[Freeciv-Dev] Re: civ3's answer to smallpox

["Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>]

As a prelude, the basic growth ideas behind ICS are applicable at
different levels and for other game elements than purely population.

To really max out Civ growth, you need to understand when a new 
strategy is more effective to push some element, and that this is
more effective long term than any other particular option.

To start, despotic ICS is the only real strategy and you need to 
pursue this.

At some point, others may become available, and if you can take 
advantage of them, you switch play modes. 

But when these peter out or run up against some of the penalty barriers, 
you switch again or back to the original mode with hopefully an enhanced 
multiplier.

To beat a pure ICS player, you have to pick the moment to move your
Civ to a new level of play. If you do it successfully your new growth
rate should be multiplicatively better and you will blow the inflexible
strategy out of the water as eexponential processses take over.

At 01:23 AM 01/10/21 +0200, Reinier Post wrote:
>On Sat, Oct 20, 2001 at 10:39:00AM -0400, Ross W. Wetmore wrote:
>> I missed Arien's earlier post so will piggyback on Renier.
>> 
>> Again ICS doesn't work and Arien's "faster" analysis is wrong because ...
>> 
>> Assume you start with a size 1 city with food +2, prod +4 which *really*
>> weights things in favour of ICS (double average production).
>
>OK.
>
>> You will generate a new settler every 10 turns, and if your 2nd city is
>> as productive and not that far away it will take 30 turns or so to get
>> 4 cities plus some noise (up to 10 more turns) to get them all to size 2.
>> During this period you are totally defenceless :-).
>
>Note that I was discussing the standard setup here, in which you are
>unlikely to meet anyone the first 50 turns or so. Most games are
>played with generator 2 or 3 which makes it even more unlikely.

All the better in some ways. No defense means you can concentrate on
pure growth improvements.

But the lack of defensive capability of an ICS player is one of the 
critical weaknesses in real play. The player with a slight tech edge
and a few military units, can devastate an undefended ICS Civ in very
short order, and really slow it down if production is diverted to 
defense.

Also, cooperation between weaker players at occupying the dominant
ICSer will quickly bring the Civ down to size.

Trade is missing in Freeciv. But sharing trade (e.g. tech exchange) is 
a way to more effective growth, and restricting the sharing to dominant 
Civs again limits them.

So part of the issue may be with the artificial constraints in your
standard play mode.

>> It takes 10+15 turns to let the original city grow to size 3, and 5 more
>> to rapture it to size 8.

Remember 10+15 = 25 to answer your question below. 20+30 food at +2.

>OK, I admit, I've never tried it.  The reason being that rapture, too,
>is 4 times as attractive with 4 times as many cities.  (And rapture
>used to start at size 5, but this is no longer the case.)

Yes, Republic sprawl is a *lot* more effective than Despotic sprawl :-).

>> During this period the city can produce military
>> defence and/or unhappiness improvements like temples, or productivity
>> enhancements like marketplaces.
>
>Yes, this becomes interesting at that point.
>
>> And at this point vertical growth is moving at 1 per turn,
>
>??

Under Rapture, you add one new population point per turn per city. 
There is no associated cost other than any Lux levels needed to 
maintain happiness, but Lux is doubled or more in advanced governments, 
so they can effectively do this and still maintain a par with a despot.

>> while horizontal
>> growth still takes 10 turns per city.
>
>You don't grow cities, you build settlers.  I don't follow.

If you need to build a Settler (costs one pop point), then plant it
(costs 40 production points), your total cost is these plus the time
to build and move the unit - far less efficient though different mix
of resources used to grow each new worker == pop point. The 10 turns 
was a rough estimate for the given example with +4 production.

>> Moreover, vertical growth can add
>> 1.5 equivalent trade/production per turn with somthing like marketplace, 
>> 2 equivalents per turn with marketplace and bank, etc.
>
>All those resources could have been a multitude of cities,
>if settlers had been built instead.

A marketplace is 2 settler equivalents of production, but without the pop 
decrease and food penalty. If I add it to a size 8 city, I get the 
equivalent of 4 population points of trade enhancements. In a size 20
city this is 10 population points worth.

This vs the 2 that your 2 founders always give.

And it usually is the case that resources like trade are not evenly
distributed. It is much more efficient to enhance existing high value
workers than run off and get only a poor quality one at a new tile.

You build factories at mining cities, and research or finanical 
enhancements at road rich farming super-cities, or at oceanside.

>> If you rearrange your workers to produce food +4 and prod +2, you can
>> cut the 25 turns in half for the vertical growth.
>
>Sorry for being dense, but which 25 turns?

Look back to the 10+15 above. One can get to size 3 city in 13 turns
with these values that choose a food vs production strategy. This is
18 turns to produce a size 8 city vs the 30+ to get 4 size 2 cities
using a pure despotic ICS technique.

This is close to a factor of 2 on your exponential growth curve, i.e.
2^n raw boost.

Production is less effective than food for most growth strategies.

>> In advanced stages of the game, you can produce 1 settler per turn in a
>> large city. 3 such cities can produce a new city that starts growing at 
>> one per turn, every turn. I'll leave you to figure out the multiplicative
>> factor in the exponential growth this has over despotic ICS :-).
>
>The question is: how to reach those stages as early as possible.
>If the answer is smallpox, there is no more room for such a city.

As stated in the introductory comments. You start with (at least some
form of) ICS until you have reached the stage where an alternate strategy
becomes available. Concentrating on Monarchy/Republic/Mysticism/Trade
research gives you the growth techs, and the subsequent penalty reducing
techs, or productivity enhancement techs to exploit your growth, will
help you hit this point sooner.

Once you have critical mass, Rapture yourself to the new plateau for
as many turns as this is effective, then fall back to Republic ICS but 
on a significantly improved growth curve. Periodic puffs of rapture are
usually more effective than the steady state until you are really 
rolling, but once you have a solid financial and production base, you
can really accelerate city growth and expansion for new cities by
combining the enhanced resources of large cities to buy or build
components that take a long time to initially develop.

>> So, ICS is only a useful strategy in the early stages of the game where
>> tech has not progressed enough to take advantage of other growth means.
>
>No doubt about that.  But as I said, the game is usually decided in that
>early stage.

I should clarify here, despotic ICS via settler production only. 

One should be able to overcome even fairly significant leads in raw
size with an accelerated growth curve. Tanks against Legions is really
quite effective at reducing seemingly insurmountable odds. Chariots
are really ineffective against Riflemen fortified behind city walls.

>> If your game never moves beyond a despotic bloodbath, it may be the only
>> winning strategy, but that is up to the players to decide :-).
>
>Yes, by avoiding smallpox-inducing settings.  Otherwise, you have to go
>by a code of honour (agree not to use smallpox) or a code of dishonour
>(only play much weaker players so you can crush them using any strategy
>you like).
> 
>> ICS may be exponential, but it is an inefficient or low base rate that it
>> starts from.
>
>I must admit that rapture from size 3 changes the equation, but it
>isn't clear to me how you make growing a city a better investment of
>time and shields than pumping out size 1 cities.

The initial growth mechanics should be clear if you think through the 
above.

The more subtle effects are that doubling the basic terrain resource
accumulation can be done through terrain improvement or buildings. The
amortized cost of a building goes down with every additional population
point that can use it. Settler costs and terrain improvements in themselves
are constant.

Adding resource multipliers is really what the game is about, and except
for the final pop size component of the score, population is really just
for collecting resources. Besides, population in vertical growth is
rated significantly higher in the final scoring.

Despotic sprawl is very ineffective at utilizing its population. The
exponential growth loss from fractional utilization will kill it against
a more efficient Civ.

That is if the despot doesn't kill you first, but that is just the initial
challenge of the game :-).

>> Cheers,
>> RossW
>-- 
>Reinier

Cheers,
RossW
=====