Venezuela, most likely. I doubt America'll let any of the West Indies, both integral territory and vassals, go Societist black.
Look to the West Volume VII: The Eye Against the Prism
- Thread starter Thande
- Start date
Venezuela, most likely. I doubt America'll let any of the West Indies, both integral territory and vassals, go Societist black.
I mean it's literally bought by the Combine so there's that, I'm not sure about the others though. Given the history of the caste system on the subcontinent maybe Bengal?
I think it's already said that Venezuela is gone, but Faulkner never really "had" it in the first place. I was thinking of maybe Guinea or Bengal.
Well, when you introduced that, what, 4-5 years ago, I rhapsodized about it, so I went looking for the post, 233 in Book V. There are no threadmarks, because I believe it was started long before the innovation in site software allowing them, and so it took me half an hour or longer paging through the thread to find the post. I therefore include a hot link to it.
Being more versed in details and nuances of psephology (we use this word OTL too) now than back in 2016, I can now identify Thande's APR as a variant on region-based MMP, omitting the FPTP districts and omitting the systemwide collation of votes for party, and indeed the separate party vote found in classic MMP as in Germany or nowadays New Zealand. Or simply proportional representation, but drastically compromised by the small number of seats to be won, the separation of voters into districts, and the use of what is in fact Jefferson/d'Hondt (which to be sure is widely used OTL, but I think it should not be for reasons I will show shortly. It is poorly inclusive, versus Hamilton's method which is most inclusive).
The small regional divisions (apparently something like 3 seats per district is fairly normal, though that example might be on the small end of the range, it could hardly get much smaller without turning into FPTP!) are in fact assigned proportionally by party using what is in fact Jefferson's Method in OTL terms, widely known as d'Hondt's method when used for elections.
Thande presented the iterative form, and did not call it Jefferson or d'Hondt because this is an ATL, I forget if Thande said Cooke actually invented the method in the ATL; if so, it is Cooke's Method there I guess. Unless someone else invented the spreadsheet form first (this is how Jefferson did it, long before, for apportionment of seats in the US House of Representatives) for some other problem!
If countries like the USA were to adopt "APR" exactly as Thande has it, we'd expect a closer approach to proportionality than FPTP, which is quite awful and perverse, or IRV with ranked choice, but perhaps not better than Single Transferable Vote with RC. With all the above hurdles, not really a very close approach, but certainly better than FPTP or IRV anyway.
I think the better and more obvious thing to do if one values proportionality is to have all the candidates in the entire system's elections identified as party members, or generically allies, and sum them up all over the entire system for total votes for each party.
Then as Thande says, having determined the shares for each party, in the entire (OK, I forgot the name of the state-like entities comprising a Confederation) or Confederation or Empire, grab the top N vote winners for each party.
I propose and have been thinking of for many years a variation on this theme, which I discuss elsewhere.
To give some idea of what APR would do in practice, here first of all is what the outcome of the 2016 US House of Representatives election was OTL, out of 435 seats:
Democrats 194
Republicans 241
Here is what Jefferson's method for 435 seats would make of the party balances:
Democrats 212
Republicans 217
Libertarians 5
Greens 1
And Hamilton's method gets us this:
Democrats 209
Republicans 214
Libertarian 6
Green 2
Constitution 1
Independents:
David Walker Oregon CD 3 1
Alan LaPolice Kansas CD 1 1
Preston Picus California CD 12 1
See how much more inclusive Hamilton's method is to begin with. Cooke's system could apply Hamilton's method by the way--divide all party votes by total votes in the district, multiply by three, and then observe how many whole numbers there are. These seats are won for each party, then we see how many seats are left, and assign them to the parties with the highest remainders. In a hotly contested district, where no party gets a full third of the votes, this amounts to giving one seat to each of the top three parties, their top candidate in each winning their party's seat.
One thing to realize about 2016 (and many other House of Representatives races) is that despite the near-total futility of running as or voting for a third party candidate, some Americans do persist in doing that anyway, and as a result quite often neither party wins a majority of popular votes. Any system calling itself proportional should aim to reflect such gross outcomes, and not award a party that failed to get a majority a majority of the house seats--with 435 members a majority of House seats is 218. Actually in this year even Jefferson's method does not give the Republicans that many seats, being one short, though the Hamilton outcome is more representative of the gap they should face. FPTP of course gave them 241! I call that a spurious majority, and FPTP can do worse--it can flip the outcome, as in 2012 where neither party got a majority but the Democrats got more popular votes than any other party--but still wound up with just 201 seats to the Republicans' 234! But other so-called PR systems, such as STV, can be perverse in that way too if we chop the electorate up into separate bailiwicks with no cross connection for general proportionality.
Now, Cooke's complete method works differently in that it only looks at the party weights within each district, and the example Thande gave is just three seats per district. As it happens I have district by district data on each of the 435 races in 2016, and I can show what would happen if we were to elect 1305 Members to the House using it, though it will be tedious getting the separate multipliers for each district! Then I can divide by 3 to see how that compares to the above integrated methods.
Doing this, I find the raw result (times three if we actually want just 235 seats)
Democrats 660
Republicans 641
One each Libertarian and Green
of the three independents, only Alan La Police of Kansas would be elected--the second largest vote winning independent, versus David Walker of Oregon who would not win a seat, and another California independent with far fewer votes than Preston Picus would win instead. This happens because turnout was low in that candidate's, R. G. Gonzalez, district so his smaller vote was proportionally larger versus the Democrat he ran against.
Now at least all the parties who appeared under Jefferson's rule nationwide show up here, but notice that if we divide by 3, all the minor party and independent candidates would vanish, and we would have 220 Democrats, and at most 215 Republicans. So in addition to being far more restrictive than Jefferson's rule applied nationally to the whole party votes, we have a flip! Now the Democrats dominate the House, albeit by a small margin, despite more people voting Republican.
I get a similar outcome when simulating Single Transferable Vote.
Part of what is going on here is the outcome of partisan gerrymandering boomeranging; one way to dilute the prospects of electoral victory of a party one is against if one controls district lines is to crowd your opponents into as few districts as possible; once they are over 50 percent, they can't win any more seats under FPTP so creating districts that vote 100 percent for one's opponents is good (for winning more seats, not a good thing to do IMHO). But if we suddenly use the same districts for multiseat races, the high concentration means the formerly discriminated against party now has an advantage and so what were systematic advantages for one party flip into benefiting the other. Does this problem go away if we greatly increase the number of reps per district? Let's crank them up to 9 and see what happens!
That turns out to be tedious indeed--I have a way to semi-automate tweaking 435 districts, but with 9 seats, it is quite difficult. I gave up, but when I did, with about 30 districts still one short, and some other irregularities involving having to look at very minor candidates and parties (while larger ones are still left out) the Democrats were ahead of the Republicans, and I expect that would not change. Probably the Democrats would wind up with a spurious majority. Meanwhile, while I daresay some of the districts in the ENA would be larger than 3 member I doubt they would want to go for 9 member districts!
And yet, having even that many is roughly equivalent to a threshold of 10 percent, which in the world of proportional representation practice is generally considered pretty severe. (I think such arbitrary hurdles have no good purpose whatsoever and should not be used; Jefferson's method is already pretty stringent, more than it ought to be, we don't need extra barriers).
Looking at using the Hamilton, that is Greatest Remainders, method, still with the nation chopped up into 435 districts, the answers just pop out of the slot! There are still some issues involving tracking down what appear to be viable small candidates who possibly are and possibly aren't, but just a handful, and the outcomes are for 3 candidates per district
Democrats 636
Republicans 654
Libertarians 7
Greens 1
5 independents win, and possibly three more, otherwise those seats will go to one of the large parties.
This gives the Republicans a bare majority, if we divide by three, of 218 seats out of 435--if we divide by three of course all the small fry except the Libertarians disappear however. Recall the Republicans should be ahead but should not win a majority. This outcome is far closer to proportional than the Cooke's APR described--and that is down to using Hamilton's method, despite the awfully stringent filter of 3 member districts. Which to be fair, I imagine the ENA has larger districts perhaps--but if they vary much in number of seats to be won, then we have what OTL the 14th Amendment would call unequal protection.
I can as easily do 9 member districts in Hamilton's method as 3 or any number you like. I observe an interesting thing though, if I put in different numbers per district, then the outcomes, divided by that number, oscillate. Using 1 gets us OTL FPTP results--all proportionality algorithms break down to awarding victory to the plurality winner when there is only one seat to be won.
Democrats Republicans Libertarians
Seats per district
1 194 241 0
2 234 199 0.5
3 212 218 2.3bar
4 218.5 212.5 1.75
5 219 210 2.6
6 214.5 213.85 2.16bar
7 217.43 210.86 2.86
8 214.25 212.125 4
9 216.1bar 210.2bar 4.3bar
10 213.9 211.6 4.4
So the Republicans coming out ahead with 3 is really a fluke, by and large they trail the Democrats despite winning more votes, probably because of the OTL structural advantages they enjoy being spread out more evenly over more seats working against them here. We see the Libertarian share rising, this due to the lowering effective hurdle as we go up in seats per district.
Clearly then, Cooke's APR is an improvement of sorts over FPTP, but not a tremendous one.
It would be tremendously improved if all the votes being counted for a given party were consolidated together system wide, and the largest vote winners for each party were simply named the winners, making up each party's share. Concerns that a district could go unrepresented could be addressed by guaranteeing whichever candidate wins the plurality in a district is elected, and subtracting their wins from their party's total, or alternatively the top candidate of the top party in the district--but as I said, when there is a single winner, proportionality always boils down to plurality winning, which is why, as a fractional part of a larger election system, a single winner for a district should be the plurality winner, period.
If you add in giving voters the power to vote for any candidate anywhere in the system, you would have a version of the system I generally recommend. That's necessary since there is no other way to vote for a party than to vote for a candidate running for a party, and so the only way to empower every voter to have the same party choices as every other voter is to allow them to vote out-district. I'd rule that such votes do not affect the 1/3 of seats that are decided by district plurality, but do count toward party totals and toward which candidates are elevated to the body by being the top vote winners for their party (who did not win a plurality seat--this spreads party representation out among the most districts).
We would not want to give voters two votes, one for a district candidate and one for a party, because that opens the door for a way to game MMP systems that is not desirable to replicate. If party vote remains fused to candidate vote, on the theory that if a voter likes a given candidate, they probably ought to respect that candidate's judgement as to which party they run with after all, then there is no way for parties to "double dip" by running candidates in several "sock puppet" parties.
I admire two things then about Thande's APR--that the single choice each voter has fuses party and candidate vote, and that the party's winners are chosen by the top vote winner(s) for each party. These unusual features are things I think serious electoral systems should incorporate.
-----
Editing note--when I drafted this, I failed to note that the "modification" in MAPR was to give all the seats in the district to any party that got over 50 percent of the vote there, which I probably think is a more horrifying thing than even Cooke himself would have. With that "modification" combined with the weak weak PR of the basic system, which I discussed (I just went and deleted all the M's, unless I missed some) it is a pretty weak sauce improvement over FPTP. The very evil that inspired Cooke to develop APR would not be checked by MAPR.
Last edited:
Hold up, I was just joking about an Irish colony in North America.
The West Indies are smack in the middle between South America and Carolina though, which seemingly *will* fall to the Combine.
The West Indies are smack in the middle between South America and Carolina though, which seemingly *will* fall to the Combine.
Thande
Donor
Thanks for the analysis - I invented MAPR from scratch, but @OwenM tells me it is similar to the electoral system used in (IIRC) Luxembourg? I wouldn't say it's necessarily very similar to the regional part of MMP, except in the very broad sense that that is a form of PR. That uses party lists whereas this doesn't.
Also, if you want to see this system applied to OTL elections, check out this thread: https://www.alternatehistory.com/fo...k-elections-under-a-new-voting-system.374445/ You can read more about the philosophy behind the system in the OP of that thread, but essentially it is based on a philosophy that 'it is unfair for a party that gets 30% of the vote across a city or county to get no seats, but it is not unfair for a party that gets 3% of the vote across the whole country to get no seats, even though that's the same number of votes'.
What I'm getting here is that Faulkner is the "Seinfeld is Unfunny" of ENA Presidents. So much of his domestic reforms became accepted as common facts of life that nobody appreciates how innovative they were in their infancy, or how hard it was to pass them even with a divided opposition.
Someone should look deeper into and codify the common elements of that.
Bulldoggus
Banned
In the United States that's even more the case. Attlee is just a weird blip on the radar (his name- which is a strange one to American ears- likely doesn't help) who somehow beat the Savior of Britain.
On a related point, Clement Attlee is not the only case of a highly important but obscure reformist British PM. Other cases of this are Earl Grey and HH Asquith, and the latter seems the most like Faulkner. Other countries probably have similar ones, though the only one I can think of at the top of my head is VP Singh of India.
Bulldoggus
Banned
Even in your country, it seems as though St. Laurent and Pearson are often drowned out in between the titanic personalities (and egos) of King and Trudeau.