If 58% of 58% is 33.64% as indicated in the first post (corrected to remove the roundoff), then in the newspeak of recent technical reporters, since the difference between 58% and 33.64% is 24.36%, 58% is actually 124.36% of 38% which is 41.834%. So with 58% actually being only 41.834% it is obvious that 58% of 58% is actually only 58% of 41.834% which is a mere 24.26372%, unless we assume that both of the 58%s are really 41.834%s in which case 58% of 58% is about 17.5008%.

And I know that this is an accurate analysis because I made it up myself just now. Confidence is additionally proven by the fact that nearly all (98% at least) news reporters I've read recently obvously use this kind of pure and unsullied logic.

The nice thing about statistics is that almost any SPGWK can come up with any answer wanted, without actually knowing anything at all - about anything at all.

John