This was inspired by some astrological charts comparing the horoscopes for famous victims of assassination. It occurred to me that now that we have huge amounts of data via the Internet and massive data mining power, people have looked for correlations between just about everything. I was wondering if any have been found that are so ridiculously impossible in terms of causation that they would have to amount to divination- yet are eerily on the mark?
How on the mark do you require for something to be eerie? After all, if you check 1000 randomly-picked pairs of data sets, you’d expect to find one that appeared to be significant to a 99.9% confidence level. Does that make that random one you find “eerie”?
YMMV, but I would use the rough and ready standard of “an apparently random correlation that has continued since it was originally pointed out”. 