[Freeciv-Dev] Re: ICS and corruption (waste)

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

At 12:32 PM 02/12/07 +0000, Gregory Berkolaiko wrote:
>
>In some recent emails I felt a sentiment that corruption increasing with
>distance from the capital can help battle ICS.
>
>I think it is very wrong.  Non-ICS strategy involves building cities a 
>decent distance from each other, so they all have enough resources to 
>sustain a large population.  ICS strategy involves packing cities as close 
>to each other as possible, which in particular reduces the effects of 
>corruption.
>
>So corruption (and waste) help making ICS more profitable.

The situation is a bit more complex than that. You might want to 
reevaluate your position :-).

If you consider the twenty-odd tiles around a city at distance "X" from
the capital and populate it with one city or 10 mini-cities, then the
difference is that all tiles in the first case are corrupted at the
same value given by the center tile. In the mini-case the corruption
varies by each individual city.

So for mini-cities closer to the capital, there is a benefit you see,
but for cities further out the benefit is reversed.

If the waste were linear in distance, then there might be some minor
benefit by having your mini-cities sit on the closer of their two 
tiles, but this amounts to at most shifting the center one tile inwards.

But waste, especially in the lower forms of government is non-linear.
This means that you lose more from tiles further out in an increasing
amount. Thus the weighting over all the mini-cities hurst ICS.

Also, if you lose 50% of a single shield, or however high you set the
roundoff threshold, the mini-cities can be completely wiped out, while
the single city will still produce at ~50%.

It hurts ICS as well in that large cities can buy cheap improvements
(1/10th the capital and ongoing costs in the above example) to reduce 
their corruption by factors of 2 or more. The big benefit from vertical
development is almost always asociated with improvements.

>Here comes a proposal: for corruption calculation, take "distance to 
>capital" to be number of the city in the list ordered by the real 
>distance.  Example:
>capital A,
>distances:  
>       B to A  5
>       C to A  7
>       D to A  4
>       E to A  10
>Then "corruption distance" of B is 2, of C is 3, of D is 1 and of E is 4.
>
>I am not saying this idea is very good, but it is maybe worth considering.
>
>G.

Cheers,
RossW
=====