C3C 1.13b - sinple FP & Corruption Test (Graphics intensive)

[Yeti]

That's very odd.

Distance for distance corruption is still being calculated using nearest capital, so FP position does still very much matter.

The question is, is the code correct, or are the patch notes correct?

FWIW: The way it is coded now seems pretty balanced to me (assuming that having it work similarly to PTW means it is balanced). Fixing the duplicate rank issues meant that something else needed to be done to reduce city rank / OCN corruption, which they did (FP increases OCN more than it used to). Removing the effect that the FP has on distance corruption would have made the FP weaker (and given the AI a boost).

 

[Arathorn]

It appears to me that corruption now works somewhat like this:

- Distance is calculated from nearest capital building for distance corruption
- Rank is still determined by distance from the one true capital for rank corruption
- The FP has some special rank or coding to keep its corruption minimal
- OCN is bumped way up by the FP to mask the effect of the rank corruption if the FP is "near" the capital

If I'm right (or close), this has at least the following repercussions:
- Optimal distance/# of cities between FP and capital will exist (I have no clue what it will be)
- Building a fully corrupt city can now worsen an old city of similar distance (old city is near FP and suffers minimal distance penalty and reasonable but high rank penalty sees its rank penalty increased and the new city has such bad distance penalties that it can't take advantage of its better rank, being in an opposite direction from the FP). This is something I don't like, personally.

Can someone test to see if my hypotheses are accurate?
Arathorn

 

[TedJackson]

Looking at the FP ring Cities (all at distance 4/4.5 from the FP) listed in reverse creation order (so presumably highest ranked first we get:

Nagsu = 4-1spt, 14-5gpt = 20% waste, 36% corruption
Kutha = 7-2spt, 11-4gpt = 28% waste, 36% corruption
Kuara = 5-2spt, 16-6gpt = 40% waste, 37.5% corruption
Marad = 9-4spt, 10-4gpt = 44% waste, 40% corruption
Kua = 11-4spt, 11-3gpt = 36% waste, 27% corruption
Der = 6-2spt, 17-6gpt = 33.3% waste, 35% corruption
Kisurra = 8-4spt, 18-9gpt = 50% waste, 50% corruption
Zabalam = 6-3spt, 16-7gpt = 50% waste, 43% corruption

Kutha is the closest city to the Capital and Nagsu the furthest. Which doesn't seem to fit with with either Rank or Distance corruption calculated purely from the Capital

Perhaps I need to increase the tile Shield & Commerce values and make a pure BG map and try again to get better data to work from


Ted

 

[Arathorn]

How did you determine which city is closest to the capital? With your given coordinates in another thread (19,61) and (71,55) on a 100X100 map, the city nearest to the capital will be WEST of the FP (left on the map), to the northwest, around (17,58) or so, as I count things. Since you were at almost exactly the break-even point between east and west, determining which city is closest to the capital (by wrapping or whatever) is going to be very difficult.

I suggest you check again, as the sheer randomness appearance of your numbers makes me think something else is going on. I suspect the wrapping issue. Can you post exact locations of each city? Or test true distance from the actual capital yourself?

Arathorn

 

[TedJackson]

Good point Arathorn,

I'll go and have a bite to eat and then rework this. I only started this to get a quick idea of what was happening but it's getting interesting now


Ted

 

[anarres]

Originally posted by Arathorn
It appears to me that corruption now works somewhat like this:

- Distance is calculated from nearest capital building for distance corruption
- Rank is still determined by distance from the one true capital for rank corruption
- The FP has some special rank or coding to keep its corruption minimal
- OCN is bumped way up by the FP to mask the effect of the rank corruption if the FP is "near" the capital

If I'm right (or close), this has at least the following repercussions:
- Optimal distance/# of cities between FP and capital will exist (I have no clue what it will be)
- Building a fully corrupt city can now worsen an old city of similar distance (old city is near FP and suffers minimal distance penalty and reasonable but high rank penalty sees its rank penalty increased and the new city has such bad distance penalties that it can't take advantage of its better rank, being in an opposite direction from the FP). This is something I don't like, personally.

Can someone test to see if my hypotheses are accurate?
Arathorn

This is accurate. I like the new system btw. It reduces the benifit of building the FP on the far side of the world and still allows a moderately far-away FP to be a very useful corruption fighting measure.

In my current game with my FP build 10 tiles from my capital I got an increase in income from 480 to 535.

 

[The Last Conformist]

I'm too tired to think this true ATM, but am I right that bigger map means bigger optimal P-FP distance since the OCN is higher?

 

[Roland Johansen]

If Arathorn is right, then yes.

 

[akots]

Originally posted by The Last Conformist
I'm too tired to think this true ATM, but am I right that bigger map means bigger optimal P-FP distance since the OCN is higher?

Not true, because rank corruption does not depend on map size. Considering OCP-placed cities, if FP is placed too far away, rank of the surrounding cities would be too high to compensate for decrease in distance component. Maximum benefit also is unrelated to OCN. More research is needed to determine location with maximal benefit.

 

[TedJackson]

Here's what I came up with to try and clarify what's actually happening.

Test Criteria:
Standard Map (100*100) all BG (modified to 3 food, 10 Shields, 10 Commerce)
Place Ur (Capital) at 25,49
Build ring, starting from 12 O'clock and moving clockwise, at distance 4/4.5
Sumer, Lagash, Kishm Umma, Bad-tibira, Agade, Erech & Isin
Place FP at 49,49
Build ring, starting from 12 O'clock and moving clockwise, at distance 4/4.5
Zabalam, Kisurra, Der, Kua, Marad, Kuara, Kutha, Nagsu
Give each city 2 Spears. Give the Capital & FP site 2 leaders.
Make each city pop 10 & give each city 100 culture points.
Place an extra settler 2SE of Capital & FP site
Place an AI start at 75, 25
Add Resource colonies (1 of each)


Start game as Sumeria, 1 AI opponent, Regent level, default rules, AI agression normal

4000BC:
Set sliders to 10.0.0
Fortify all units in place
Set all builds to wealth except Akshak (set to FP)
Save game
Wake leader in Akshak and hurry FP

Press button...

Note results

Before FP

After FP

The ranking comes from sorting the cities by production in the F1 screen.

In particular I don't understand the result for Zabalam, which at 3 tiles North of the FP has unexpectedly risen in (effective) rank, and Kua (3SE of FP) which has fallen in (effective rank).

It's good to see that the FP site isn't 2nd in the list after the FP has been built.

The 90% cap is also quite visible in the pre-FP listing (although at slightly less than expected.


Ted

 

[Roland Johansen]

Originally posted by akots


Not true, because rank corruption does not depend on map size.

That is not true, at least it wasn't true in PTW and Vanilla Civ3.

A rank 5 (just picked a number) city on a huge map suffers less rank corruption than on a small map (see Alexman's corruption formula for PTW 1.27f)

 

[MrWhereItsAt]

Question - is this 1.13b patch different to the 1.13 I have seen so far? Or is it the 1.14 patch reported elsewhere?

 

[TedJackson]

I just stuck the b on to indicate beta.

Not heard of 1.14.


Ted

 

[Dr Elmer Jiggle]

The reason the palace ring cities improve is because of the increase to the OCN. If you look at alexman's corruption formula, rank corruption is essentially Ncity / (2*Fn*Nopt). Since the FP increases Nopt by 50%, rank corruption will decrease. That means that any city where rank corruption dominates (i.e. a city that's very close to the palace) will improve.

 

[TedJackson]

Thanks Jiggle,

I'd forgotten the increase to OCN


Ted

 

[Dr Elmer Jiggle]

Also, the cities that unexpectedly rose in rank didn't really rise in rank. They were tied in the "before" data, so they're sorted alphabetically. They could realistically be in the same order in both cases.

 

[Dr Elmer Jiggle]

Originally posted by Dr Elmer Jiggle
they're sorted alphabetically

Or randomly if you use a normal alphabet.

 

[TedJackson]

Never mind


Ted

 

[akots]

Originally posted by Roland Johansen
That is not true, at least it wasn't true in PTW and Vanilla Civ3.
A rank 5 (just picked a number) city on a huge map suffers less rank corruption than on a small map (see Alexman's corruption formula for PTW 1.27f)

It might be they fixed it but it did not depend in 1.12b: here

 

[akots]

Originally posted by TedJackson
Here's what I came up with to try and clarify what's actually happening.

Your system is too complicated to analyze it normally. To see the true effect of rank on corruption it is necessary to subtract distance corruption from all values and not from 90% but from the true-over-90% number which is used for calculations and which you don't know. Since some of the cities of the RCP-versus-FP before FP was built were at different distances than the RCP-versus-Palace cities, distance component would be different for each of them. And this must be also calculated. Either this is incomplete and may be wrong or, I did not understand the test. There are slightly more variables in this than it seems.