Includes resolutions even if no one corrected the market beforehand. (As long as the resolution was accurate.)

For example:

• A market that was at <=1% and then resolved YES.

• A market that was at >=99.5% and then was suddenly bet down to <=0.5%, and remains at <=0.5% for several days, or resolves to NO shortly afterwards.

Only counts binary markets. Does not include any market with less than 10 traders at the time of the swing. Does not include any market with a structure designed to make this happen. (Like the creator offering to pay back all the traders or something like that.) Does not include someone spiking the market and it being corrected shortly afterwards. Does not count incorrect resolutions. Does not include any swings that occurrred before this market was created. (Though if this has already happened I'd still like to know about it.)

Markets on various swing sizes:

• 95%

• 98%

• 99%