Sir Donald Bradman is not overrated. He averaged one hundred (okay, don't get pedantic now!) which is incredible. You assemble the worst bowlers today and ask today's best batsmen to play against them and I doubt they'd average much more than 70-80. You could make a good case for him being the greatest sportsman ever (Wood, Federer, Jordan, Pele, Ali etc being the other contenders).
HOWEVER, I do have to ask the question : can the fact that he was so much better than his peers alone mean that he would have been able to maintain that average if he'd played in the modern era?
Let's take an example. A,B,C take a test on numbers. A scores 100/100, B scores 50/100 and C scores 30/100.
D, E, F take a test on calculus. D scores 50/100, E scores 52/100, F scores 55/100.
Can we automatically conclude that A is better at Math than D, E and F? I am not so sure. Of course, it is a bit disingenuous to insinuate that the difference in difficulty between Bradman's era and the current era is the same as the difference between a simple numbers test and a hardcore calculus test, but it is worth a thought. The thing is, comparisons in the same era can be extremely difficult and meaningless, therefore comparing across eras is more often than not on the verge of impossible.
I would also like to ask my fellow PPers this:
George Lohmann picked up 112 wickets@10.75 in tests. I am not making this up I assure you, you can check it here :
http://www.espncricinfo.com/england/content/player/16337.html
Does this make him a better test bowler than say, Imran Khan or Malcolm Marshall? Half the average roughly, which is roughly the same advantage Bradman holds over the Laras and Tendulkars. It would be interesting if longevity, difference in eras, etc. are used because you could make the exact same arguments for favouring a Tendulkar or a Lara over Bradman.
NB : I am playing devil's advocate here. Whilst I don't think Bradman would average 99 in today's era, I think he'd still be the best batsman by a fair distance.