Atheism as the null hypothesis.

[Simon Moon]

Atheism is not the default position. It is the agnostic's position that should be the default position.

1. There is no God.

This is either true or false. Atheist believe this to be true. Agnostics don't know. Theist believe it to be false. Verificationists think the statement meaningless. Even if evidence and argumentation fail at proving there is a God, that does not mean there is no God. To say there is no God is a claim to knowledge and must be justified.

If you are defining atheism as an absence of belief in God (as I believe many of you like to do), that would only be a psychologically state. In that case, "absence of a belief in God" cannot be the default position because that would include a wide range of views: traditional atheists, agnostics, and verificationists and therefore incoherent.

Agnosticism is, of course, a statement about knowledge; atheism is a statement about belief, or, rather, the lack, positive or negative, of a particular belief.

Your statement above is quite correct.

A true agnostic would regard the existence of god, any god, to be a 50/50 proposition.

This is not correct.

There is nothing within any definition of agnosticism that requires any sort of probability calculation. Nor, is there really any way to calculate such a probability. 

To calculate probabilities, data points are required. I'd be very interested to know where your hypothetical agnostic acquired their data points. As in: how many universes did they study in order to determine how many of them have gods vs how many do not? 

The classic, formal definition of agnostic is, "a person that does not know, and/or believes it may be unknowable, if a god exists".