The only way to have "mechanically caused" undervotes, or to shift votes from one candidate to another, is to tamper with the gears or levers themselves. This would be readily apparent at the end of the day, at the polling place, where the votes are observed and recorded, because some viable candidate would end up with zero votes on that machine. There is no way to shift some of the votes without shifting all of the votes. These are levers and gears, not computer programs. Forensic evidence of the tampering would be left behind. Lever machines are mechanical devices. Tampering with a lever machine is like tampering with a car.
Rady Ananda has well explained the fundamental differences between mechanical devices and software engineering. "One is physically lockable; the other is electronically mutable. Physically locked levers will reveal tampering; software driven systems will conceal it." Lever machines are "testable, durable, and very difficult and time consuming to rig. Proof occurs in the testing of its ability to increment correctly. ... The machine stops counting and that is immediately obvious."
"When you open your screen door, by pushing on that door handle button, you can't see what's happening inside the mechanical device. You can't see the spring. You can't see that the button you pushed is pressing into something else that unlatches the door. But you know it works properly because the door opens. This is the beauty of a mechanical system the proof is in the function."
A number of unfounded attacks on lever machines have circulated lately on the internet. Most that I have seen are hopelessly uninformed rants by activists who have never seen a lever machine, much less voted on one. I regret that I must respond to them at all.
Writer #1 alleges that lever machines "produce" exit poll discrepancies. This is silly on its face. Exit polls are quite separate from the machines. Moreover, in the infamous 2004 presidential election, when the exit polls in nine of ten battleground states showed Kerry winning a much higher percentage of the vote than he was awarded in the official results, there was no discrepancy in New York State. The exit poll had Kerry winning by 17.3%, and the official results had Kerry winning by 17.3%. The same writer alleges that "programmers" "know" that lever machines are not to be trusted, and that those who do trust the lever machines are "ignorant." To the contrary, it is Writer #1 who is ignorant. Lever machines are not "programmed" at all. They are mechanical devices, not computers.
Writer #2 states that although the "lever voting machine is not quite as inscrutable a closed box as an electronic voting machine and does take more work to reprogram in bulk, such things can still be done." He gives no examples of such reprogramming of vast numbers of lever machines. I doubt that he can. They are mechanical devices, not subject to "programming" as are electronic devices. The only two vote-rigging methods he suggests "zeroing the vote incorrectly, or messing with the adder gears" are exactly what election observers check for. The writer claims that lever machines are "an invalid mechanical solution" without offering any proof. He later acknowledges that he does not have research to support his position, and then restates it anyway, claiming that "The machine count cannot be observed without taking apart the machine," and that this would have to be done while the votes are being cast, which would violate the voter's right to a secret ballot. This argument is specious. It is rather like saying that cars cannot be trusted to function properly unless somebody is looking under the hood while the car is being driven.
Writer #3 alleges that lever machines "break the chain of custody," and that they "sit between the person and their vote." This is an unfortunate misuse of terminology. The lever machine tallies are observed and recorded right at the polling place, in full public view. The chain of custody is not broken unless and until the lever machine is taken away from the polling place and removed from public scrutiny. The writer later states that "lever machines transfer voter intent to a ballot," and then the ballot is deposited by hand into a ballot box. Clearly this writer has never seen the lever machines that she so vociferously attacks. There is no ballot at all. The votes are tallied by levers and gears. The writer then raises the case of a candidate (obviously Paul Harmon) who lost a close race in Licking County, Ohio. The writer claims that Harmon "forced a look at the actual lever machines (and) when he took them apart he saw where the teeth that were supposed to poke holes in the ballot were filed down to a nub and those ballots (only on the Democratic levers) were seen as undervotes where a person did not vote at all." This writer is confusing lever machines with punch card machines. Licking County used punch cards in the 2004 election to which this writer refers. Moreover, punch card ballots are punched by hand, by the voter, not by some mechanical device hidden within the machine.
Writer #4 attacks lever machines because "voting cards" can be laminated "to make some punches almost impossible to make, while others would almost drop out, of themselves." The writer makes the same mistake regarding Dan Rather's expose of "clearly defective punch cards" in South Florida, concluding that "the history of levers is a very sullied one," and that "lever machines are entirely riggable and unreliable in their performance." Clearly this writer, as with Writer #3, does not know the difference between lever machines and punch card machines. I rest my case.
I generally make it a point not to publicly criticize election integrity advocates. But these rants are just plain wrong. I have voted on lever machines in New York State since 1972. I know what they look like, I know how they operate, and I surely can distinguish them from punch card machines. These misguided activists who are slamming lever machines without ever having seen one, and who do not even notice when the accounts they are citing are actually describing punch card machines, should just shut up before they cause further damage. They are playing into the hands of a federal government that is presently suing New York State to force us to abandon our beloved lever machines and replace them with optical scanners or electronic voting machines. We don't want or trust either of them, and we don't like being told what to do by the federal government or by people who don't know what they are talking about.