Resolves YES if by the end of 2023 it is possible to undo or change a market resolution. Otherwise NO.

For example, if the author or Manifold admins can change the resolution to a different one, or change the market from resolved back to open, that counts as YES.

Any capability to undo (even if it's just a few seconds) will count for the purposes of this market.

People accidentally resolve markets wrong by mistake sometimes - it's nearly impossible to avoid. Mistakes can include accidentally hitting the wrong button, or resolving based on a misinterpretation of news/data, or resolving based on a news/data that turns out to be wrong and is later corrected.

I think it's important for Manifold to be able to fix resolutions that were honest mistakes. This could be done by supporting undoing a market resolution - you could undo all of the payouts made by the incorrect resolution (undoing transactions is not a new concept, it's similar to how N/A undos all of the trades). Then the author can resolve to the correct resolution, or the market can stay open if it wasn't supposed to be resolved yet.

This could be subject to reasonable limitations - e.g. can only fix the resolution within some grace period, or perhaps it could require approval by the admins or other parties.

Even a hypothetical 100% mistake-free author would benefit from being able to change resolutions. Today, authors often have to wait for weird, very unlikely edge cases in order to resolve, or wait some time for anyone to raise unexpected counterarguments to the resolution. This delays resolutions unnecessarily which is bad because it locks up people's investments, which means people will be less likely to want to trade and probabilities will be less accurate. Once modifying a resolution is possible, they can resolve earlier and then modify it if needed.

For example, markets on who will win an election are typically resolved when a decision desk calls it, but these calls are typically incorrect about 0.1% of the time - which means statistically it's fairly likely for at least one election market to end up resolved wrong. But waiting for the election results to be officially certified would introduce a huge delay. Similar concerns apply to many other questions, and being able to change resolutions would greatly reduce these sorts of unnecessary delays in resolutions.

Related question on whether it will be possible to override the author's resolution without their agreement, to address intentional misresolutions in addition to mistakes: