On the fraud issue:
Mobile polling in army centres with little oversight.
I would definitely not rule out some fraud there, or some intimidation, etc. The problem is, these don't add up to a 63% against 34%.
Again, it's a bit like Greg Palast's case that the 2004 Bush reelection was stolen. He
does find some real instances of wrongdoing. But by any conventional mathematics, they don't add up as having changed the overall outcome of the election.
Counting possibly suspiciously fast.
I'd like to chip in on that one. First, the questions of how fast the counting really was, seems to have been the subject of some whispering rumours. First I heard something about an hour or so. Now somebody in this thread quoted something about results being out before polls closed.
Next question then is how fast the counting would be. We have paper ballots here in Sweden too, and we had an election recently (june 7). Here, too, results were in quite soon after the polling places had closed, I think it took about an hour. Not all results of course, but it took quite some time in Iran as well before all the votes were in.
Karoubi getting fewer votes than he has party members -- possible, but....
Again, I would want to know where those claims are from. I must say I'd be quite impressed if he really has more than 300000 party members.
Personally, I'd be interested in seeing results broken down not just by region, but so that figures were given for each place where votes were counted. If those stats were made available (or they already are), then I can't imagine wholesale fraud being possible. Conspiracies that size just don't work.