Hacker News new | past | comments | ask | show | jobs | submit login
The Electoral College and the Knapsack Problem (mike.place)
82 points by williamsmj 16 days ago | hide | past | web | favorite | 55 comments

Now, what if Republicans try to react by moving between states themselves to prevent Democrats from accomplishing their mission? Then, we have a kind of game-theoretical knapsack problem. I could imagine that generalizing and formalizing this scenario in detail would actually be a neat scientific contribution. Any comments that point to relevant publications on this are much appreciated.

If you do this with perfect knowledge and both sides playing optimally, I believe that just ends up being the popular vote, so the democrats win. Whoever has the popular vote simply has more people to shuffle around.

I suppose what you mean instead is that each side gets one round, and the new goal is for the democrats to create a distribution which the republicans cannot win by moving people around.

I believe that ends up being identical to the knapsack problem since you now simply have the democrats wishing to maximize their margin in each state based on its number of EVs; in fact, I think that's an even more straighforward knapsack problem than the one in the article since we don't have to take the complement.

Perhaps you meant something else entirely different, in which case I don't think you described it clearly enough.

Agreed. I wasn't describing the problem clearly at all, it is just a rough idea. Now, I agree that the problem is only interesting if we keep some constraints, a) at least as to how people can move and possibly also on b) what information people have on the strategy of the opposing team, as well as c) how teams can coordinate among their own members. Another fun constraint would be to model mixed democrate/republican households of different sizes that can only move together.

>Democrats win 21st century elections by relative landslides, while Republicans win by very narrow margins

Republicans would need orders of magnitude more people to move

Really cool little project; I'm cheered to see it.

There seem to be one or two people who are taking this as a serious idea. Stating the obvious; it is faster and cheaper for political parties to change their policies to things people will vote for. Nobody is an [Party X] because they were branded at birth.

The reason the margins are so close is because the politicians are purposefully doing just that, and it is knife-edge which does a better job. 2008 and 2012 represent colossal failures by the Republican party to campaign effectively/Obama was an election winning machine.

On the contrary the Republican party does an amazing job campaigning effectively, gerrymandering, disenfranchising, misinforming and scaring their voters.

It's amazing the divergence between the policies that the US population supports and what gets enacted by their government and that's down to a lot of hard work and low morals on the behalf of the Republican party.

2/3rds of voters who approved of Trump in 2016 think Obama is a Muslim. That's truly impressive work and should not be dismissed as an accident.

Isn’t the bigger issue that it actually matters whether Obama was a Muslim or not?

All of the polls show that Republicans and Democrats are not that far apart on non social issues. Trump went after people who wanted to “protect their culture”.

>Isn’t the bigger issue that it actually matters whether Obama was a Muslim or not?

Yeah, but the implication in calling Obama a Muslim was that he was also sympathetic to Islamic extremism (see: the meme that he "founded Al Qaeda",) and that he was possibly a Manchurian candidate (after all, why would he hide his Muslim status if he didn't have nefarious motives?,) and that was just a variation on the theme of "angry, violent black man" scaremongering.

Remember that the Republicans tried to paint Obama as a radical Christian black separatist at first through guilt by association with Bill Ayers and the fiery rhetoric of preacher Jeremiah Wright. That didn't stick, so they pivoted to "secret Muslim" in order to take advantage of post 9/11 xenophobia.

None of it was intended to speak to any part of the populace that didn't already consider "Muslim" a pejorative by default.

The people who believed that weren’t going to vote for Obama anyway. I honestly don’t see how Trump won’t win the electoral vote in 2020.

Honestly, I’m as far as you can get from Trump’s base, but all of the Democrats scare me as being too far left. I would love to see a Democratic President and a Republican House (the Senate approves judges, etc.). The fewer laws the government can pass the better.

The Democrats are learning the lesson of the last election, which is that voters want outside the mainstream radicals, not centrists to follow the status quo. Whether that's the correct lesson to learn remains to be seen.

The lesson the Democrats need to learn is that the swing voters don’t care about climate change and they care about pocket book issues. They are also afraid of “big government” that doesn’t focus on them.

I'm sorry but that seems an awful lot like "the lesson Democrats need to learn is how to be more like Republicans."

Caring about climate change and a belief in regulation and social welfare is the Democrats' entire purpose, it's what they're selling. UBI is a "pocketbook" issue, as is socialized healthcare and loan forgiveness.

I agree the Democrats fail utterly at communicating this to mainstream and rural voters, though.

Well, if they want to get elected, they need to do what it takes. All of the high minded stances will keep them from doing anything.

You can believe in social welfare without burdening business with social policies -- i.e. don't increase minimum wage but increase access to the Earned Income Tax Credit and make it easier to get. The same with healthcare, don't force businesses to provide it, the government should provide it. If it takes higher taxes to do it -- I'm okay with that.

Then you're not leftist. It's that simple.

> so they pivoted to "secret Muslim" Who is 'they'? The politicians or random citizens?

Until the last election, it was random crazies that weren't mainstream. In fact, McCain shut them down at rallies when they mentioned it. Romney didn't shut them down as forcefully but he didn't actively partake in it.

I'm certain the crazies weren't entirely random, just as the sudden rise of the Tea Party at the time wasn't entirely grassroots.

Donald Trump was a large part of this disinformation campaign. The whole Kenyan birther campaign was a racist dogwhistle.

I’ve wondered before why billionaires don’t just pay people to move to help their party win. I bet it wouldn’t even cost that much.

A couple of reasons come to mind:

- it would probably cost more than you think per person to get someone to pick up their life and move to a state of your choosing

- it's not clear ahead of time how many people you need to get to move

- it's not even really clear ahead of time which places you want them to move to. How many people would have guessed that moving Democrats to Wisconsin ahead of 2016 would be the most efficient path?

Some other issues:

1. How would you verify a moved person voted a particular way?

2. It cuts both ways... there are are billionaires on all points of the political spectrum.

3. If you're a billionaire, why would you even care that much? Politics has an effect on your bottom line perhaps, but there's simply nothing that's inaccessible to a billionaire. To that point, the politics common people are really not that important except when it can impact business.

> 3. If you're a billionaire, why would you even care that much?

The Koch Brothers' political network supposedly budgeted $889M for 2016 campaigns [0]. Clearly, some billionaires care quite a bit. Whether they're zealots or expected a positive ROI for their businesses, who knows.

Clinton + Super PACs spent $1,184,100,000 on the 2016 campaign [1]. Divided by 77,747 = $15,230.16 per person. All other problems aside, it seems within the realm of possibility that a PAC of Billionaires could raise enough money to move enough people to sway an election.

[0] https://www.nytimes.com/2015/01/27/us/politics/kochs-plan-to...

[1] https://www.bloomberg.com/politics/graphics/2016-presidentia...

$15k to move to a different state (especially given job prospects are not great in places where that can go far) are not so great.

Inversely that could pay for a couple of months of a staffer on the ground going around and trying to convince locals to change their vote. That feels like it could go a bit further than the 1 vote the $15k buys.

Plenty of struggling people out there might be pretty happy to go struggle somewhere else with $15K in their pocket.

And if a group of Mega Billionaires expecting a positive ROI believed they could lock up an election by convincing ~77,747 sympathetic voters to move, how much more money do you imagine they'd be willing to deploy?

A better exercise would be to look at this map [0] and figure out how many people need to be moved to traditional swing states to assure a win for either side.

[0] https://www.270towin.com/maps/same-since-2000

The $15k per person ignores the number of votes spending that original billion influenced.

The other major oversight is: it's almost just as easy to pay voters to leave a given state. You don't have their idealogical sympathies to exploit, but I bet $15k no-questions-asked offer to move out of ie Florida is going to get some takers.

But right now you have county level voting data, many red places weren't THAT red, just a few percentage points away from blue, but far enough red to ensure the whole state would be predictably red.

I am starting with this angle because the coastal elites live on the coasts which are blue, and we are assuming they want blue.

The numbers probably arent that high and if you also imagine that there is some herd immunity to adding more blue ideology people in, then you might get more to flip as well.

I’ve been wondering if blue people moved to those places would see different needs of the people there and begin voting red. They came for their civil rights causes that they have been assuming republicans consciously think about and choose to undermine, but notice something completely different.

It's probably cheaper to just run influence campaigns, shape the overton window, than it is to physically relocate voters. It also creates the appearance of being more democratic. Paying people to move to get the election outcome you want makes a mockery of the democratic process.

I've wondered about this too.

I'd love to host some of y'all in California here in Atlanta to help turn our state blue again. Your state doesn't need your votes, but ours desperately does.

We're going to draw the district lines this election that will impact the next ten years. Seriously, please help us undo the gerrymandering and win two Senate seats.

(Atlanta isn't half bad, either!)

Can we keep the fuckup that is California politics limited to that state as much as possible, please?

Allow these Californian expats to continue voting on California issues in exchange for voting in non-national elections in their new states.

Quite sure this is electoral fraud and illegal. Downvoted.

Probably not? You vote where your primary/legal residence is. If you stay somewhere long enough, that's your primary residency. Congrats, you're now a legal resident and able to vote!

Now if you don't actually establish Georgia as your legal residence, that's a whole other animal. But there's nothing stopping someone who rents in LA from ending their lease, moving to Georgia for 6 months, working remotely and voting in the GA election. And what they do with their life after the election is their choice.

Is this a moral gray area? Yea, probably. Is this a legal gray area? No, it's pretty well settled.



18 USC §597 says:

> Whoever makes or offers to make an expenditure to any person, either to vote or withhold his vote, or to vote for or against any candidate; and Whoever solicits, accepts, or receives any such expenditure in consideration of his vote or the withholding of his vote— Shall be fined under this title or imprisoned not more than one year, or both; and if the violation was willful, shall be fined under this title or imprisoned not more than two years, or both.

While it would be legal to switch your residence to a different state to vote tactically, it is definitely against the law to pay someone to do that (and to be paid to do that, too).

echelon made no offer of any expenditure, so your point is valid but orthogonal.

Update: It’s also unclear that enticing someone to move and paying them to do so falls within that definition, because you’re not paying them to vote or withhold a vote. They’re free to vote, or not, as they see fit.

It's perfectly legal to recruit someone thru very careful Facebook advertising of just the right ideology and demographics and then offer to relocate them. So long as you don't discuss voting, or how they will vote it's not electioneering.

You have been relocated to prison in state X. Please vote

You can trade votes though.

I had that exact thought when Bezos was deciding where to put HQ2.

> just pay people to move to help their party win

If they can disguise it as something else - "Temporary relocation grant to help small communities" or some other made up name, and then ensuring that everyone involved understands the real reason, but nobody leaks it out. But that seems morally and legally wrong. Otherwise, they could just hand out restaurant coupons or prepaid credit cards to local people asking them to vote a certain way? That's cheaper than paying moving costs, rent, and finding jobs for all those people to come from another part of the country.

Helping someone move to help elect their party, is quite different from bribing someone to vote against their own preference, especially because ballots are secret and bribes are unenforceable.

> bribing someone to vote against their own preference,

The preferences are not as strong I suspect. If someone out there offered tens of thousands in relocation packages, stipends, other assistance, I imagine a lot of people would be joining that party, whatever the party is.

"Go tell this org that you are voting this way and you need help moving and you get $50k to move to Georgia" you don't think the party won't find a lot of "helpers" all of the sudden.

At that point that's not very different than handing out coupons or gift cards locally.

Just like you can wrap up the relocation package in a language that makes it "legal", can wrap up handing out gift cards in a similar language. "We just want you to go out and vote, here is a restaurant card, and a free Uber trip, and a free, an amusement park ride tickets,..." etc but you can reach a lot more "voters"t that way.

Because it is more cost efficient to use other means to the same end (votes)

You could be right but I don't know. Let's look at some numbers. Wyoming is the least populated state with 577k people. Trump beat Clinton by a 46% margin in 2016. If you can find a population that is 80% likely to vote for Democrats you can shift the state to roughly even by bringing in about 330k people. Let's call it 400k to be safe. At $10k a move that would cost $4 billion. That gets you two Senate seats plus some house seats and EC votes for a fairly long time frame. What's the return on investment from other methods like get out the vote campaigns, etc?

The cost of a vote through traditional ad spending is less than $100 [0]. Moving people can't compete with that.

[0] https://www.cato.org/blog/dollars-vote-2016-election This analysis gives ad spend as $3-20 per candidate per vote. Assuming only 20% of voters were undecided, the marginal cost can't be more than $100.

It's probably cheaper to change minds. You could probably hire full time campaigners for better value.

Even cheaper to just get out the vote. You'd be amazed how many people don't vote because they can't get to a polling place because they don't have the time or can't get a ride.

Another big one think that their vote doesn't count, a drop in the ocean logic.

it's probably cheaper to just pay off politicians, and if you are really cheapskate billionaire, just run for office yourself(like trump)

And in 10 to 20 years time move them all again

Free State Project.

Could this provide a way to measure the fairness of redistricting vs gerrymandering? I'd expect a good set of electoral subdivisions would be insensitive to people moving, but gerrymandered districts would be much more sensitive.

Interesting idea!

I haven't thought this through completely, but I think I'd expect the opposite. Gerrymandering works by designing districts in which your opponent wins by huge margins (among other things). Close races can happen for lots of reasons, but they don't generally happen in an situations we recognize as gerrymandered.

What is the easiest and fastest way to "move", in order to vote legally, and move back afterward? Asking academically, of course. Can you create pop-up towns?

I think you'd find that if enough people were able to swing a vote by this method whatever limit that is in place right now would get a lot longer.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact