What if...the Russians hack voting machines for Trump?

[Fanatical Meat]

Why would you presume you needed any certain critical mass of machines to be successful? That would depend entirely on the margin you needed to overcome. You could have changed the results of the 2000 election with access to just one voting machine in Florida. And probably a handful of machines in the 1960 election. The Electoral College system is almost ideal for this purpose because polling can tell you with a high degree of precision exactly where you need to focus your energies, it's not like you have to randomly pick the correct machine out of 100s of thousands of possibilities.

Plus you don't even need to flip the election to be successful, you just need to discredit the process. Clinton will already be a historically disliked President, throw in an election win tainted by allegations of election fraud and you can deny her much of the little political capital she already would have possessed and thus ability to implement much of her agenda.

True but how do you predict what area to mess with moments before voting begins?
I'll agree its possible but very, very unlikely

 

[dullard]

Florida in 2000 was a rare case. How often do you expect a state with that much population to come to a 400 vote margin? Yet even in that case, the tamperers would have to predict where they need to tamper and how much because they wouldn't know in advance exactly what the unrigged outcome would be, precinct by precinct. The closest example comes from fiction, the TV show "Scandal" where supposedly they rig an election by tampering with voting machines in one key bellwether county in Ohio. Yet they never explain how they knew in advance they would win by tampering with just that one precinct in that one state. Because it's TV and it doesn't have to make sense. Reality has to make sense.

I agree that there really isn't any evidence of rigging. But I disagree with your contention that it is rare to have close states in close elections. Even in the 2000 Bush Gore election, New Mexico was actually a closer vote, with Gore winning by just 366 votes. While not quite as close, in that election Iowa was within 4144 votes, Wisconsin was within 5708 votes, Oregon was within 6765 votes, and New Hampshire was within 7211 votes. That is five states that could have been swung with just a couple compromised machines in that election. Maybe even just one compromised machine in some states.

And close elections aren't that uncommon. In about 8 of the 57 elections we have had (14%), one state could have made the difference (there were two ties before the 12th amendment). So there are fairly regular opportunities to throw an election if one wanted to do so.

I don't understand your point that a person has to select the correct precinct. A rigged vote in any precinct in any close state could sway that state. You'll have months to prepare since you'll know that the race is close and which states are close. Find a precinct that seems more lax than the others.

 

[Mandres]

OK, where do you count (just the state would be fine unless you want to be more specific) and what is your job title?

This is a topic close to my heart - let's talk about it. I'm the tabulations supervisor for my County in TX. It's my job to collect/receive the electronic ballot data from all of the voting precincts, load it into the central tabulating computer and generate the County-wide results reports. I'm very interested in hearing about how the machines/system can be hacked. That's understandably a scary idea, but I don't think it's realistic.

If you, the Russians, or any other scary group were going to hack my County's electronic voting system how would it be done? Let's test the controls in place

An additional question in this round would be were you taught to question authority or were you taught to blindly follow authority?

I said a grown-up conversation ...

 

[woolfe9998]

I agree that there really isn't any evidence of rigging. But I disagree with your contention that it is rare to have close states in close elections. Even in the 2000 Bush Gore election, New Mexico was actually a closer vote, with Gore winning by just 366 votes. While not quite as close, in that election Iowa was within 4144 votes, Wisconsin was within 5708 votes, Oregon was within 6765 votes, and New Hampshire was within 7211 votes. That is five states that could have been swung with just a couple compromised machines in that election. Maybe even just one compromised machine in some states.

And close elections aren't that uncommon. In about 8 of the 57 elections we have had (14%), one state could have made the difference (there were two ties before the 12th amendment). So there are fairly regular opportunities to throw an election if one wanted to do so.

I don't understand your point that a person has to select the correct precinct. A rigged vote in any precinct in any close state could sway that state. You'll have months to prepare since you'll know that the race is close and which states are close. Find a precinct that seems more lax than the others.

I said 400 votes in a state with that much population, and hence the high number of electoral votes. None of the examples you provide invalidates that statement. New Mexico has 1/10th the population of Florida. In relative terms, the vote was about 10x closer in Florida. This matters because the likelihood of a state like Florida changing the total election outcome is much higher than for New Mexico. It IS an extreme rarity that tampering with a few hundred or even a few thousand votes will tip an election.

Sorry if I was unclear. They don't have to select the correct precinct. But they do have to select the correct state or states, knowing in advance the state will be close and by approximately how much. You could go by polling, but if aggregated polling is off by even .5%, which it usually is, that means at least 10's of thousands of votes randomly shifting in one direction or the other. Depending on the state, the random variance from polls to outcomes could be 100's of thousands of votes. To guard against this uncertainty, you'd have to over-compensate by tampering with a lot more votes than you might actually need. And since you can't predict which state would be a tipping point, and indeed, it isn't all that common that only one state tips the election, you'd have to tamper with multiple states. In the example I gave from the TV show "Scandal," the campaign doing the tampering got absurdly lucky that their minimal tampering in one precinct actually changed the outcome, because they couldn't have known it would all come down to the tiniest margin in one state. And since the likelihood of their tampering actually changing the outcome was low, would they have actually taken the risk of getting caught?

If you had close polling, which we do not have in this particular election, at a minimum you'd want to tamper with machines in at least several states and shift 10's of thousands of votes in each state. And even that would be nowhere near a guaranteed outcome. The requirement to tamper with machines in multiple states is a real kicker. It means that the tampering could not be merely opportunistic, i.e. there happens to be a corrupt election official in one state who is someone'e brother, friend or whatever, and is willing to give would be tamperers access to voting machines in that state. Rather, it would have to be a coordinated effort crossing state lines, involving multiple corrupt officials. Probably a bribery scheme.

There is literally zero evidence of coordinated election fraud in national elections in this country, at any time in its history. "Voter fraud" so far is individuals taking it upon themselves to cast a vote using someone else's identity. Even this doesn't happen often, and there is no evidence that one side does it more than the other, so in the long run, the already infinitesimal impact of it gets cancelled out.

Sure, we might have a campaign or group of powerful interests who risk serious jail time by bribing a bunch of corrupt officials, themselves risking serious jail, all to give one candidate a somewhat better chance of winning an election. Consider the risk-benefit of such a scheme.

But hey, anything is possible. It's just a matter of probabilities.

 

[dullard]

I said 400 votes in a state with that much population, and hence the high number of electoral votes. None of the examples you provide invalidates that statement.

There is literally zero evidence of coordinated election fraud in national elections in this country, at any time in its history.

I saw that you wrote the "with that much population" qualifier. I decided that it was not a useful qualifier. Yes, in 2000 only Florida was close enough to swing the election, because the other 4 close states were already in Gore's pocket. But, in ~15% of the elections even a tiny state could have swung the election. It doesn't need to be Florida sized to change the outcome.

Yes, to be absolutely guaranteed to rig an election, you probably need multiple states. But, that isn't the question. The question is can an election be rigged. The answer is yes, if you select a state with close polling, it can be rigged if you can hack to one or two machines. You might not succeed because you chose the wrong state. But it can happen, and close states and close elections are common enough that it should be a concern.

That said, you are correct that there is zero evidence of it.

 

[sdifox]

It is really impossible for a machine to be left alone with a hacker to be hacked and not be verified, double verified and triple verified before it is put into use.
Then you have people from both party and election officials that will detect irregularity. Then parties will launch challenge if there is suspected tampering in a district. So theoretically you can hack a machine, but that doesn't mean you can influence an election.

 

[woolfe9998]

I saw that you wrote the "with that much population" qualifier. I decided that it was not a useful qualifier. Yes, in 2000 only Florida was close enough to swing the election, because the other 4 close states were already in Gore's pocket. But, in ~15% of the elections even a tiny state could have swung the election. It doesn't need to be Florida sized to change the outcome.

Yes, to be absolutely guaranteed to rig an election, you probably need multiple states. But, that isn't the question. The question is can an election be rigged. The answer is yes, if you select a state with close polling, it can be rigged if you can hack to one or two machines. You might not succeed because you chose the wrong state. But it can happen, and close states and close elections are common enough that it should be a concern.

That said, you are correct that there is zero evidence of it.

The qualifier was quite relevant because if the states are small, such that a smaller margin is probable, you'd need to rig votes in more states. Getting a larger state, perhaps if you're lucky, that one state might be enough, but then, the margins aren't nearly so small in the larger states. Florida in 2000 being a very rare exception.

I doubt one or two machines in a state would do it. How many votes does one machine count? Because the unrigged results are not known in advance, to have a margin of safety in your rigging, even with close polls you'd need 10's of thousands of rigged votes in a small state, hundreds of thousands in a large state. Even if that is potentially just one or two machines, the results of those machines would look rigged. Those machines would have at or near 100% of the votes being for one candidate. In order to really pull it off, you'd need to tamper with a smaller number of votes on a large number of machines. The dilemma is, fewer machines means a few machines that look really rigged. More machines means...having to access more machines, making the entire scheme harder to pull off.

The risk of being caught in such a scheme is quite high. It's probably why there is no evidence that it has ever happened. Because no one would take that risk, especially with the outcome nowhere near guaranteed. You admit there is no evidence, so ask yourself why. It certainly isn't because there aren't enough unethical people in politics. It's more likely that it is just too hard to do without being caught.

So sure, it's theoretically possible. Lots of things are. Just not terribly likely.

 

[dullard]

The qualifier was quite relevant because if the states are small, such that a smaller margin is probable, you'd need to rig votes in more states. Getting a larger state, perhaps if you're lucky, that one state might be enough, but then, the margins aren't nearly so small in the larger states. Florida in 2000 being a very rare exception.

I just don't think Florida is that rare. It happened in the Wilson Hughes election. If Hughes had swapped California (13 EC votes at the time), he would have won the election instead of Wilson. Instead Woodrow Wilson won by 3773 votes out of 999,603 with Benson and Hanly acting as 3rd party spoilers.
https://en.wikipedia.org/wiki/United_States_presidential_election,_1916#Results_by_state

In Tilden Hayes, any one of many states with narrow margins that could have swung the election. 1000 more votes in Florida (back then, Florida was a small state with very few electoral college votes) and Tiden would have won its 4 EC votes (out of 369) and would have won the election. Heck, ~1000 more votes in any of these: Nevada, South Carolina, or Oregon would have done it too. Although 1000 extra votes in Nevada or Oregon would have been suspicious. Even California would need a fairly small number of fake votes to swing it in Tilden's way.
https://en.wikipedia.org/wiki/United_States_presidential_election,_1876#Results_by_state

I'm not even going to get into the electoral college ties that have occured. Any state changed, no matter how small, would have swung that.

But even worse, is that you don't need to ADD 400 votes in Florida. You only need to switch 200 votes in a machine. By switching a vote, you can do half as much and it would be even harder to be noticed.

 

[woolfe9998]

You're still ignoring the fact that no matter how close the margin ultimately turns out to be, it cannot be predicted. If you were to try rigging an election, you could never count on those razor thin margins in any particular state. You'd have to rig a lot more than 100's of votes, and you'd have to do it in many states. Hindsight is just that...hindsight.

 

[dullard]

You're still ignoring the fact that no matter how close the margin ultimately turns out to be, it cannot be predicted. If you were to try rigging an election, you could never count on those razor thin margins in any particular state. You'd have to rig a lot more than 100's of votes, and you'd have to do it in many states. Hindsight is just that...hindsight.

I'm not ignoring it, we are just talking about different things. You are talking about 100% chance to change an election. I'm talking about a very good chance to change an election.

If an election is a blowout, you pretty much can't change it without corruption in many levels. But if the election is close, we know that well ahead of time and we know exactly which states are close well ahead of time. If you really wanted to swing an election, take a shot, and it is likely that you'll choose correctly.

For example, I don't think this 2016 election is close enough to alter fraudulently. But, if you wanted to, only three states are really in the air (Ohio, Arizona, and Iowa). Ohio is the biggest by far. The Ohio polling average is 42.7% Trump, 42.5% Clinton. So, if someone wanted to fraudulently influence this election, they should focus on Ohio. End of story. No magic hindsight needed. One compromised machine in Ohio MIGHT do it. You are correct that it might not be enough (I also think the election isn't close enough to change). But the chances are pretty high that if you were to try to swing the election, a compromised machine or two in Ohio is the way to go.

 

[sdifox]

I'm not ignoring it, we are just talking about different things. You are talking about 100% chance to change an election. I'm talking about a very good chance to change an election.

If an election is a blowout, you pretty much can't change it without corruption in many levels. But if the election is close, we know that well ahead of time and we know exactly which states are close well ahead of time. If you really wanted to swing an election, take a shot, and it is likely that you'll choose correctly.

For example, I don't think this 2016 election is close enough to alter fraudulently. But, if you wanted to, only three states are really in the air (Ohio, Arizona, and Iowa). Ohio is the biggest by far. The Ohio polling average is 42.7% Trump, 42.5% Clinton. So, if someone wanted to fraudulently influence this election, they should focus on Ohio. End of story. No magic hindsight needed. One compromised machine in Ohio MIGHT do it. You are correct that it might not be enough (I also think the election isn't close enough to change). But the chances are pretty high that if you were to try to swing the election, a compromised machine or two in Ohio is the way to go.

how will a compromised machine evade detection when election officials test it?

 

[blankslate]

Then the hawks will get the war with Russia that they want....

but seriously maybe people will wake up about the sheer stupidity of using ELM to count votes....


___________

 

[senseamp]

They don't need to hack anything. GOP is already rigging this election for Trump with voter suppression.

 

[piasabird]

It is more likely the democrats have already hacked all the voting machines. They are afraid some hackers might figure that out. What tests for accuracy can be performed with an anonymous voting system?

 

[piasabird]

I think when you vote you should get a receipt with a serial number on it that you can go back and check for accuracy.

This way a quality control team could pull 100 ballots out of every voting location and check them for accuracy.

 

[Mandres]

I think when you vote you should get a receipt with a serial number on it that you can go back and check for accuracy.

This way a quality control team could pull 100 ballots out of every voting location and check them for accuracy.

That sounds interesting but I'm not sure I follow. What would the serial# link to? How would a quality control team check electronic "ballots" (bits in a computer database) for accuracy?

 

[sdifox]

It is more likely the democrats have already hacked all the voting machines. They are afraid some hackers might figure that out. What tests for accuracy can be performed with an anonymous voting system?

Are you shitting me? Do 5 sets of 100 votes with differing distribution and check result. Mandate paper trail, not electronic trail if you fear tampering.

fear the boogeyman? Use paper ballot with scanner counter.

 

[woolfe9998]

I'm not ignoring it, we are just talking about different things. You are talking about 100% chance to change an election. I'm talking about a very good chance to change an election.

If an election is a blowout, you pretty much can't change it without corruption in many levels. But if the election is close, we know that well ahead of time and we know exactly which states are close well ahead of time. If you really wanted to swing an election, take a shot, and it is likely that you'll choose correctly.

For example, I don't think this 2016 election is close enough to alter fraudulently. But, if you wanted to, only three states are really in the air (Ohio, Arizona, and Iowa). Ohio is the biggest by far. The Ohio polling average is 42.7% Trump, 42.5% Clinton. So, if someone wanted to fraudulently influence this election, they should focus on Ohio. End of story. No magic hindsight needed. One compromised machine in Ohio MIGHT do it. You are correct that it might not be enough (I also think the election isn't close enough to change). But the chances are pretty high that if you were to try to swing the election, a compromised machine or two in Ohio is the way to go.

Yes, that would be the best way to go, but the odds of it turning the election are very small. First of all, your candidate might have won that state anyway. With that polling, it's a toss-up. The penalty for that, by the way, is taking a major risk of imprisonment and de-letigimizing a candidate who would have won anyway. Also, the tipping point state(s) may have turned out to be elsewhere, so your candidate still loses the overall election. Second, the other candidate could very easily win the state by more than the amount of altered votes. With state polling being less reliable than national polling, you should consider it +/-2 from where aggregated polling stands. Even if the opponent leads by 1% in unrigged results, that is ~80,000 votes (8 mill votes in Ohio 2012) so you need to flip 40,000 to have an OK chance to change that state. If national polling were actually close this time to make it worth a shot, at a minimum you would want to rig in all three states, say 40,000 in Ohio, 20,000 in Arizona, and 10,000 in Iowa. You'd want to spread all that rigging around among multiple precincts and machines in each state so that there is no suspiciously lopsided result in any one place.

All that would give you a fair chance of changing the outcome. Also a more than fair chance of getting caught. Another approach is go with the 5 closest states, and double the relative numbers I gave above for each state. By now, this is a massive undertaking with totally unacceptable risk, but that would make it actually probable to change the outcome. Even the 3 state scenario is too risky to be worth it. Which is why it hasn't ever happened.

 

[dullard]

how will a compromised machine evade detection when election officials test it?

How did VW evade emission cheating detection for years? Answer: VW programmed the equipment to detect real use vs testing use.

I don't know election machines at all. But I could imagine several ways that it might happen:

• Program it to switch 10% of the ballots but only from 8 AM to midnight on election night, then don't have paper backups so it can't be recounted.
• Program it to switch 10% of the ballots if the batch size is >1000 and then sell test ballot packs of size 100 to 500.
• Program it to work normally if the test ballots come in a large batch (the way they are probably tested), but not when the ballots come in spaced by several minutes (like it would in real elections).
• Program an erase code that deletes the data (and of course don't have paper backups), enter the code only in areas that are heavily favored by your opponent (that would be detected, but no one could do anything about it as the votes are lost).
• Program it so that it won't turn on when given a signal. Lines in that district are so long that far fewer people vote. Oops, the machine broke. No one suspects a thing. Oddly 90% of that district would have voted for your opponent, if only the machines were functional.
• Etc.

Without knowing the specifics, I cannot give exact answers. Someone who has worked on voting machines could probably tell you much better ways. Or, someone who has read the accidentally released Diebold code could probably figure out far more devious methods. Just Google the Diebold voting machine encryption key for examples: F2654hD4.

Again, I don't think a specific voting machine is being hacked. A much better way would be to hack into the voter registration rolls. Oops, everyone named Jose isn't registered on voting day. Or better yet, every other Jose isn't registered, so the winning politician can use their favorite anecdotal evidence way of "proving" it wasn't rigged by going to the media and saying "Here are 10 Jose's that did vote."

 

[sdifox]

How did VW evade emission cheating detection for years? Answer: VW programmed the equipment to detect real use vs testing use.

I don't know election machines at all. But I could imagine several ways that it might happen:

• Program it to switch 10% of the ballots but only from 8 AM to midnight on election night, then don't have paper backups so it can't be recounted.
• Program it to switch 10% of the ballots if the batch size is >1000 and then sell test ballot packs of size 100 to 500.
• Program it to work normally if the test ballots come in a large batch (the way they are probably tested), but not when the ballots come in spaced by several minutes (like it would in real elections).
• Program an erase code that deletes the data (and of course don't have paper backups), enter the code only in areas that are heavily favored by your opponent (that would be detected, but no one could do anything about it as the votes are lost).
• Program it so that it won't turn on when given a signal. Lines in that district are so long that far fewer people vote. Oops, the machine broke. No one suspects a thing. Oddly 90% of that district would have voted for your opponent, if only the machines were functional.
• Etc.

Without knowing the specifics, I cannot give exact answers. Someone who has worked on voting machines could probably tell you much better ways. Or, someone who has read the accidentally released Diebold code could probably figure out far more devious methods. Just Google the Diebold voting machine encryption key for examples: F2654hD4.

Again, I don't think a specific voting machine is being hacked. A much better way would be to hack into the voter registration rolls. Oops, everyone named Jose isn't registered on voting day. Or better yet, every other Jose isn't registered, so the winning politician can use their favorite anecdotal evidence way of "proving" it wasn't rigged by going to the media and saying "Here are 10 Jose's that did vote."

You understand these are not connected to the internet right? And VW got away with it because EPA doesn't actually vigorously test vehicles. It lets manufacturer say whatever they want.

Voting machines on the other hand are vetted to kingdom come.

 

[senseamp]

Republicans would cheer if Russia hacked the election for Trump. They would say things like, we may have gotten a Putin puppet president, but SCOTUS is too important.

 

[dullard]

You understand these are not connected to the internet right? And VW got away with it because EPA doesn't actually vigorously test vehicles. It lets manufacturer say whatever they want.

Voting machines on the other hand are vetted to kingdom come.

You do understand that they are easily accessible even though they aren't internet connected: http://www.politico.com/magazine/st...w-to-hack-an-election-in-seven-minutes-214144

And until we get the actual code from the manufacturers, we just have to trust that the manufacturer didn't put in anything bad.

 

[sdifox]

You do understand that they are easily accessible even though they aren't internet connected: http://www.politico.com/magazine/st...w-to-hack-an-election-in-seven-minutes-214144

And until we get the actual code from the manufacturers, we just have to trust that the manufacturer didn't put in anything bad.

Pretty sure the states dont let anyone near the machines and open them up so they can do a rom swap.

The states dont demand code review (under nda)? That seems extremely unlikely.

 

[IGBT]

If the Russians wanted to influence the US election all they would have to do is contribute to any number of the fake clinton charities or that money laundering fraud called the clinton world initiative.

 

[dullard]

Pretty sure the states dont let anyone near the machines and open them up so they can do a rom swap.

The states dont demand code review (under nda)? That seems extremely unlikely.

Unfortunately, voting machines often are left alone. Here is a blog where someone goes around photographing voting machines that are unguarded:
https://freedom-to-tinker.com/blog/felten/e-voting-links-election-day/

In general, states aren't allowed to see the code. The companies claim it is "proprietary". California was the last to do it a decade ago:
https://en.wikipedia.org/wiki/Electronic_voting#California_top_to_bottom_review
"The security experts found significant security flaws in all of the manufacturers' voting systems, flaws that could allow a single non-expert to compromise an entire election...The Premier Election Solutions (formerly Diebold Election Systems) AccuVote-TSx voting system was studied by a group of Princeton University computer scientists in 2006. Their results showed that the AccuVote-TSx was insecure and could be 'installed with vote-stealing software in under a minute.'"

Only about half of states audit the results after the election.

Keep in mind that I don't think this is actually happening. But we can put in simple steps to prevent it from occurring. Keep machines locked and guarded. Loctite the screws. Audit all results with a required paper trail. Ban use of machines where the code isn't available for review. Etc.

But, then we'd be left with the real source of rigged elections: laws (and even the original constitution) that make voting harder for certain groups. There has been so much of this form of rigging that we needed Amendment 15, Amendment 19, Amendment 23, Amendment 24, Amendment 26, and even additional proposed amendments to the constitution. Rigging is far easier by just preventing your opponent from voting (or making it very difficult to vote) than by stuffing a ballot box or hacking a machine.