favorite or your least favorite a higher or lower score in order to help or hurt them based on what you think other voters are going to do. If your candidates can still be ranked in honest order of preference from looking at your ballot this is not a dishonest strategy and it's not harmful but if you actually change the derived rank order this can be a dishonest or destructive strategy to the whole, even if it helps the individual voter.

 

Up Voting or Down Voting Strategy: this is a strategy that can be used in score or rank voting systems where you give a candidate who is not your favorite or your least favorite a higher or lower vote in order to help or hurt them based on what you think other voters are going to do. Otherwise known as Tactical Minimization or Maximization. If your candidates can still be ranked in honest order of preference from looking at your ballot this is not a dishonest strategy and it's not harmful but if you actually change the derived rank order this can be a dishonest or destructive strategy to the whole, even if it helps the individual voter.

 

Voter Satisfaction Efficiency (VSE): VSE is a system used to measure the accuracy (utility) of various voting systems. This is a rebranding of the earlier term "Social Utility Efficiency.” VSE allows for a detailed look at how different systems perform in mathematical simulations. It can also measure how well each system functions using honest voting as opposed to various strategies. VSE can be used to predict how effective a given strategy is in each voting system and how often that strategy will backfire for the given voter.

In the field of voting theory, there are many desirable criteria a given voting method may or may not pass but it’s been shown that it’s impossible for a method to pass all desirable criteria (Arrow’s theorem), so trade-offs are necessary. VSE measures how well a method makes those tradeoffs by using outcomes. Basically, instead of asking “can a certain kind of problem ever happen?”, VSE is asking “how rarely do problems of all kinds happen?”.

VSE is expressed as a percentage. A voting method which could read voters minds and always pick the candidate that would lead to the highest average happiness would have a VSE of 100%. A method which picked a candidate completely at random would have a VSE of 0%.

Voter Satisfaction Efficiency is inversely related to Warren Smith’s Bayesian Regret models. VSE is explained here in depth: http://electology.github.io/vse-sim/VSE/

 

Vote Splitting: an effect where a majority coalition can be defeated by a much smaller minority because the coalition's votes are split between two similar candidates, usually from the same side of the political spectrum or from the same party. This is well known as the "Nader Effect" from when Green Candidate Ralph Nader ran in the general election alongside Al Gore. George Bush Sr. was elected and Nader was blamed for being a spoiler though this is actually debatable for a few reasons in Nader's case. Regardless, the experience had a lasting effect on voters who still remain very reluctant to vote 3rd party for fear of the vote being split.

 

3-2-1 Voting (321V): A new voting system that seems to be very accurate while still being quite simple. This system was invented by Jameson Quinn of electology.org. Voters rate each candidate using one of three ratings: Good, OK, or Bad. There are three rounds where candidates are pared down to the top 3, the top 2, and then the winner. The 3 semifinalists are the candidates with the most “good” ratings. (No two may be from the same party, and all must have at least 15% “good”.) The 2 finalists are the semi-finalists with the fewest “bad” ratings. The winner is the finalist rated higher on more ballots. If this system had a name that could be googled we’d be more excited about it.