Granted BOTH TPGs have some serious problems when it comes to variety attribution. Pretty much anybody who has submitted more than a couple varieties for attribution will attest that MANY mistakes are made. There have been many threads on the PCGS forums detailing the problems with variety attribution at PCGS. I don't read NGC forum so am unsure whether or not there has been much discussion re over there.
I seldom submit coins to NGC, but because the certified populations for many varieties are very low at NGC I decided to try to cross over a PCGS attributed variety. The Franklin half 1957-D/D FS-501 has a FBL population of 95 at PCGS, whereas the NGC pop for same is 16. It stands to reason that an NGC example could sell for more than a PCGS example simply because there are far fewer available.
So, greedy me decides to send an example for crossover, already properly attributed as FS-501 in a PCGS slab. To my great surprise, the coin returned from NGC with notation "not an RPM, machine doubling only". The fact that NGC's variety team failed to properly attribute the variety is bad enough.
But the real icing on the cake is this: NGC proceeded to cross the coin over into an NGC holder sans the variety designation. That is totally unethical. PCGS would never do that. PCGS procedure in a case such as this is to contact the submitter and explain that in PCGS's opinion the coin is not the requested variety, and therefore, they need permission to cross the coin without the variety designation.