Silviu Itescu The Legend, page-280

To pretend that this is amazing and a sure thing is delusion. If they had 20 predetermined subgroups then statistically one of them would incorrectly have a P <0.05.

few corrections (and to illustrate how overinflated your statement is. images courtesy of a prior post by stockrock)

from left to right (where statistically significant).
p-value = 0.019, thats a 1 out of 52 chance the null has been incorrectly rejected
p-value = 0,003, so thats a 1 out of 333 that would be incorrectly rejected


p value = 0.001, 1 in 1000...
p value = 0.035, 1 in 28...
p value = 0.0017, 1 in 58...


1 in 22..., 1 in 200...

funny how multiple of these sliced, diced and subdivided groups show such strong signals. so statistically multiple of these groups have incorrectly rejected the null? what are the odds? 1 in 20?