Maybe it's just my relatively ancient RS voltstick but I find I have to hold it just so at the back end of the handle and hold the translucent probe bit just so in relation to what I'm testing to get a reliable indication. Thus I calibrate it by holding it close to a known live and adjusting my hold to get it to light up then, without changing my grip, move over to the conductor I'm testing.
(Then check with the neon stick and flick of the finger.)
It should detect higher impedance leakage than necessary to light even a neon (since the volt stick measures almost pure electric field, through air + its plastic cap + whatever insulation is on the cable etc). However there may well be situations where there's a live conductor whose field is shielded by a grounded conductor: I suppose a hypothetical example would be coax with mains on the centre & ground on the braid. (I'm sure you'd never find anything like that in the real world, of course :-)
John S