As I and everyone else reported, the Berkeley study found "an excess of from 130,000 to 260,000 votes [for Bush] in certain counties using e-voting equipment."

Keith Olbermann has a clarification on what those two numbers mean—

This has led to the assumption in many quarters that Cal Tech [sic] has suggested as many as 260,000 Florida votes could swing from Bush to Kerry (enough to overturn the state). In fact - and the academics got a little too academic in summarizing their report and thus, this kind of got lost - the two numbers already consider the prospect of a swing:

a) There may have been 130,000 votes simply added to the Bush total. If proved and excised, they would reduce the President’s Florida margin from approximately 350,000 votes to approximately 220,000;

b) There may have been 130,000 votes switched from Kerry to Bush. If proved and corrected, they would reduce (by double the 130,000 figure - namely 260,000) the President’s Florida margin from approximately 350,000 votes to approximately 90,000.

Previous post
"Something went awry"—UC Berkeley on Florida voting (updated)