Sadly, Balognius Maximus was one of the only ones I never got to see :(
The exit poles were unreliable during this election because of the sample of voters taken.
An exit poler doesn't stand outside a poling place and ask every voter who he/she voted for- thats the purpose of the election- they stand and ask a sample of people who voted, during a certain period of the day (maybe several times at varying polling places).
Then they use statistics that say a cross section of 1000 people can predict the actions of 1,000,000 people. That can work- but only if the right sample is selected. They also take into account the percentage of voters who share certain characteristics, such a sex, age, ethnicity, etc as the sample. The GOP claims (and did claim all through the election coverage on election night) that the exit polls were unfairly accounting for percentages of the population. For example, the told an NBC anchor that they belived women were being incorrectly associated with a larger portion of the vote than was actually happening.
I think what happened was due to the larger turnout this year, the exit pollers samples were distorted, and some groups that would have accounted for a higher percentage in other election did not factor as greatly as in the past.
Also, the sheer number of voters meant that a good majority of people did not get asked for exit poll data- thus destroying the argument that simply because exit poll data showed Kerry ahead means the elction should have followed suit.
*Minor note- the reference to challengers is not credible. Every major news networked took great pains to metnion how much of a non-factor they were, as they simply did not challenger that many votes- period.