3 ms·
> Considering only the answers with the high-confidence R3 rating, we found that the probability of providing the correct response was significantly elevated co
by srtjstjsj 10y ago
> Considering only the answers with the high-confidence R3 rating, we found that the probability of providing the correct response was significantly elevated compared with all responses (0.60±0.03, P=0.0010),
For "high-confidence", that's a massive false-positive rate.
- frozenport 10y agoNice catch, these error bars are enough to call bullshit on the whole thing.
- Noseshine 10y agoYou misunderstand and subsequently misrepresent the whole experiment and paper. Nobody but headliners (people only reading the headline) and overexcitable readers even tries to make this about "we can reliably detect single photon events". Yes that's what the headline seems to say. The detection is clearly much better than chance. This is about exploring the absolute lower limit of human perception. It is not about "we can 'see' single photons", even though that's not even wrong - we sort-of can, the experiment shows we have a chance higher than random. On that low level it isn't about absolutes but about statistics. I'm really disappointed in the comments here, the first time I saw this - was it on HN too not that many days ago? - comments were a lot better. Do't use a tabloid-paper interpretation and do read the actual paper too, as always headlines are useless.