The esteemed Dr. Clam and I had discussed the differences between the electoral math between the US and Australia and Dr Clam came up with something I couldn't have said better myself:(context here)
Preferential voting and proportional voting in the senate makes minor and single-issue groups not associated with a major political party much more important and successful in Australia. Third parties can play a constructive role,not just act as spoilers to cause civil wars. If there had been just one Democrat candidate in 1860 , that extremist yokel would never have gotten in with 39.8% of the vote.(Australia wins!)
*On the other hand, compulsory voting reduces the importance of single-issue groups associated with major political parties, which are valued in the US because they can get out the voter base. This means single-issue groups have less influence on the policies of the major parties here, and the parties better reflect the mainstream. (Australia wins again!)
* Finally, the US has been comprehensively gerrymandered on the state and federal level so that there are very few marginal seats, and parties can concentrate their resources even more disproportionately. (Australia, once again, wins!)
Basically, the jist of point two is that candidates on the extreme ideological edge of the major parties are an asset in the US, while they would have absolutely no chance in Australia (and would have to try for minor placings in the senate)- the mainstream would eat them up for breakfast.
Thus Sarah Palin, by dint of her being more of a wing ideologically compared to the mainstream, appears to be a blunderous choice (as it would be, in Australia). Any political commentator with any nous for the US electoral mathematics would realise that the mainstream just has to not hate her enough to be bothered voting against her.
I've been meaning to add an entry in Principia Marconomica about electoral maths, as I have done for a few other tidbits of late, the jist of it being that it doesn't matter so much that elections give spurious results, but that in the long run the electoral system is self-adjusting and stable.