New Mexico flag is just an ugly version of the flag of Madison, WI. Change my mind.

THAT was abacus-ly beautiful... Thanks 😁

I wonder how much more time it would take to get the spreadsheet to where the only interaction required to calculate would be inputting the population of each state.

Where's DC?

So this is a "casual video"?

I wish more of his videos were covered in such depth as this one is ..

It seems that this system is ideal for preventing corrupt political parties from drastically and rapidly manipulating house seats (by giving more seats to states that vote for their party), by incentivizing population redistribution.

Instead of doing it the way youre doing it in this spreadsheet, is there an alternative approach? Can you distribute seats according to population proportions and then with some special algorithm redistribute them?

Watching this at 5 am. Slowly faling asleep

I would like to see how the seats are allocated using the first method by giving them to the state that is the worse off.

It is so strange. I live in Brazil and the state capital with the least population is 306k, and 1.2 mi for the whole state. That means it has about the same population as whole states in the US.

... why it isn't just "one vote per 1.64million people" and recalculated every census year I'll never understand.

It would really be nice to see this for 2024 electors

I'm not sure if this method it objectively better tha the D'Hondt method, since they are trying to do different things. The Huntington method tries to minimize all the differences, while the D'Hondt method tries to reduce the biggest overrepresentation.

See yall when youtube has veered into obscurence and is then trying to recommend people old videos by once popular creators who have left the site due to its meany pitfalls.

Cool spreadsheet. Great insight.

Things I learned watching this video:
1. You can put the dock on the left.

Why no District of Columbia with its 3 electoral votes and award-winning flag in this spreadsheet? 23rd Amendment, people!

I literally fell asleep. 10/10, would spreadsheet again.

Wanted to know what was wrong with the process detailed at 7:40 such that the US had to use a more complicated system. I copied the spreadsheet & did a comparison. Looking only at "Population Per Representative for Calculation" rather than "Population Per Electoral College Vote", it actually ends up better on one end & "more skewed", I guess, on the other.
The Huntington-Hill method, on the high end has 994,416 per House Seat in Montana, vs 749,112 per House Seat in Pennsylvania using the 7:40 simpler method. Which makes the method seem better, until you look at the other end, where Huntington-Hill produces 527,624 per House Seat in Rhode Island, vs 409,881 per house seat in South Dakota.
Honestly, what the 7:40 method seems to do is take some representation away from the larger states, in favour of making it easier for the smaller states to get more representation, where they'd be on the cusp of deserving another representative in the HH method. To be honest, looking at the results, I don't really see exactly how HH is "mathematically better". It's kind of funny to see that in our current system, if there were a 436th spot available in the house, Delaware is 7th in queue with the 2nd worst census population vs house seats ratio.

24:40 This is my favorite part

In interesting exercise would be to find what would the electoral collage look like if the two extra votes for just "being a state" were removed and the 102 votes this creates shared out following this method

I was wondering how far off the method of just giving every seat to the least represented state, turns out, it gives the same number of seats for 35 states, and is of by 1 (+ or -) in 13 cases, gives 2 fewer to Texas, and 3 fewer to California.

It has now been nearly a year since this viodeo got released, but here I sit and watch a man go through and calculate a spreadsheet for over 10min.
And i am proud to say that it was Grey wo did it

I’d like to see this with the total number of seats reflecting the cube-root rule, which results in 693 seats.

I enjoyed watching this more than I feel like I should have.

My take away which could be wrong: The simple algorithm of dividing the total population of the US by the number of seats and then dividing the state population by the population per seat to get the number of representatives doesn't quite work because there is always a remainder. Most of this video is how to distribute the seats that remain to be apportioned after the simple algorithm is applied.
Total number of electoral votes: I didn't make it to the end of the video but does he ever mention that Washington DC gets three electoral votes so that the total number of electoral votes is three more than the sum of the number of senators plus the number of representatives? And if he does, does this effect anything he was talking about?

I actually watched the entire video. Who knew a spreadsheet could be so interesting?

Matt Parker

I love how Grey pronounces Are Kansas ;)

what is this program called?

I wonder what would happen if you would include US territories like Puerto Rico or Guam to this

I actually thoroughly enjoyed this. Very relaxing somehow. Seeing everything fall into place. It's still completely stupid as a problem, bc popular vote is obviously more fair, but oh well lol

And to think that just dividing the states respective populations by the population of the smallest states would be easier. But we'd need to not have a cap on House seats.

beautiful work - thx

20:30 Three hundred and eighty-four :)

what program is he using?

I really enjoyed this walkthrough! It's amazing to see even a glimpse of the work you put into your videos, and you still explained everything extremely well. I gotta admit I was suspicious when the video first started with only 5 columns!

Why can't 1 representative just be the smallest state, then every states' population is just divided by the smallest state to get closest to the most fair number?
Why is there a 435 cap???

Because you'd have a crazy number of people. Also it's probably to the advantage of one party to have theis system where less populous states matter significantly more so then won't want ti budge on it.

I'm curious, how would the method you initially thought of compare? Would the results be somewhat even or do some states get way more / less than they currently have?

would be intereting to see how many seats you need to make it even closer

"as part of making videos, I often make spreadsheets"
Accountants: nod in unison

Why did I sit through the whole thing?

Ah yes, “so does arKANSAS at 106”

Did he just pronounce Arkansas as "Are-can-zuss"?:)

I'm imagining a system where it minimises the total misrepresented people. Misrepresentation would be calculated by checking the number of people that voted X when the result was Y. An algorithm could find the optimal distribution of electoral college votes across states using this as the measure by which each distribution is assessed. The number of total potential distributions of electoral college votes across the states is very large (combination with replacement with 538 votes and 50 states gives about 9.6 * 10^71) but a good algorithm could limit the relevant distributions to a small number. To mathematically prove a method to be perfect, a test could be done with both this potentially unfeasible algorithm and the proposed "perfect" method on a scenario with only a few states and votes to be feasibly computed. There do seem to be other factors than this, so its not as simple as it seems really, but at least it seems there's more to it than those papers you brought up in this video.

This was a great sleep aid. Also motivated me to get started with a thing I've been planning: a spreadsheet app with visual scripting and higher order functions so that nobody has to go trought this again.

I'm not sure how to contact CGP Grey directly to convey this information, but the book "Mathematics and Politics: Strategy, Voting, Power, and Proof" explains The Hill-Huntington Method of Apportionment (page 157), and a few others that have been used over the years. It also gives a theorem (similar to the Balinski-Young theorem) that there is no apportionment method that satisfies the
monotonicity property, the quota condition, and the population property (page 158).
This book is a great resource for making videos on voting, because it has many theorems (both positive and negative) with proofs and many algorithms related to voting, apportionment and related things.

Grey seems like the guy to do every extra credit assignment... in class... the same day it was assigned

Has anyone done this with the projected census results? It would be cool to see the big winners and losers.

It would be interesting to see what happens (and what would have happened in past elections) if congress did something crazy like making the house have 2,000 seats. I.e., how many seats would congress have to add before democrats wont he 2016 or 2000 elections? (I suspect a HUGE number for 2000, but probably only a reasonable number for 2016).

Oh jesus you put the whole loop into the columns.

I'll be honest implementing the ranking in excel is pretty impressive.

Anyone know what the justification for limiting to 435 seats was? Did I missed it in the video?

Why does the winner takes all system even exist. It's ridiculous. You'd have a far better system if a) you get rid of the EC, or b) make states EC votes proportional to the ballots received. (ie so if candidates receive 50% of votes in a state, the state allocates 50% of its EC votes to each candidate)

I love spreadsheets and excel but I think this level is called fetish.

I call this issue Microwaving it... idk how it works but it keeps making my food warm...

Ar-kansas 26:00 lol

You're telling me you don't completely understand the mathematical proofs made by people that have been studying math for years and have probably written it as part of their PhD? How absurd!
Hahaha, but actually major props for trying to follow the official proof. Math is it's own language and there's definitely a learning curve there.

"Regular Kansas..."

I can't believe it's Saturday night, and I'm watching this entire video, and it's both educational and soothing. One of the best videos out there.

Mighty Spokane!

I don’t see DC on this list. If you left DC out, the numbers would be slightly off. DC gets electoral votes, but doesn’t have a voting seat in the house.

I still dont understand why this is the best system or why we cant use something that is more in line with an equal democracy?

I'm not sure I can satisfactorily answer but the naive apportionment method Grey describes at 6:38 does have a higher standard deviation of population per representative by state than the Huntington-Hill method so at least in this one case it appears to have been the better algorithm.

Did you try adding seats to lower the representative divergences? I wonder how many seats would be needed to have near the representative congruence. Great vid although I clicked through the seat distribution. You have some wicked spreadsheet skills, wow

Doesn't matter with the winner take all system

This is at least my 3rd time watching it in full since it's been released

This makes me want to program the calculation to just take in population. I love this.

The differences between this method & Canada's method for allocating seats in the House of Commons are pretty fascinating.
Notably, rather than capping the maximum number of seats, Canada specifies the _minimum_ amount (without a cap). This means granting additional seats to one province doesn't come at the direct expense of another—it just raises the total number of seats.
Next, while the House of Representatives begins each state at its constitutional-minimum, allocating the remaining seats from there; Canada instead begins with an 'electoral quotient' (total population divided by total seats), and assigns seats proportionally based on each province's population. Then additional seats are granted in order for each province to reach its minimum guaranteed by the constitution.
While I believe this method runs into the same problem Huntington suggests in his paper (non-whole number results), because the House of Commons doesn't have a maximum number of seats the elegant solution is to just round all the numbers up.
Now I'm not prepared to say which method results in less difference in representation or better representation overall, but I find it interesting that Canada's method becomes *much* simpler mathematically by not fixing the total number of seats from the outset.

28:00 POV: You are the Census Bureau mathematician in charge of calculating apportionment

I wonder how different your original system would look compared to the real one?

Looking at these population numbers and the side these states usually vote the republicans would actually benefit often from a national popular vote🤷🏽‍♂️ - but of course they dont want to because its less corruptable😪 the same for the democrats btw...

Can two or more states ever tie for the same queue position in an iteration? If so, how is the tie broken? Or is it mathematically/statistically impossible for that to happen?

very unlikely because population are large numbers and so the chances of the proportions lining up to the last digit at an unfortunate time is very low.

I believe the "largest remainder method" will give almost the exact same result. The only moment I can think where it wouldn't is when a state would end up with 0 seats with said method. If you eliminate states with 0, give them 1 and just redo for the rest, it just give you pretty much the same result
What I can assure you is that if you calculate the proportional number of representatives and round down, you can start this sheet from there

Anyone want to volunteer to do an animated graph of this being done?

Here are some facts about USA history, the Electoral College, and the civil war. The sources of this information are the USA Constitution and actual events in USA history:
Slavers are terrorists. Slavery is terrorism.
The Electoral College was written for only one purpose.
The Electoral College was written by terrorists(slavers) to be nothing more than a "welfare benefit" for themselves and other USA terrorists. The E C (+ the 3/5ths clause) awards excessive national governmental and political power to terrorists(slavers). The Electoral College encouraged and rewarded the terrorism of slavery. The Electoral College allowed terrorists to dominate the USA national government until around 1850-1860. The USA's "founding fathers" were the USA's first group of "welfare queens". Ten of the first twelve presidents were terrorists.

What happened around 1860 when abolition and the prohibition of slaver terrorism in the new territories and Western states greatly reduced the "free stuff" to which the terrorists had become so accustomed?

One of the biggest blows to the "terrorist welfare queens" was the prohibition of slaver terrorism in Western states. That's one of the reasons you hear that old csa/kkk terrorist propaganda phrase, "WE DON'T WANT TO BE RULED BY THE COASTS!".

What happened when the terrorist "welfare queens" lost their "free stuff" from the USA government?
What happened when the terrorist slavers could no longer easily dominate the USA national government and national politics?

The csa/kkk was just a low-life, MS-13-type gang of butthurt "welfare queens".

After causing the civil war, the Electoral College became a "welfare benefit" for states which suppress voting. I wonder which states LOVE to suppress voting .......... might they be the former terrorist states and terrorist sympathizer states?

Eliminate the Electoral College. It has poisoned the USA!

Everyone's favorite political math rabbit hole to dive down.
But really, the solution is to dump the cap on the size of the house, and go back to the last Constitutional rule of 1 rep per 50,000 people, and turn Congress into a Galactic Senate-like body of nearly 7,000 seats. Dilute those feudal 100 votes from the Deante to hell and gone.

Still seems a bit unfair to the more populous states.

Oh, so what happens when DC finally earns Statehood? Or PR? .......: Downloads spreadsheet:..

Me: ahhh a long day of working with databases and Data mining just finished let's see if I can find something relaxing on youtube to take my mind of things.
Oooh spreadsheet's!

Everyone: what spreadsheets
Me: yeah but what sexy program is this that so elegantly depicts depicts those gorgeous spreadsheets?

With the exception of Minnesota (7.4776714853
to 8) and Rhode Island (1.4846604037 to 2) a much simpler formula gives the same results:
Census State Population / Census National Population (309183463) * number of all seats (435) = Numer of seats per state (this has to be rounded).

Now I assume you know how the votes are distributed but you made a spreadsheet for something simple.
Electoral votes are distributed per state with house representatives +senate representatives. Now house members are still not perfect but it’s close enough. So instead of this spreadsheet to fix it just remove the senates votes. 100 less votes and now it’s more fair.

Edit: I mad this before fully watching the video... sue me

I wonder if the amount of federal land per state also connects to this in some way, based on one of your recent videos

Grey and 5upp have got to be the same person!

What would the difference be in electoral college votes if you just added one to the least representative state each time.

I wouldn't consider this method better than giving to the least proportionally represented state. The main problem is there is not enough seats to give everyone a proportional votes, and for democracy's sake, it makes a lot more sense to give enough seats (say 1 per 200,000 people) to give both fair and more complete representation. The sited excuse of the house chamber not being big enough is a very poor excuse to cap the representative count. The Huntington-Hill method favors smaller states earlier than they would be otherwise, yet the US has the Senate - a system already designed to give smaller states very high relative voting power, adding to the favoritism is enterally unnecessary for the house of representatives. If given enough representatives to fairly represent the population, both systems will reach a point of fair representation, but again, there isn't enough seats to reach it. Letting states that just don't reach the population threshold to stay relatively low to bring the overpopulated states into fairer terms is preferable.

rest is more ranty stuff about other parts of government, skip if you will.
Over time, this could lead to the senate portion being a bit to heavily negated for the electoral college, though that is a problem of the electoral college itself and not with how the house or senate are represented. The Senate, or more accurately, states borders don't accurately represent various people grouping as the senate was designed to represent, with the EU's equivalent working by country doing a better job, with few exceptions (like Utah). If the senate accurately represented groups, you'd expect senators for the black belt, Chicago to be separated from the rest of Illinois(who would likely be merged with parts of the surrounding states), California and Texas being split into multiple senate jurisdictions(with some going to each party), senators for various native American countries/reserves, along with various groups being united across state lines, ect.

Please do a video on what it would be like if the Congress did not cap the seats at 435. Like what would the Electoral College look like if there was 800, 1600, 4000 seats to allocate??? Maybe if you also did a video about how the Electoral College is dumb by redrawing lines of the States and showing how it represents more land then people??? TY

TL;DR
The better the state flag, the more people for each electoral college vote

Anyone else hear him say R-can-sis instead of R-can-saw

grey likes apple

I now understand the how, but I still don't understand the *why*. The states _are already guaranteed representation_ through the senate, why is there a need for the electoral college? Why can't at least one part of the legislative branch be democratic?

Jeez America, get yourself proportional representation. It'll fix so so much. And also get more reps even Ben Franklin didn't think this many people per seat was a good idea.

I wanna know what this looks like with Puerto Rico added in as a state

I always thought it was just the Congress Seats. Isnt the +2 for the Senate?

Spreadsheet at the beginning : Cute kitten
Spreadsheet at the end : Sauron

26:16 Ar-Kansas?

As I understand it, the electoral college votes are apportioned as House seats + Senate seats (that's where the 2 comes from,) which makes sense why Maine and Nebraska give 2 votes to the winner of the statewide vote (since those 2 votes are given for the senetors, which represent the state as a whole) and 1 additional vote to the winner of each congressional district :)

No wonder when you went into the past they didn’t want to see your spreadsheet, BOOOOOORING!
All jokes aside that was actually really interesting and I learned a lot today lol thank you for your research Grey!

I sat through nearly 37 minutes of this to hear him say Are-Kansas

Did I just watch a 42-minute video on a spreadsheet?
Yes. Yes, I did.
And I liked it.