As to the Fudge and Boehner comparison, they both ran unopposed and they both got roughly equal numbers of votes (i.e. in the 240,000 to 250,000 range). So given that that each district contains a roughly equal number of voters it seems very strange that Fudge's district would have an explosion of votes for both Obama and Sherrod Brown while Boehner's district wouldn't see the same counter-effect for Romney.
If I remember correctly, zero and all DEMS were at the top of every ballot. I’m just wondering if people just voted the first name on the ballot?
All the Cuyahoga numbers are here: http://boe.cuyahogacounty.us/pdf_boe/en-US/ElectionResults2012/11062012UnofficialResultsbyPrecinct.HTM When I look at an average Romney-zero precinct, say Cleveland 08-J I see 482 ballots cast (55% turnout). Johnson-L got 1 vote, Obama got 472 votes, Stein-G got 1, and there were 4 each of undervotes and overvotes. The undervotes I assume means someone left President blank. Overvotes I have no idea since the machine only lets you pick one.
Sherrod Brown won 420 to 5 (and 8 and 1 for the others), and there were 48 undervotes. So right away 1/10 people stopped checking boxes. Fudge got 367 votes with 115 undervotes. The D state rep ran unopposed and got 349 votes with 133 undervotes. The county attorney won 300 to 53 with 127 undervotes (he was a D running against a no party candidate). So there was a bit more voting in a contest with an actual opponent. The state board of education vote got 178 undervotes. Supreme court justices had 186 undervotes, 198 undervotes, 139 undervotes, unopposed appeals judge had 221 undervotes, but the contested appeals judge had only 198.
The pattern is that some voters dropped out right after Obama but others went down to hit some of the contested races. Fudge was uncontested so she got skipped in many cases.
I don't think there's any mystery in how the voters can drop out through rest of the ballot with upticks for some contested races. I don't think it's much of a mystery that Romney can get zero votes when there are hundreds of other precincts where they tried to give him zero votes and weren't quite as successful (he got one or two votes in some cases, but less than 10 in hundreds of cases).
Out of 1000 precincts there were 16 zeros (not counting precincts with 20 or less votes), 41 of 1 or less, 77 of 2 or less, 100 of 3 or less, 128 of 4 or less, etc. IOW, there were 200 or 300 precincts were they tried to give Romney zero votes and succeeded in 16 of them. I think that is entirely possible without messing with machines, but probably requires some gatekeeping and/or coaching. With some of those turnouts in the 50's it might suggest gatekeeping (the person night be somehow discouraged from voting, e.g. their name isn't on the list).