[EM] Cost of Manual Counting vs. Machine Counting

[Abd ul-Rahman Lomax]

At 01:16 PM 5/25/2007, Chris Backert wrote:
>Any analysis of tabulation cost must consider the length of the ballot. In
>the United States in particular you'll find the number of races and ballot
>complexity are some of the primary factors in analyzing tabulation costs.

It might be noted that the calculations I used determined the cost 
per race, based on hand counting, then an assumption about the number 
of races was used in the return on investment calculation. Obviously, 
the more complex the ballot, the better machines will look. However, 
with moderate complexity, comparable to what I've seen in elections 
here, the machines are far from justifying their cost. And the whole 
thing is a red herring. Paper ballots can be scanned and counted 
automatically, at very low cost, with standard, multipurpose computer 
equipment, commonly available, so using dedicated special-purpose 
voting machines isn't necessary at all for efficient counting.

I was just using the manual count costs as a baseline. We can do a 
lot better than that without "voting machines."