It seems to make up an arbitrary example (or how did it ever get that graph?) and then goes on to say
Quote
In practice it’s worse than this, because this is assuming everyone is investigating exactly the same question.
I don't really follow this. What if many studies are more like the red curve?
(http://i.imgur.com/27UV20q.png)
What about the uncertainties that are usually presented by the individual studies themselves? The studies themselves would be curves like that representing each question with uncertainties depending on the method. But sure some of the things mentioned are legit problems.