2012 general election
Unlike nearly every past biennial election, 2012 saw no pervasive red shift for presidency and Senate. A red shift did appear, meanwhile, for further downticket House and state legislative races. It appears, based on election forensics and circumstantial evidence, that a top-ticket rig was planned but nixed at the last minute. Top-ticket races likely used a real-time MITM approach which was subverted, while lower-ticket races used a static pre-set rig.
The exit polls had no significant red shift for the presidential or Senate races, bucking a consistent trend since 2000. (It's worth noting, however, that exit polls were cancelled in 19 solid red and blue states, removing a comparative baseline.) At the time, Election Defense Alliance (EDA) wondered whether the lack of red shift indicated no fraud or that the exit polls were pre-adjusted to anticipate fraud. Given that red shifts have shown up in all past and future biennial elections (2002-2010 and 2014-2016), it's likely that 2012's lack of a red shift was an aberration indicating a mostly clean election.
In 2006 and 2010, the fraud appeared to target competitive elections. This makes strategic sense, as more competitive elections are easier to flip and have the results stay plausible. EDA applied this logic in an attempt to validate the exit polls. They aggregated the mostly-noncompetitive House races across each state and compared them to the highly-competitive presidential and Senate races. EDA expected the House elections to serve as a non-fraudulent baseline for the top-ticket races that would be more likely targets of rigging.
To their surprise, the House races in Ohio and Pennsylvania were red-shifted compared to the presidential and Senate races. This validated the exit polls showing no top-ticket fraud, and hinted at a reverse of EDA's theory: presidential and Senate races weren't rigged, but the House races (presumably the handful of competitive ones in each state) were. And if House races were rigged for the GOP, it's a reasonable conclusion that state legislature elections were as well.
Pre-set and real-time rigging
Why did the election riggers target House races, but opt not to rig the high-profile contests for the presidency and Senate? All three races had been targeted in past elections, with the presidency and Senate more likely to be rigged than downticket races. The likely answer is somewhat complex.
Election fraud has developed into two main techniques: pre-set and real-time rigging. Pre-set rigging involves a predetermined vote flip or election result being programmed onto the voting machines or central tabulators before the election. For example, there could be a routine to flip 5% of the vote from the Democrat to the Republican. Real-time rigging involves precisely modifying vote counts as they're tabulated to achieve a desired result. In addition to reprogramming the tabulators, a real-time rig requires the ability to intercept and alter vote counts in real time, like the SmarTech man-in-the-middle in the 2004 Ohio election.
Pre-set rigging is much simpler and less visible than its real-time counterpart. However, it requires guessing in advance how much to swing the election. When the rig is undercalibrated or the political winds change, a pre-set rig can prove inadequate, as happened 2006 and 2008 when Democrats overcame election fraud. Real-time rigging allows adapting to changing political circumstances, but it has a greater risk of being exposed, as it nearly was in the 2004 Ohio election.
The two types of election fraud are best suited to different types of races. When it's possible to predict the outcome in advance, or when you only need to win a bunch of contests in a group rather than a single one, pre-set rigging will usually suffice. This would include the lower-profile House and state legislative elections. When there's a single race whose outcome is much harder to predict, real-time rigging increases the chance of success. (It's also worth noting that real-time rigging may not always be possible to set up, especially in states that don't use central tabulators.)
Some years' top-ticket elections might work with pre-set rigging. But in 2012, the presidential and Senate races were too highly-volatile to predict in advance. Pre-election polls were wildly fluctuating, and political events like Romney's "47%" gaffe kept changing the electoral dynamics. It would have been most advisable to target the top-ticket races with a real-time approach. House and state legislative contests, on the other hand, were static enough to employ a pre-set rig for.
That would leave the presidential and Senate elections depending on a real-time rig, while the House and state legislative elections would only need a static pre-set rig. If something then happened to deter or disable the real-time rigging, the presidency and Senate elections would be clean of fraud while the downticket races would red shift. This would explain the odd red shift pattern that appeared. And the evidence strongly hints that a real-time rig was planned but averted.
Perhaps the best clue about the 2012 election was Karl Rove's "meltdown" on election night. When Fox News called Ohio for Obama, Rove publicly challenged them, citing a nonexistent rapidly-closing gap and grasping at straws to justify a Romney comeback. Rove, a well-disciplined GOP strategist, was so clearly expecting a different result that he made a fool of himself on election night. It made little sense, unless Rove was expecting a fix that never triggered.
He also mentioned that Ohio's election reporting site had crashed. This was reminiscent of 2004, when the site went down at 11:14 PM before the results were rerouted through SmarTech and Bush got an unexplained surge in votes. In fact, the site went down at 11:13 PM in 2012, just a minute earlier than in 2004. And just as tabulators were rigged in 2004, likely to allow interfacing with the SmarTech man-in-the-middle, there were last-minute central tabulator patches in 2012. The same kind of MITM setup employed in the 2004 Ohio election appears to have been prepared for 2012.
EDA confirmed that SmarTech had once again contracted with the Ohio SoS to handle the results reporting. And they learned that several other states also hired SmarTech for the same purpose. That would fit with their conclusion that a national real-time rig was planned, likely in key swing states.
However, the MITM setup ultimately failed to rig the election. Rove clearly expected a fix, as (it appears) did Romney, who was infamously so confident he didn't even write a concession speech. They were both left shellshocked when Obama won, implying the real-time rig failed to trigger. And EDA's analysis indeed found that the MITM-targeted presidential and Senate races weren't red-shifted. But why didn't it work?
One possibility is that the rig was called off for fear of its exposure. In Ohio, the SoS was sued over the last-minute tabulator patches that likely formed part of the MITM setup. The judge didn't halt the patches, but he left the door open to an election challenge that might have audited the patches. The FBI was also on the ground in Ohio on election day, including in the SoS's office, watching for signs of electronic tampering. The risk of pulling off election fraud might have simply been too great, and the fix was nixed.
Alternatively, white-hat hackers like Anonymous may have learned of and disabled the vote rigging infrastructure. Before the election, Anonymous warned Rove not to steal the election, and after the election, they sent a letter taking credit for stopping Rove's fix.
What actually happened may never be known. But all the evidence in the 2012 election points to an attempted repeat of the SmarTech MITM from 2004 that was averted at the last minute.
Main article: 2012 Ohio general election
In many ways, the 2012 Ohio election exactly mirrored what happened in 2004. There were last-minute patches to central tabulators, networking of those tabulators to a SmarTech man-in-the-middle, and late-night server issues with the election reporting website. Unlike 2004, however, the election wasn't stolen from the Democrat. Ohio was at the epicenter of why the MITM fix seems to have failed. A court challenge to the tabulator patches and the FBI presence in Ohio heightened the risk of committing fraud. But while the top-ticket races remained untouched, there were red shifts in House races.
2012's most visible prizes, the presidency and Senate, were taken by Democrats thanks to an averted rig. But beneath the radar, House and state legislative elections still succumbed to theft. Overall, very little actually changed politically: Democrats kept the presidency and Senate, while Republicans kept the House. It took a partially-disabled election fix just to maintain the status quo.
- Election Systems & Software, "Oakland County Reports Election Results with Speed, Efficiency and Accuracy", 2012/11/21 - yet another late patch by ES&S, this time in Michigan
- Code Red: Computerized Election Theft And The New American Century by Jonathan Simon (2014 link; 2018 link)
- Interview excerpt with Jonathan Simon, found on p.194 of Presidential Puppetry by Andrew Kreig: "The inside scoop I've heard is that additional infrastructure (see SmarTech) has been installed in key swing/Senate races to "process" votes on off-site servers a la Ohio 2004."