Emails (excerpted), responses to PolitiFact Texas, Elchanan Mossel, professor, Statistics & Computer Science, University of California, Berkeley, Sept. 6, 2012

610 pm

Sept. 6, 2012

The statement you quote is not correct on the math. As you say if you make additional assumptions then it can be made true. But I think it is fair to categorize the statement as nonsense.

For example, if we make the assumption that all other states vote exactly the same as the last election then this statement obviously true. Of course in that case it doesn't matter what is the outcome in Texas :)

Not sure there is much to add to that

Elchanan Mossel, Professor

Statistics & Computer Science, U.C. Berkeley

http://www.stat.berkeley.edu/~mossel

120 pm

Sept. 11, 2012

It's really a question of using the expression "mathematically impossible" in a speech.

As a professional mathematician I would prefer to use something like  "extremely unlikely".

Such an expression indicates (as you mentioned) that some assumptions are being made.