My own wild guess is that perhaps the two-full-dose protocol raised too many antibodies to the adenovirus vector itself, and made the second dose less effective. This has always been a concern with the viral-vector idea.
It sounds like there was a medium-sized prior that the lower first dose would be better—why else would they have tested it?
I was confused as to why they did this too—alternative guesses I had were to increase number of available doses or to decrease side effect severity.
However the site you link to has been updated with a link to Reuters who quote AstraZeneca saying it was an accident—they miscalculated and only noticed when side effects were smaller than predicted.
Derek Lowe writes:
It sounds like there was a medium-sized prior that the lower first dose would be better—why else would they have tested it?
I was confused as to why they did this too—alternative guesses I had were to increase number of available doses or to decrease side effect severity.
However the site you link to has been updated with a link to Reuters who quote AstraZeneca saying it was an accident—they miscalculated and only noticed when side effects were smaller than predicted.
I guess the lesson being: Never attribute to n-dimensional chess that which can equally be attributed to stupidity.