Can math save us from electoral fraud?

The new City Paper offers us a feature article on a mathematician-cum-political consultant who hopes to use statistical analysis to verify the results from the hot fall elections. The concerns stem from odd effects seen in battleground states in 2004, but it's hard to say whether cross-referencing voting patterns against registered party affiliation, local turn-out, and other measures can offset the potential for tampering with modern voting machines. The article raises more questions than it answers, but it will be interesting to see what comes of attempts to oversee the voting system from outside.