STAR Voting does a good job at both the Later No Harm and Favorite Betrayal criteria

In election science there are a number of desirable pass/fail criteria. In order to get a "pass", the system would need to pass the criteria 100% of the time. Some of them, like the "Favorite Betrayal" criterion and the "Later No Harm" criterion are mutually exclusive and it's been proven that it's impossible for any voting method to pass all desirable criteria 100% of the time. In voting reform competing priorities need to be balanced. 


Favorite Betrayal and Later No Harm assess strategic voting incentives for voters. 

1. Favorite Betrayal tells us if it's always safe to vote or your favorite. If so, it passes the Favorite Betrayal criterion. 

2. Later No Harm tells us if it's always safe to support other candidates beyond our favorite. If so, it passes Later No Harm. 

While it's impossible to pass both criteria 100%, it's absolutely possible to excel at both in practice. STAR Voting strongly incentivises honest and expressive voting, which has been clearly demonstrated in the peer reviewed literature.  

In contrast, most other voting methods badly fail one or both of these criteria.


Favorite Betrayal Performance: Our current system is terrible at Favorite Betrayal and the incentive to not vote for your favorite if you don't think they can win can be incredibly powerful. Ranked Choice Voting also fails this criterion, and ranking your true favorite first can backfire badly in any election with three or more competitive candidates. STAR Voting doesn't pass this criterion 100% of the time either but in practice, a voter should always give their favorite a full 5 stars.

Statistically speaking, STAR Voting gets an A in this category. The only scenario it might not pass is edge cases where the election is essentially a dead tie and where predicting the scenario and leveraging it strategically would be impossible.


Later No Harm Performance: The second criteria, Later No Harm only applies to voting methods where voters can support multiple candidates. Ranked Choice Voting passes this criteria, but doing so requires the method to ignore all down ballot rankings until after their first choice is eliminated, even if counting them might have helped that voter. STAR Voting counts all of a voters ballot data, so it does not pass this criteria, but it does accomplish the goal behind the criteria. In STAR Voting, studies show that voters are strongly incentivised to not only score their honest favorite 5 stars, but also to show their honest preference order between all relevant candidates. Doing so ensures that no matter who the finalists are in a given race, the voters' vote will go to the finalist they prefer. 

Statistically speaking, STAR Voting gets an A in this category too. The scenario it might not pass is when a voter may want to give a zero, not a 1, to a strongly disliked candidate, particularly if other strongly disliked candidates are essentially irrelevant anyway. In this case, the vote would still be considered a "viability aware" honest vote. 


A look at the data: The paper "STAR Voting, Equality of Voice, and Voter Satisfaction: Considerations for Voting Reform", which was peer reviewed and published in Constitutional Political Economy in 2023 takes a close look at the strategic incentives covered by Favorite Betrayal and Later No Harm. As you can see in the chart below, Favorite Betrayal, Burial, and Bullet Voting are all strongly disincentivised in STAR Voting. 


Later No Harm and the Spoiler Effect: Another mind blowing fact is that it's impossible to pass the Later No Harm Criteria 100% of the time and also eliminate vote-splitting and the Spoiler Effect. In order to eliminate the Spoiler Effect a voting method can't ignore ballot data that could be relevant. In order to pass Later No Harm a voting method has to ignore that additional ballot data, even when it's highly relevant. Selectively ignoring some ballot data comes with a number of serious potential repercussions that can and do hurt a voting methods accuracy, corrupt voter intent, and lead to voter disenfranchisement issues.


Takeaways: Some people are concerned that because STAR Voting doesn't pass Later No Harm that voters wouldn't have any incentive to score candidates beyond their favorite (known as bullet voting.) That's simply not true. Again, the STAR Voting runoff specifically encourages voters to show their honest preferences to ensure that no matter who the finalists are, their vote will go to the finalists they prefer. STAR Voting also strongly incentivises voters to give their honest favorite a full 5 stars. 

Groups like FairVote (and other Ranked Choice lobbyists) who try to weaponize the science by cherry picking criteria their proposal satisfies while obscuring the existence of criteria that Ranked Choice fails, do so disingenuously. In the real world, if a proposal got 99% on two competing metrics we would not call that "failing". 


FAQ: Does STAR Voting fail the Later No Harm criterion?

Learn more about Favorite Betrayal, Later No Harm and the Spoiler Effect in this ground breaking article

Peer Review: Constitutional Political Economy. STAR Voting, Equality of Voice, and Voter Satisfaction: Considerations for Voting Reform