There will always be some noise in the system, and therefore a margin of error.
There sure is a lot of room for improvement, though. They should clone the architecture of ATMs (the local polling-booth machines connecting to regional servers and uploading the vote when cast) plus have the local machines generate a physical record with a sequential voter ID# that can be correlated with the electronic one on server if questions arise.
Fifteen percent would cause a massive problem if the available infrastructure wasn’t prepared for it. Imagine adding 15% more cars to a big city rush hour.
Well, being content and accepting are two different things. I’d prefer every vote be counted accurately, but until someone can figure out how that’s done, I have to be content with the system we have now. Remember, there have been inconsistancies in US voting since the 1700’s.
No snarkiness here, but if you know a better way, call any election commisioner and pitch the idea. You could stand to make millions with the next great idea. (Just give me 5% for mentioning it) 
This is the sorta stuff I’m talking about…
I know about Time’s Arrow and all, entropy increases, yadda yadda…
but surely we can do our damndest to get it right and never, ever, ever, look at it as ‘ok’ that a single American citizen was disenfranchised.
Oh, the ledgers may add up … but there are many shenanigans along the way. And “Enron-style” accounting is hardly a rarity – their error was more one of degree than of manner.
If you ever meet up with someone who works for a billing center, a bank, or a credit card company, ask them about “manual adjustments”.
A practical impossibility, not a theoretical impossibility.
A big impediment in perfect vote counting is the human factor. Training to run polling places can be particularly perfunctory and superficial, to start with – there aren’t very many real polling professionals who perfecet their “craft”. As for counting votes: punch cards can stick together, and not be noticed by the human operating the automated counter. A push-button voting machine could conceivably malfunction for a short while with no one noticing and fail to tally votes – once the error is discovered, the voters who voted but weren’t counted are long gone. With Scantron-type ballots, human voters can pencil in the blanks too lightly, or in the wrong spot – and counters can have a few ballots turned upside-down in of a stack of 2000 or more.
And no matter how many people check, and how thorough people are, this kind of stuff still gets through.
Personally, with all the errors, I think voting is currently very close to as good as it’s going to get. I, personally, agree with you about increasing precision where possible. The trouble is, the Law of Diminishing Returns kicks in, and the beancounters get in the way: “You want to spend double of what we spend now on elections, and can only deliver 10% more precision? No way we’re funding this!”
Well, how about a variation of what AHunter3 suggested?
Give me massive redundancy.
Give me an electronic record tied to social security numbers, give me a physical printout, give me immediate updates to a server, give me backups in hardcopy of this server, give me a second server that’s not connected to the net that can check the first server, give me hardcopy readouts for this server, etc…
And, hey, we could always go with ten pound marble slabs 
Not true. I live in Franklin County, and while overall there is a majority for the Democrats, and a even bigger majority for them in the City of Columbus, there are other cities in the county where the Republicans would have a very clear majority. Just because a county goes one way does not mean that every precinct in that county goes the same way.
Will do, I am quite curious now.
It just strikes me tha there must be a method of voting which is as close to %99.999 as possible. Hell, our phone networks have that reliability, why can’t our voting?
And yes… it is sad that beancounters might get in the way, but what’s more important to our republic than citizens having the right to vote and be counted?
They do and they don’t. The accounting standard is “materiality” – that is, is a discrepancy significant enough to matter to the average investor. Or, to put it another way, is it close enough?
What is or isn’t material depends on the size of the company. A $100,000 discrepancy is huge for a company with revenues of a few million. On the other hand, for a company like GE, with revenues of around 50 billion, it’s not even a drop in the bucket.
At some point, chasing down every last penny doesn’t become worth the cost – both in accounting fees and in lost productive time that could be spent in revenue-generating activity. So the financial world says “close enough.”
It’s a lot harder than you think. My wife is a Big 5/4/3 auditor, and she’s spent many a late night just trying to be sure that figures meeting the materiality threshhold match up. Drilling down to the last penny would require a small army of accountants, and wouldn’t really give investors much additional benefit.
But at what cost? What if it costs a billion dollars to have perfect accuracy, and history tells us that elections are almost decided by margins well in excess of a few votes?
The typo didn’t just show Gahanna 1-B going opposite of the county … it showed that one polling place going 95% to 5% opposite of the county.
That raised a huge red flag, and invited further scrutiny. Your point is taken in the sense that a single polling site’s abberation proves nothing alone. But when we compare the figures from all of the polling sites in Gahanna 1 to one another, and then compare the erroneous total precinct-wide Presidential vote count to the actual voter count in that precinct, the total picture becomes clear.
I noticed AHunter3’s post after I submitted mine. The reason it won’t work is tied to your suggestion about SS numbers, and is two pronged.
First, as it is now, even in my state, the only one that doesn’t even have registration, you still have to show ID to prove you can vote in that district. The way they make sure you only vote once is to check your name off the list, or if you’re in a new district, they write your name and address down in 2 ledgers to match up later if needed. After that, anyone that voted is cross-checked statewide. Both previous and new voters. Works for us and we’re just a bunch of dumb, redneck North Dakota hicks. Yet it works.
Second, tying it to a SS number is the easiest way to commit voter fraud. I won’t explain how, but it is insanely easy to get a valid SS# that will never cause a flag to be thrown in government, ever. And I’m not talking about stealing a neighbor’s identity. And don’t ask how. It’s illegal as hell, and no, I’d never do it.
And 10 pound marble slabs? :eek: First, the disabled will sue. Second, people can’t bother to get a state ID to vote and say they’re disenfranchised. :rolleyes: Third, the bulemics and anorexics (sp?) will toss them as the extra weight is unacceptable.
Dewey Cheatem Undhow: thanks for the clarification. I have to admit, when it comes to economics I’m pretty clueless.
As for what the costs, when elections generally aren’t decided by only a few voters… I have only to say 1) sometimes, they are 2) every American’s vote should count becuase that’s the foundation of a free republic.
(to me, it’s certainly worth the cost of the war in Iraq to ensure the functioning of American democracy. Maybe for the next war we can shave a few billion off of illegal contracts to Halliburton and instead upgrade our voting system)
Duffer:
I was only tossing out ideas. I’m no computer maven, nor have I studied Information Technology. I leave it to someone else to figure out exactly how.
Hell, if they need to, print out special ID cards with pictures and non-copyable holograms or something… I dunno…
~chipping away at some marble slabs~
I just want to add something about our backward state. We use optical scanners You fill out your ballot, put it in the machine, and it tells you immediately if the ballot was counted. While standing in line an elderly woman inserted hers and the machine gave a loud beep. The (really hot) election aide manning the machine explained the error was showing the lady voted for 2 presidents. What happened was she voted for Kerry then filled in the write-in circle to write John Edwards. (In Northe Dakota, voting for the presidential candidate by default adds a vote for his chosen running mate. We’re renegades)
Anyway, hot chick showed her which button to press to get the ballot back, then had her put it in a shredder over a mesh basket to show it was destroyed. The woman looked distraught that she wasted her chance and the hot chick told her she had up to 3 ballots to get it done. (Hell, if someone needs 4 tries do you want them deciding your representatives?)
We not only have historically one of the highest consistant turnouts, but even elections decided by fewer than 50 votes have rarely been contested. Works for us, even if we are all idiots.
I don’t know. It seems to me that if an election is close enough to literally be decided by something like less than 0.1% of the voters, it doesn’t really matter enough who wins to go to all the trouble to count every vote exactly and correctly. If it’s a practical toss-up, let the flawed system just pick one. It really won’t matter much.
That’s the way I felt in 2000. The election was really close and the system was (and is) flawed, but whichever side had won the other side would’ve bitched about it and nitpicked flaws in the system that we long ago had tacitly agreed to live with. It was so close it didn’t matter who won as far as the populace as a whole was concerned.
Don’t even need the system.
(Now, if anyone can tell me or email me one more time how to get that to read "Don’t even need the <hyperlink>system</HL> I’m sure everyone would appreciate it) :smack:
Loopus: I just can’t get behind that reasoning. I just can’t.
Duffer: <url=http://www.blahblahblah> text text blah blah </url>
replace the < >'s with 's
Thanks Finn, I’m gonna write it down this time. [Homer] Damn ADD! [/Homer]
I’m sure there are. Wouldn’t it be crazy if some people were more interested in the integrity of the vote then who won?
Everybody benefits if the system is fixed.
Yeah, I don’t know if you noticed, but we’re discussing that right now.
Just out of curiosity, has anyone ever ‘taken back’ a concession speech if massive result-affecting fraud was later revealed? I don’t think it’s at all likely in this case, but generally?