Why did Einstein change his prediction about bended light?

[Bjarne]

When Albert Einstein published the general theory of relativity he predicted that light would bend.

But before the measurement experiment was executed he changes that prediction. The last prediction was a factor 2 larger.

So far I remember he claimed that this was because also the whole Universe bends also bends (which I find hard to believe).

What was really the reason to that he changed his prediction, (he could not know that the whole Universe would cause exactly the double effects) – there must have been mathematical reason that had force him to change his prediction, - but which?

 

[bcrowell]

It's true that the universe has curved spacetime (if that's what you meant by "bends"), but that's on a cosmological scale, and it has no effect on solar-system tests of general relativity.

Einstein's original prediction was made after he'd published special relativity, but before he developed general relativity. It's simply the prediction you get if you use Newtonian mechanics and treat a ray of light as a particle. After he figured out general relativity, which is the relativistic theory of gravity, he figured out that the prediction should be twice as big.

 

[Bjarne]

After he figured out general relativity, which is the relativistic theory of gravity, he figured out that the prediction should be twice as big.

But why did he do that?
Which equations did show he should?

 

[Mentz114]

There is a good explanation here.

http://www.mathpages.com/rr/s6-03/6-03.htm