Civ-Specific Strategy: Americans

[Dominae]

Maybe my preference in this matter arises from the AU mod, where Specialists are "doubled" (Taxmen produce 2 gold, etc.). When I conquer cities that are mostly corrupt, it's more efficient in the short term to just irrigate away, putting pop through the roof, and taking off Laborers when pop is maxed in order to produce some Gold or Beakers (unfortunately, this requires the rushing of an Aqueduct). I'm not sure if this is the way to go in stock Civ3, where Taxmen and Scientists are pretty bad. Of course, this strat is highly reflective of my "right now" attitude towards empire-building.


Dominae

[Arrian]

I'll grant you that my corruption-fighting efforts pay off down the road. They are an investment (much like culture), the opportunity cost of which is difficult to quantify.

Speaking of specialists, when (if) you starve down a captured city (except for very early conquest, I starve most captured cities down to size 1 or close to it, unless I rushmove my palace near there and 1. want the productivity of the citizens asap and 2. have almost no flip worries), do you remember to make the citizens into taxmen or do you leave them as entertainers?

-Arrian

[BRC]

Of course, this strat is highly reflective of my "right now" attitude towards empire-building.

Haha. Maybe I was psycho analyzing.

My experience is that I can usually get most, if not all, of a 4-5 civ continent (originally that is, before I killed the 3-4 neighbors) to be productive. By that I mean less that 50% shield loss. A really good city (w/RRs, factory & hoover) will put out around 100 shields (the best ones, most end up in the 70-90 range). So lots of my fringe cities, in WLTKD end up producing 35-50 shields. It definitely takes work, and doesn't really pay off until RRs, factories and hoover when the game is probably already won, but...

It definitely takes a lot of work. The debate here could be whether it's overkill, or even whether it's worth the work. However, from my limited experience, this is all true. When you can turn out a tank a turn from a few of your cities, you have a pretty good thing going for you. And then your fringe cities turning them out every 4 or 5 turns is the icing on the cake. The AI will have a hard time keeping up with this kind of production, even with the bonuses. However, if you own your whole continent, then the game may be almost over....

This is why I'm such a luxurymonger, and why I love the Sistine.

I hate having to trade for luxuries, but I will if I can't take a city that has access to it. I will send two Galleons worth of troops to the far side of the other continent to take one city in the tundra with no chance of growing above size 3 if I can get access to a luxury that I currently do not have in my possession. In fact, once I get a map of the whole other continent, I assemble the force.

[joncnunn]

Myself, most of my cities are in the middle range where the cities greatly benifit from both a Court House & WLTPD.

I do put a great deal of importance into building Court Houses throughout my empire. (If I'm going to have a Horse-Men rush, I'll have a massive Court House campaign on all cities outside the first two ring in a tiny map, 2 rings on a normal map)

If I've decided not to have a Horse Men rush given how much territory (or it's quality) I was able to peacefully expand into, then the massive Court House campaign starts right after I'm reasonaly competent I have enough military units to keep the neighboring AIs from getting upity.

During the Court House building campaign forests will be cut down to speed the process along.

[Charis]

Corruption and WLTKD Analysis

I think Arrian and Dominae are correct in what they're saying, but have slightly different priorities (and whether it's "worth it" to spend a lot of effort squeezing out efficiency when the absolute payoff is low)

Still, there seems to be some difference here in the actual shield capability of mid-long range cities.
Arrian suggests 35-50 shields from far off same-continent cities, Dominae thinks they're totally corrupt - which is a dispute over facts, not priority.

Arrian - do you tend to space your cities out relatively far apart, no overlap, such that the total number of cities you have after taking over a 4-5 civ continent is not more than, oh, 1.5 * OCN? Or... do you figure on rushing your FP in a far off location from your palace (or palace hop) such that your palace+FP radii do actually cover a large chunk of the continent? I think one of these two cases must be true or else 35-50 shields is a massive exaggeration.

Dominae - it's important to realize there are not three types of cities, but four - this is a key point of WLTK-lovers. Core cities (highly productive), outer ring (mildy productive), *currently* corrupt cities (1 shield) and ultracorrupt cities. The latter are so far away in terms of distance and number of cities closer to the palace or FP that they will NEVER crank out more than one shield, even with courthouse, police station and WLTKD. It's critical to build NO infrastructure out there besides a temple and occasional barracks. A courthouse in an ultracorrupt city is horrible.

With "currently" corrupt cites, they look just as ugly as ultracorrupt - you're getting 1 shield. But usually they're also small, for example you might get 1 shield out of 6. That city might be capable of 30% efficiency, which times 6 is rounded down to 1. Add a courthouse, police station, grow the city to size 12, and WLTKD, and boom, you have a 40 shield producing city! Neglecting these "relatively corrupt" cities is easy to do, but a big mistake. It's easy to neglect because the 'first' step of those 4 (courthouse,PS,growth,WLTK) may not get it out of 1 shield status.

For most cities, distance is not so big a factor as its ranking in how close it is to either capital or FP - so spreading out the FP and Palace can increase those "middle" cities.

How many cities are in each category? Let's consider a 'large' map where optimal number of cities is 24, and the formula for %corruption due to rank goes as:

Fc = NC / (2*Nopt) for NC < NOpt
= NC / Nopt- 0.5 for NC > NOpt

The effect of CH or PS or WLTK is to increase effective Nopt by 25%, or by 6 on large map.
(Nopt = OCN * (1+NH/4))
Distance-based corruption adds separately, = 3.6 * distance / mapsize, adjusted for govt and courthouses.

Let's say you have two rings of 12 cities, with cap or FP surrounded by two rings, one of 3-4 cities, the other of 7-8 cities. Then many more cities past that.
The 'midrange' cities are non-corrupt even without their courthouses, and go out to about 1.25*Nopt cities.

No courthouse, PS, WLTK...
3 Inner ring cities, 95% productive
8 Outer ring cities, 55-75% productive
8 Mid-range cities, from 5% to 55% productive
ALL remaining cities past this currently seem corrupt, generating just one shield. If you have a big empire, this will be tons of cities. If you only have around 32 on a large map, only your furthest or smallest cities might seem corrupt. For illustration let's say you have 48 cities total. Then 28 cities SEEM worthless.
So the core (inner and outer rings) produce 85% of your GNP, the midrange 14%, and far-flung, 1% or less.

Add courthouse, PS, and WLTK...
3 Inner ring cities, >95% productive
8 Outer ring cities, 80-95% productive
8 Mid-range cities, from 50% to 80% productive
18 Relatively corrupt cities, from 5% to 50% productive
8 Absolutely corrupt cities

So for such an example:
- For core cities, either CH or WLTK will do, PS is overkill, and you get very high productivity either way
- The cities that "looked" midrange earlier, now that they have the support they need, are ~66% productive, capable of 50-60 shields per turn! Before: <20 shields
- A fairly high number of cities, 18, or one-third of your empire, can get around 25% productivity instead of 5%.
- Around 8 of the cities (if you empire is larger, all the rest) are ultracorrupt, and bldgs were a mistake

What's the "total civ benefit" of these additions:
Core cities have a +15% GNP increase,
Midrange have a +30% GNP increase,
Relatively corrupt have a +33% GNP increase,
Absolutely corrupt are still worthless.
These numbers are not highly dependent on the size of your empire, although the number in the ultra-corrupt range will increase directly once you get big enough to have any. (Caveat during edit - these increases would scale down if those mid- and outer cities are much smaller than the core ones, but I was taking a long-range view where they do get attention and grow)

I think these bear out that the overall benefit to your empire can be quite signficantly increased in cities which are barely productive 'now', by adding CH, PS, WLTK.
Why the stress on doing ALL of these, including WLTK?
The corruption factor might start out 1.4, all corrupt. Adding one drops it to 1.12, still corrupt, adding the third, 0.93 just a smidge higher than minimum production, then finally the fourth sees the factor drop to 0.8, for a 20% productive city, a four-fold increase in the number of shields produced after seeing NO increase after adding both courthouse and police station.

Hopefully that wasn't too much math :P

Charis

[Dominae]

Originally posted by Arrian
Speaking of specialists, when (if) you starve down a captured city (except for very early conquest, I starve most captured cities down to size 1 or close to it, unless I rushmove my palace near there and 1. want the productivity of the citizens asap and 2. have almost no flip worries), do you remember to make the citizens into taxmen or do you leave them as entertainers?

During starvation, I've trained myself to put all citizens to either Taxmen or Scientists. As you may well know, if a city starves, the governor will step in and try to remedy the situation (i.e. by transforming Specialists into Laborers). This is really annoying, because you have to go back every turn and 1) take off all the Laborers, 2) tranform them all into Taxmen or Scientists. What a hassle. But, if you're trying to get the most out of your totally corrupt cities, this is a good short-term way to do it.


Dominae

[Arrian]

Charis,

Holy ****. That was a long post.

To answer the questions you asked, though:

I use very little overlap. Maybe a few tiles. Particularly once I'm "far out" from my capitol in the expansion stage, I space a full 4 tiles between cities (no overlap). My first few cities may be at 3 tile spacing to boost my early expansion phase. Spacing is also clearly dependent on rivers/lakes, etc. Finally, when I conquer AI civs, I generally keep any cities I take, so their spacing becomes my spacing... resulting in some overlap.

I'm all about optimal Palace and FP placement. The goal is to have little or no totally corrupt areas on my continent. If it's a huge continent with 6 civs on it, no way. If it's me and three neighbors (standard map, btw), it certainly can be done.

Shall I post an example? I don't think I have a really good PTW example, but I'm sure I can come up with a good 1.29 game. I know I can, in fact.

-Arrian

[Charis]

Good idea

Holy ****. That was a long post.

I was just getting warmed up -- good thing I cut it short!
Actually, I knew qualitatively the results before I started, but Dominae's question on how much of a *total* effect on the civ's productivity was such a good one I had to work out an example to see for myself.

(Makes note to self - remember to be more concise, Charis)

[QOUTE]I'm all about optimal Palace and FP placement. The goal is to have little or no totally corrupt areas on my continent.[/QUOTE]

Aha! I knew you had either great FP placement or an extremely sparse to claim that level of productivity!

Shall I post an example? I don't think I have a really good PTW example, but I'm sure I can come up with a good 1.29 game. I know I can, in fact.

Great idea - if you have one where you did conquer a 4-5 civ continent, got a nice FP/Palace placement, and were getting good production out of even your outlier cities (with CH, PS, WLTK), that would be excellent.

Thanks,
Charis

[Arrian]

Charis,

I have several of those, one of which has its own thread ("Ultimate Power"), but I want to make sure I find a good 1.29 example. The "UP" game was 1.21 or 1.17, IIRC.

I know I have one somewhere - an Egyptian game on Emperor where I took over most of a 7 civ continent (me and 6 others, standard map) and nearly all of it was productive. It's just a question of whether it's on my computer or my gf's.

-Arrian

[alexman]

Charis, after reading your post, I believe we have just found a new corruption nazi!

[Dominae]

Here's an example from AU202 (try to imagine the map if you've tried it):

First, some background. I did an early Palace-jump to a city about 4 tiles West of the starting location. For various reasons, I placed my FP about 15-20 tiles South of my Palace city (close to where the Iroquois start), mainly because I wanted the FP up fast and that was the best location at the time. So, my core was a big blob right in the middle of the continent.

Enough background. Now consider all the cities up where the Germans were originally located (before I purged them from my continent...), specifically Berlin. I haven't done any calculations, but my Civ3-sense tells me that that's a pretty corrupt city. So I guess my question for Charis is: Will Berlin remain forever corrupt, or is it one of the "redeemable" cities? I'm not sure.

Let's just say that it's redeemable. What exactly needs to built there? Well, Courthouse, Police Station, Temple, Cathedral, Marketplace, and maybe some others (if you want to get it really productive, I suppose you need a Factory too). How are you going to go about building those? I assume some if not all of these improvements are rushed with gold. If the strategy is to drop into Communism (yay! Religious) and pop rush, then that's pretty good, but I haven't heard that argument yet. Similarly, if you're doing some major Forest planting and chopping, I'm impressed, but no mention was made of this. So, again, I'm assuming that at a lot of gold is spent over a large period of time to redeem Berlin at some future point.

Now, here's what I did with Berlin. I irrigated everything, made sure they it had an Aqueducts and that it wouldn't revolt. Then I took off the extra Laborers that I didn't need and made them Specialists. Irrigated Grasslands with Railroads produce a bunch of food, so I could get at least 5-6 Taxmen or Scientists in Berlin. In the AU mod, Taxmen are worth 2 gold, so Berlin was producing 13gpt (and 1 Shield, but who cares). Multiply this by about 5 similar cities, and I had a good 60gpt coming in from those ex-German cities.

And here's my point: I got this setup up and running about 10 turns after I conquered the German cities. So, with no gold input, no disbanding of units, almost no Forest-chopping, essentially no effort other than Worker actions, I got 50gpt (minimum, more like 80) out of those cities. Multiply this number by the number of turns between this point in the game and the point where I would have redeemed them through Courthouses, etc. and I think I've got a pretty strong case that my "right now" mentality was a good idea on this map.

The net effect was that after I defeated the Germans, I pulled ahead of every other civ by about 8 required techs. But even if the Babs or Egyptians were contenders in that game, I would still have preferred my "right now" strat, simply because it would have kept me in the race (and, when in the Industrial Corridor, you simply must win the race).

Note that I did "redeem" some other cities, but they were only mildly corrupt and generally close to my capitals. So, I'm not against "redemption" per se.

I'll grant that this may only be possible because of the AU mod (with doubled Specialists), in which case I'm happy since one of the purposes of that mod is to give more strategic options to the player.

Sorry this was so long; I hope I've made my point somewhat clearly. I would love to go back and do things differently for scientific purposes, but I deleted the saves already (don't ask why...).


Dominae

[Arrian]

Now that does make sense, I admit.

-Arrian

[Arrian]

Charis & Dominae,

I am attaching a game that I think illustrates the point fairly well (the one I wanted to attach is, in fact, on my computer, not my gf's, and I'm always at my gf's these days). I think it's a Monarch level game (1.29, not PTW)

The continent originally had 7 civs. Me (china), Japan, Rome, India, Egypt, Greece, Aztecs.

I have conquered Rome, Japan, India, Egypt and a chunk of Greece. So throw out the Greek territories, because the discussion was of a 4-5 civ continent. So if you cut my empire off at the chokepoint (if you look at the map, you will immediately know what I'm talking about), the rest is a pretty good approximation of a 5 civ continent.

There are a few cities that are useless, but the overwhelming majority, even those pretty far from the Palace/FP locations, are in good shape.

Case in point: Rome. It's neither "core" nor extremity, yet it is producing 40 shields. Not too shabby.

The only reason I'm building up my Greek cities is that I can: I'm pretty rich, and the game was won long ago. I never did play it out.

-Arrian

[Charis]

Alexman, from you that's very high praise :P
I actually was 90-95% done with a very extensive corruption guide, way back when, then you and bamspeedy blew the lid off that big mystery! I was too bogged down in getting the last details figured out first - got the OCN thing, the dual slope formula, effect of map size, and how distance was calculated. Couldn't get precisely the role of connection to capital, commercial trait, or govt, and let it go back burner. In any case, nice job figuring out and explaining all that good stuff

Dominae, super post!! I had the feeling that there was a straw man (or two) being argued against here. I was addressing the question: "Can you improve cities to the point where a very large improvement is seen in the civ's total production" -- i.e. do I 'bother with' corrupt cities or not. Your post was more functional in addressing "what CAN you do with highly corrupt cities to get you benefit, and without sinking a lot of time, or the civ's treasury, into doing it?"

To be honest, I think in most folks games, including many of mine, corrupt cities get too little attention, as we're too busy sweeping past them and conquering new, even more corrupt cities to notice. Nor do we want to send our workers anywhere near those front lines (and too often, we have too little workers!) What you describe is a very practical and easy to implement tactic to 'redeem' the value of the citizens in those town. Arrian describes the only effective way to 'redeem' the production capacity of that same set of towns, in a more UP approach that will take longer. (I just looked at a save file, same era in game, back before I understood corruption, large map also, and I made these kinds of mistakes - a ton of mines in distant corrupt cities in vain hope of productivity. And far too many 'relative' corrupt cities that weren't getting PS or WTLK)

Arrian, I looked at your save file. Wow!! UP indeed, I did I real double-take looking at all those wonders in your core cities! As for finding the palace and FP, I looked at the map, asked where would be great spots for them, and lo, there they were. Excellent job! If I look to see how many cities have a net production have 30+ shield production it's almost 30.

This seems to be a large map, and I count 47 cities (wow, was 48 a good assumption for my example or what?) which is almost exactly 2.0 * OCN.

With no CH/PS/WLTK, full corruption will hit around 1.25 to 1.5 OCN, or after 30-36 cities. With all three of those,
you don't hit ultracorrupt until 1+3/4 times that, or about 56 cities. You'er not even close to that. So for this continent and this excellent FP/Palace placement, the formula say you've avoided ultracorruption, and a look at the save file shows... that's correct! Sparta is about your worst, and it's still about 25% productive. Veii is at a distance that marks the 50% productive point, and looking at the map that's "3rd ring" from the palace.
First ring, second ring, third, fourth, fifth run then show 99%, 94%, 75%, 50%, 25% productivity. That's RIGHT on target with the numbers I gave (phew, didn't mess it up :P ) In fact, that's kind of a easy progression to remember - if you have FP and Palace with nice rings around them and you don't have more than about 2*OCN. How many 'third ring' cities are there? I was guessing 18? Only 14 here due to outstanding FP/Pal placement, it would be higher in my games :P

Now let's look at those identical cities:
Tsingtao (FP), Xinjain, Tatung, Veii, Cumae.

We'll SELL off the courthouse and police station, and kick them OUT of WLTKD. What happens to productivity?
Before: 99%, 94%, 75%, 50%, 25% productive
After..:. 97%, 84%, 54%, 12%, and single shield.

Right in line with my example

Now let's see what income we could generate with them
using a food-max taxman approach. Those outer two cities, switched from mines to irrigation, produce 8gpt and 6gpt - the latter despite the city being fully corrupt. That matches just what Dominae said! (Well, after you divide by two to cut down the specialist bonus of his mod) In fact... looking again at his save directly, Veii's commerce was only 7gpt (if you swap the one guy giving excess food to a taxman), which is less than the 8gpt irrigation-taxman approach. Cumae's commerce is 3, vs 5. So because WLTK only affects shield-waste, not cash corruption, he recoups most but not ALL of what you could get with taxmen.

In conclusion, Arrian has shown that with tender care and excellent FP and Palace placement you can get your whole empire on a large continent to have over 25% productivity, and most over 50%! Dominae has shown that with proper care and enough workers and knowledge how to get the most of of specialists, you can get a very quick gpt return for even the most highly corrupt cities. (Charis has shown that sometimes math actually works, and that someone actually read all of Alexman's gory details, but that's beside the point!)
And were all showing that good stuff can lurk in tangents that have little to do with the title of the original post

Thanks for digging out that save and for the other excellent posts,
Charis

[Dominae]

Originally posted by Charis
In conclusion, Arrian has shown that with tender care and excellent FP and Palace placement you can get your whole empire on a large continent to have over 25% productivity, and most over 50%! Dominae has shown that with proper care and enough workers and knowledge how to get the most of of specialists, you can get a very quick gpt return for even the most highly corrupt cities. (Charis has shown that sometimes math actually works, and that someone actually read all of Alexman's gory details, but that's beside the point!)
And were all showing that good stuff can lurk in tangents that have little to do with the title of the original post.

I'm pretty sure your post excels mine in the "excellence" competition; good work!



Your analysis and conclusions are right on, IMO (I actually like math too, but when other people are willing to do it for me, that means I have more time for Civ!). It would very interesting (as I mentioned above) to do a "alternate universe" analysis for "corrupt city redemption", just like with Arrian's recent choice of what to do with a super-early Leader. The point would be to see if a "right now" strategy would be more helpful to overall success that "redemption".

Question: Is it better to place Palaces centrally on landmasses so that there are many cities around the first ring but at the cost of having overlapping rings, or is better to make sure the first rings do not overlap at all so that the distance from any city to a capital is minimized? For instance, in AU202, my Palaces could have been placed further apart, but one of them would have been on a "neck" in the continent. In the second case the number of cities around an inner ring would have been reduced because you can't build cities in the water. But, some of the outlying cities would have been less corrupt overall.


Dominae