Firaxis: Corruption Breakdown

[anarres]

I would like to say again that I love the FP as it is now!

Well, I'm almost in love. To counter the decreased effectiveness of the FP I think the OCN bonus should go from 10% to 40 or even 50%. This would make it much more powerful, and would still have the affect I am coming to love...

 

[alexman]

Yes, that would be a great solution.

And I think the Communism bonus should also be increased from 20%, so that it's still more than for the other governments.

A good solution would be a flag in the editor, in the governments tab, where you can set the FP OCN bonus for each government.

 

[planetfall]

Originally posted by Tavis

For the coders, here's the expanded section on what was posted earlier - we'll see how things go before dropping code bombs

// city ranking
if (cityRank < optimalNumberOfCities)
{
cityCorruption = ((shieldsOrTradeArrows * cityRank) + 1) / 2;
}
else
{
cityCorruption = ((shieldsOrTradeArrows * ((cityRank * 2) - optimalNumberOfCities)) + 1) / 2;

This cityCorruption variable does not make sense to me. It seems like something is missing. It might be my head! How does corruption formula work if more cities than ONC/ONC number?

Sample data:

LESS THAN EXAMPLE city 9 of 20, shields == 10
(( 10*9)+1) /2
(90+1)/2
91/2 --> 45%


MORE THAN EXAMPLE city 29 of 20, shields == 10
(( 10 * ((29*2)-20)) +1)/2
(( 10 * (58-20)) +1 /2
(( 10 * 38)+1 )/2
(380+1)/2
381/2 --> 140%

What keeps corruption below 100% when city rank is greater than ONC?

Or does "cityCorruption" not output rate of corruption but rather available shields?


== PF

 

[SirPleb]

Originally posted by planetfall
What keeps corruption below 100% when city rank is greater than ONC?

That definitely isn't in this part of the code. Somewhere outside the code we've seen there must still be an allowance made for a number of things including courthouse/policestation, distance corruption, WLTKD. I think that after all that is done there's probably a test somewhere which in effect means "if > 95% use 95%". The code we've seen handles just the rank portion of the corruption calculation.

There does seem to be one thing in the code when rank >= OCN which is new to us. (I.e. which Alexman didn't find in his study, which previously defined all we knew ) The subtraction of " - optimalNumberOfCities" in the code for the > OCN case isn't in Alexman's work. It does seem like a good thing though. The effect of that subtraction should be a smooth graduation from city OCN-1 to OCN to OCN+1 etc., instead of a very sharp jump in corruption from city OCN-1 to city OCN.

 

[planetfall]

Interesting. I have forgotten, what does '->' mean?

So in RG, real game, assuming ranking is recalculated each turn. Is there any difference in the ranking of the top 10 cities between Play Style A which builds 20 cities 2 tiles apart and then prunes out cities later to 4 tiles apart and Play Style B which builds 10 cities 4 tiles apart. In order words, will the date of founding have a different effect on these two play styles and yield different city rankings?

If I am reading the code right, I don't think there will be difference. But I am not positive that is the correct interpretation.

Yes the '-OCN' is great info. The '+1' seems to smooth the plot and of course block any div/null errors.

== PF

 

[DaviddesJ]

Originally posted by SirPleb
There does seem to be one thing in the code when rank >= OCN which is new to us. (I.e. which Alexman didn't find in his study, which previously defined all we knew ) The subtraction of " - optimalNumberOfCities" in the code for the > OCN case isn't in Alexman's work.

Yes it is. It appears as "- 0.5" in his formula "Fc = Ncity / (Fn*Nopt) - 0.5" here.

The effect of that subtraction should be a smooth graduation from city OCN-1 to OCN to OCN+1 etc., instead of a very sharp jump in corruption from city OCN-1 to city OCN.

There's never been a sharp jump.

 

[Qitai]

-> is a pointer.

The new code will give different city different ranks instead of same rank. This solve RCP. If I read the code right. The priority of the ranking is
(1) The one with the shorter distance comes first
(2) If same, the one with the earlier date found comes first
(3) If same, the one higher up in the database ID comes first (This one is probably the founding order)

 

[SirPleb]

Originally posted by DaviddesJ
Yes it is. It appears as "- 0.5" in his formula "Fc = Ncity / (Fn*Nopt) - 0.5"

Oh right, I misread the effect of that - 0.5. It does come to the same thing as the code, assuming there's a subsequent division by OCN in some following code (which there must be.)

 

[SirPleb]

Originally posted by planetfall
So in RG, real game, assuming ranking is recalculated each turn. Is there any difference in the ranking of the top 10 cities between Play Style A which builds 20 cities 2 tiles apart and then prunes out cities later to 4 tiles apart and Play Style B which builds 10 cities 4 tiles apart. In order words, will the date of founding have a different effect on these two play styles and yield different city rankings?

No, I expect those two cases to come out identically.

 

[bradleyfeanor]

The technobabble has given me a tremendous headache. Deciphering Civ corruption is becoming more work than my job.

Question:
If I build my FP near my initial palace, and then move my palace later, does the closest city to my new palace get a new corruption rank of 1, the second closest city 2, etc.? In other words, when I move my palace does the center of the rank corruption layout move along with it?

I suppose such a model would be ok, but I would like to know, because the old strategy of building the FP in an initial core city and then jumping the palace to start a new core would definitely be counter productive. The initial core cities would be crippled. This would seem to bring another "luck" factor into the game, because in this model I would want a FP in a remote region and never my Palace. That means getting a leader at the right time to build it in most cases.

 

[Qitai]

Just let those who like to probe details worry. Give yourself a break.

Answer to your question is yes, the rank is re-calculated.

For V1.12, a general rule would be to put it near the palace, in a city which is not already fully corrupted.

 

[planetfall]

Qitai,
Thanks. I suspected it was just a pointer, but I didn't know if it had some other special meaning. I was confused by the logical naming by type. Thus pointerLoopCity pointer getWho function seemed redundunt.

SirPleb,
Thanks for the confirmation. BUT it looks like we have to pay some attention to the order we plant cities. Assuming cities are the same distance and started on same date, I don't see any method to manipulate the ranking of city #6 so it moves ahead in rank of city #5 once we have built cities. Say city #6 had gold and ivory, it would be helpful to move that city up in the ranking. Naturally pruning cities after capital of 1 to 4, would move it to rank 3. But unless you abandon city#5 and rebuild it, city #6 will never move to rank 2.

Agreed DateOfFounding would not be likely to change the rank BUT the order in which cities the same distance are initially ranked should carry forward. Thus using old RCP at 4 tiles N,W,E,S. Building cities at 4N and 4S a turn before 4E and 4W will keep those two cities a higher rank, i.e. lower number, than cities at 4E and 4W.

== PF

 

[SirPleb]

Originally posted by planetfall
it looks like we have to pay some attention to the order we plant cities. Assuming cities are the same distance and started on same date, I don't see any method to manipulate the ranking of city #6 so it moves ahead in rank of city #5 once we have built cities. Say city #6 had gold and ivory, it would be helpful to move that city up in the ranking. Naturally pruning cities after capital of 1 to 4, would move it to rank 3. But unless you abandon city#5 and rebuild it, city #6 will never move to rank 2.

Agreed DateOfFounding would not be likely to change the rank BUT the order in which cities the same distance are initially ranked should carry forward. Thus using old RCP at 4 tiles N,W,E,S. Building cities at 4N and 4S a turn before 4E and 4W will keep those two cities a higher rank, i.e. lower number, than cities at 4E and 4W.

It is true that we won't be able to control the ranking that finely and that it makes a bit of difference. But personally I doubt I'll ever worry about small ranking differences among my stronger cities. It isn't a big deal to me if reordering a few cities, swapping them up and down the list a bit, might give a slightly greater net yield.

And the only solutions I can think of for this remaining problem seem like more code than they're worth. It seems better to leave a small issue than get into significant recoding.

 

[Grille]

Originally posted by SirPleb

It is true that we won't be able to control the ranking that finely and that it makes a bit of difference. But personally I doubt I'll ever worry about small ranking differences among my stronger cities. It isn't a big deal to me if reordering a few cities, swapping them up and down the list a bit, might give a slightly greater net yield.

Agreed.
This would only be a (rather small) issue if one of the two (or more) equidistant cities would fulfil the 'rank < OCN' condition and the other one the 'rank >= OCN' condition and if one of those cities had a really big commerce/shield advantage (e.g. 1 city on gold in the middle of a certain lux resource cluster and the other one(s) in the mere desert or so).
Possibly not a very common scenario - thus not worth to implement a more complex code.

 

[socralynnek]

Originally posted by planetfall
Thanks for the confirmation. BUT it looks like we have to pay some attention to the order we plant cities. Assuming cities are the same distance and started on same date, I don't see any method to manipulate the ranking of city #6 so it moves ahead in rank of city #5 once we have built cities. Say city #6 had gold and ivory, it would be helpful to move that city up in the ranking. Naturally pruning cities after capital of 1 to 4, would move it to rank 3. But unless you abandon city#5 and rebuild it, city #6 will never move to rank 2.

Agreed DateOfFounding would not be likely to change the rank BUT the order in which cities the same distance are initially ranked should carry forward. Thus using old RCP at 4 tiles N,W,E,S. Building cities at 4N and 4S a turn before 4E and 4W will keep those two cities a higher rank, i.e. lower number, than cities at 4E and 4W.

== PF

I have a solution for this. Why use the ID for the ordering of cities at the same distance and the same founding date. Why not use the name? That's something the player can control, so he can chosse which city has a better rank and the casual player would never mind.

 

[socralynnek]

Originally posted by anarres

Well, I'm almost in love. To counter the decreased effectiveness of the FP I think the OCN bonus should go from 10% to 40 or even 50%. This would make it much more powerful, and would still have the affect I am coming to love...

With my solution, there would be no need for an OCN bonus, because if you build the FP in a region above the OCN you get a new core with productive cities that is as big as the old core, that means OCN increase by 100%

 

[Skyfish]

Originally posted by anarres

I think the OCN bonus should go from 10% to 40 or even 50%

One day these words will be looked upon as a "visionary statement"

 

[Nad]

I don't know if this has already been posted (if it has, apologies), but I think there's a very simple solution to the palace jump exploit: everytime a palace is moved without a new palace being built by shields, the civ should be plunged into anarchy.

So, if you move your palace the legitimate ways (by handbuilding a new one or using a leader to rush it), you have no anarchy because you had a full shield box for the palace...but...

if your palace is moved without a new one being built (because your capital is captured or disbanded) you are immediately plunged into a normal anarchy.


Simple, yet effective?

 

[Grille]

The main problem with a Palace jump (or for the record: rush) in vanilla/PTW is not the fact alone that you could just move your capital for a cheap cost, at least not in the first place. The real problem is incorrect rank assignment. So you could use the disband capital 'feature' (if you'd call it like that) to get a remote, isolated capital city: that way, you could get low-ranked cities around your FP. Buggy rank allocation was the root of the matter.

The discussion about banning or penalizing Palace jump sounds like if that 'feature' alone is the devil which is not true IMHO.
Heck, let's assume that getting a new Palace was only possible by GL-rushing and the rank system was still not fixed - I would almost ever try to get a GL and rush the Palace in a remote area.

I think there's nothing wrong with getting a new Palace by disbanding the capital - if ranks get (re-)calculated correctly.
Why should there be a penalty like forced anarchy when the former capital is lost? I see exploit-loopholes with such proposed penalty concepts.
What would happen if I'd conquer AI capitals - does the AI change to perm-anarchy in case I make a capital chase? Or, if losing the capital in an act of war was excepted from penalty - maybe I could then "lure" the AI to my undefended capital to get rid of a penalty?

 

[Megalou]

Originally posted by Nad
I don't know if this has already been posted (if it has, apologies), but I think there's a very simple solution to the palace jump exploit: everytime a palace is moved without a new palace being built by shields, the civ should be plunged into anarchy.
...
...if your palace is moved without a new one being built (because your capital is captured or disbanded) you are immediately plunged into a normal anarchy.
...

Good thinking! This feels realistic and fair, even though what is realistic is not necessarily most playable...

One little "but": if anarchy ensues when a capital is captured, things will really go from bad to worse for the victim. With no production for up to 8 turns, only a culture flip is likely to give him/her the capital back. The attackers will even have time to regroup and hit another city. So this detail will probably hurt the AI too much, since the human capital, I believe, is more seldom captured. I call this a detail because if an AI civ loses a capital it is usually doomed anyway. There may however be cases when it is not so, like when you attempt Right of Passage abuses against strong AIs.

EDIT: Grille, it could work if there was a distiction between abandon and capture/raze. The "luring" business aside...