See eg https://old.reddit.com/r/slatestarcodex/comments/13n6v72/did_scientists_accidentally_invent_an/

Resolves PROB to my credence in a year (which is currently, like, 5% bc base rate popular medical claims). But I'll look into it more, and will strongly take into account commenters' arguments.

This also, obviously, depends on the meaning of 'effective'. Here, I'll mean it in a sense of 'has a detectable and significant effect', as opposed to 'is an effective pharmaceutical'. In particular, if it works but tolerizes too quickly to be practically applicable, works but has too strong side-effects to be practically applicable, or something like that, that doesn't affect the resolution. If a (neither will happen but example) RCT shows it reduces some measure of addiction by .1% with 95%CI .095 - .105, this resolves NO - if it reduces it by (say) 15%, this resolves YES. That said, smaller effects are harder to detect and demonstrate so the 15% case is more likely to, in practice, resolve no

During the first day, I may change the resolution criteria to be: "Resolves to poll in a year", if commenters prefer that. But while poll resolutions work for things that are of general interest + are easy to understand, this doesn't seem to satisfy either criteria particularly well.

Less confident this market is 'good' than my other similar markets.