Another interesting topic! Obviously there is quite a difference between off the shelf and custom.
One thing that I have noticed on here, is that when guys have visited the maker, and with his help selected a cleft (largely following plenty of mallet work) the bat that is eventually produced is rarely blemish free, and might not be top grade on looks. However, the customer is largely satisfied, as he has been involved in the selection process etc. and has picked his blade mainly on performance with looks coming second. - How would the bat be graded?
Apart from this, I kind of think that the whole business of grading is a bit skewed.
For example, if a manufacturer PURELY grades on performance, who is going to buy a g3 or lower bat? Surely everyone wants a bat that will 'go'?
Overall I think it is safer to grade on pure cosmetics - especially as the eventual performance of a bat can be open to negotiation for quite some time. Even when buying from the cleft stage, it is potential performance that is tested not actual.