If you do update your p value that much through that experiment you might want to do your own experiment.
Ask DavidM for the contact addresses of those people who succeeded and draft them for two possible next round. Do one test to satisfy your own curiosity and to see whether it sort of works. If it does, you come back and we decide based on your data on an exact protocol for the next round.
After the protocol for that round is stable, the image get’s picked by highly trusted LW person like Scott Alexander or gwern.
If that works we do more math and try to make the experiment powered enough to pass Randi’s 10^{-6} standard.
If you do update your p value that much through that experiment you might want to do your own experiment.
Ask DavidM for the contact addresses of those people who succeeded and draft them for two possible next round. Do one test to satisfy your own curiosity and to see whether it sort of works. If it does, you come back and we decide based on your data on an exact protocol for the next round.
After the protocol for that round is stable, the image get’s picked by highly trusted LW person like Scott Alexander or gwern.
If that works we do more math and try to make the experiment powered enough to pass Randi’s 10^{-6} standard.