Right now, I am uploading the video to my YouTube channel.

This is the one about the sneaky plan for the Electoral College.

Now, as part of making my videos, I often make spreadsheets.

So here is the spreadsheet that I made while working on the video.

And while I'm waiting for that video to upload, I thought, Why don't I just talk you through this thing a little bit?

This isn't Ah, tight explanation or anything.

This is just going to be a scroll down spreadsheet lane with Gray as I tell you about some of the things that I did while making the video.

Okay, so for this video going up originally wanted to have something about apportionment and the Electoral College.

Every state gets a certain number of votes.

Delaware has three.

Georgia has 16.

Hawaii has four.

Uh, what are the exact details of how this happens now that part of the video got cuts and I was originally thinking it was going to become its own little thing on, and that didn't end up going anywhere for reasons that might become obvious as we talk through a bunch of this stuff.

But so Let's begin at the beginning.

Here we have all of the state's, their current estimated populations, the flag of each of these states, very important to have.

Of course, the Electoral College votes that they get.

And then the thing that drives many people to madness is how many people her Electoral college votes.

So this column down here represents how many people does it take in the state to get one Electoral college vote?

So in Hawaii, for every 355,000 people, there are four Electoral College votes.

If you go to Arizona, there's the formula.

It takes 650,000 people per electoral college votes, and you get that number by just dividing the population by the Electoral College votes there colorized here because I've divided them up into Quintiles five groups.

So you conceived, like, most proportionate and least proportionate, and we can sort that column now.

So here we are, in who is least represented to most represented as faras Electoral College votes are considered now is interesting.

Here is Texas is currently the top contender for most unrepresented by Electoral College votes, and California, which was the winner of this dubious distinction is currently in third place.

Florida is in second, and if we go down to the bottom here hoops we can see the usual.

Wyoming is the most represented.

It only takes 192,000 people to get one electoral college vote in Wyoming.

You can think of this as if you are running for office.

You need a majority of the Electoral College votes.

As you go further down this line, you need to convince fewer people in the state to vote for you to get an electoral college vote.

And if you're in the top here, you need to convince a larger number of voters to vote for you to get a proportionally smaller number of votes.

So that's how the system works.

Okay, so the question, though, is how does this number get determined?

Exactly?

Well, the number of Electoral College votes is dependent on the number of seats in the U.

S.

House of Representatives, and so the U.

S House of Representatives has 435 seats, so this 435 is going to be distributed among the 50 states, roughly in proportion with the population.

So if we take the population, put everyone in population order, you can see that roughly speaking, as we go down the population, you can see that the number of house seats is proportional ish to that number.

So the more people, the more house seats you get, the Electoral College votes is just the house seats plus two.

So that's a very easy calculation to D'oh.

Now you notice when I sort by the population over here the current population of each state, the current population per electoral college vote.

It doesn't stay in those nice Quintiles anymore.

It's gotten a little broken up.

Why is that?

Well, it's because the house seat number depends on not what is the current population, but what was the population at the last census?

So here we have now, what was the population at the last census?

And what's the current population now?

This is a little hard to look at, so we can add in this column, which shows us the population growth or in some cases, like with Illinois, the population decline of estate.

It's if we sort by this thio sort of sending so we can see that Texas and Utah Texas in Utah have grown quite a lot since last census, and Vermont, West Virginia in Illinois have all shrunk a little since the last census.

So the longer it goes between Census is, the less the current state of the House seats represents the current population, which then, of course, since the Electoral College is just this number plus two means the Electoral College also gets a little bit whacked.

So right now, Texas is at the shortest end of the stick here with having had a big population growth that isn't reflected in an increase in house seats.

Illinois, Vermont and West Virginia maybe should have fewer seats.

Well, not Vermont, because they only have one, and one is the minimum number you can have.

But Illinois in West Virginia might lose seats the next time around because their population has shrunk again.

It just useful to keep in mind that if someone gains a seat like Texas, it has to come from somewhere else because this 435 number is kept the same.

Congress passed a law saying they're not going to be any more seats than that.

So the seat numbers just have to give readjusted every census.

So then that's the big question.

How does that happen?

How are these numbers exactly?

Determined.

Well, uh, this led me down a little bit of a rabbit hole.

Okay, So the start of the process is that each state gets one seat in the house.

Let's fill that down.

So this is the way things look at the beginning.

Everybody starts with one, no matter how large or how small your state.

Now again, this process is based on the census population, not what the current population is.

So let's sorts by census population.

So next, Lee.

Unhygienic.

This column.

So here we have the census population per representative.

Four calculation.

So?

So this is thinking about the census population divided by the number of house seats.

That's what this formula is showing us here.

So these numbers start out.

Obviously hilarious.

That's Ah.

California gets one seat in the house and three Electoral College votes for 37 million people.

And then at the bottom.

Wyoming has one house seat, three Electoral college votes for 568,000 people.

So part of what happened in the process of making this video is, I had originally thought the method for apportioning the House seats was by taking this number population per representative, see who is the worst off and you give them representative, which then drops them low, sort of re sort by descending.

See who's worst off and do that again.

Give them a seat.

Sort again, you know.

Now it's New York's turn to now sort again by descending and just keep going, adding seats, always giving one seat to whoever has the worst representation based on this column, and do that until you run out of house seats.

But it turns out that is not the method by which the seats are apportioned.

That's sort of the idea behind it, but the details ended up, as they almost always are, being much more complicated and much more inscrutable.

So let's reset these couples numbers and talk about what actually happened, which is partly why there's no video about seat apportionment.

Here we are on the Wikipedia page for United States House of Representatives, and we can go to the section on apportionment.

Okay, there's a little summary of it, but let's forget that we can go to the actual page on congressional reapportionment.

So it turns out there's a bunch of different methods that have been used over time.

Okay, so blah, blah, blah, blah, blah We get down to here.

So what is the actual apportionment method?

So it turns out it's this thing called the Huntington Hill Method is what's currently used.

So if we go on to the more detailed page, this is description down here, which runs through the math of how it works.

Now this is talking about the General Huntington Hill method, but I wanted to know the exact details for the U.

S.

State House of Representatives.

So we go to the actual Census Bureau.

Who does this work?

They have a walk through of how the process works.

But as I often find looking through things, I come across these things that I think of as explanations without explanation, like it.

It tells you how a thing is done, but it it doesn't really give you a why, or it just it always feels like it's missing some key part.

So by digging around for a while, I eventually found what I guess is the first paper eso.

This is Edward Huntington of the Huntington Method.

Oh, yes, Oh, he's, he says, that the problem is basically that it's really difficult to give out the seats equally, that in theory, what you would want to do is just take the whole population of the United States divide by 435 And that's what you should have as how many people have each representative.

But the states have different populations, and it's like you just end up with this more complicated thing you don't end up with even numbers on.

I like this line here of this problem has been a subject of violent debate in Congress for the past 100 years.

Okay, so I'm just gonna point out this part here.

So none of these methods have ever possess any satisfactory mathematical justification.

So he's saying that the methods that are used or just sort of arbitrary that someone proposes them, but they don't have a great mathematical justification.

So he has this method that he's created, and I love this at the end.

This method may be called the method of geometric mean, since multipliers are reciprocal, Sze of the Geometric means of consents.

Consecutive integers The solution of the problem is thus complete.

But what I just I sort of love is that nowhere in this paper, at least to my reading do I feel like he addresses this part that none of the previous methods have any satisfactory mathematical justification.

And he talks through this method that is used and the principles that he is going for.

But I'm sort of not sold on why this is the best method compared to all the previous methods.

But anyway, I spent a long time sort of working through this and found it not easy to follow.

So I gave up on this paper because I found that later this guy who works for the Census Department rode up this much longer description of how does the Huntington Hill method actually work?

So this thing is a little bit easier to understand, and I think this was written for Congress or for the Census Department itself in order to be able to understand what's going on in this paper.

So let's talk about this just a little bit.

So here he provides the actual answer to the thing that was missing from the previous paper, which is why this method, so by the method of equal proportions, which is the Huntington Hill method.

The difference between the representation of any two states is smallest possible when measured by the relative difference in the average population per district and also by the relative difference in the individual share in a representative, these features appear to make its superior to any other method that has been devised.

So the way interpret this is that going back to my spreadsheet is not just a question about take this state that has the worst representative.

Add a house seat and make it a little better.

It's a question of if you're going to add a seat, you have to consider all of the states together.

And the question is more like if you're going to add a seat somewhere, where is the strategic place to add it?

That will bring on average all of these numbers closer together and adding a seat to the state that has the worst representative per population doesn't always achieve that goal.

Because California ends up with something like, I don't know, I think it's 50 or something.

Yeah, let's say maybe it's 53.

52 53 groups.

53.

Yeah, there we go.

California is going to end up with 53 seats.

Will just sort of cheat this.

And now let's sort by descending.

So obviously, I haven't given everybody else the seats, but, like so California has 53 it ends up with 704,000 people per vote.

There's a big difference then, between adding one more representative to California, which doesn't change its average votes per person very much and say, adding one votes to North Dakota, which drops it in half.

So the way I interpret what this paper is saying here is that it's trying to account for this effect, that adding an additional vote to a state that has a lot of votes already affects the population per representative, less than adding it to a state that has very few.

That's a thing to consider under this method.

Back to the paper.

Okay, so here's where we start to walk through the math.

We say that there are two states, A and B.

We need to give a representative to one of them.

Who should it be The basic idea is that whoever you give the additional representative to it shouldn't make the difference between the two states bigger.

You want the differences to be smaller, and that should be true for every single state that happens now.

If you're interested, there's a little math proof that you can follow in the paper, which then gets around to this idea of the geometric mean skipping, skipping, skipping.

And so, at the end, you get this formula, which just says Give the next representative to whichever state makes the difference between them smaller.

Which makes sense now with congressional method is doing is a little bit more understandable.

Is taking the formula here to calculate out who should get the next seat assignment.

So in the spreadsheet, Ah, I added a column that replicates this formula up, taking the current population dividing by this geometric mean, and that gives us this Q number, which says, What is the order right now that each of the states should get an additional house seat?

But here's the thing.

I'm gonna level with you at this point.

I can follow this, but I can't say that I truly understand why it works and this calculation of the priority number and this cue number is a little bit taken on faith now, helpfully, the census makes all of their data available.

And so here is a spreadsheet I downloaded from the census, where they show you their calculation for the priority value.

And I could check it against my own calculation for the priority value and no, like, Okay, I'm doing this.

I'm doing this right, but I don't.

But I don't really understand why or how it works.