Every registered study in the order it ran, with the result it gave at the time. Where a later study improved on an earlier one, both are here.
Infrastructure names are redacted in a few protocol files; the protocol hash recorded in each result refers to the unredacted original.