Everything about Corruption: C3C edition

[alexman]

Originally posted by Arrian
Am I right. Or at least close to right?

Yes, that's the way it works, although your actual number is a bit off because Monarch has 85% of the OCN, not 90%.

Is anyone interested in the industrial era save, to see how the FP area is doing when more built up?

That would be interesting, although if you also had a spare leader as well, so we can rush the Palace at different locations to see the effect, it would be perfect!

[Arrian]

Ah, .85 instead of .9. So 36.5. Does it round up or down? I'd guess down, so 36 cities.

I don't have a leader, no. But that is something that can be remedied

Not easily, since I have 3-4 elite units, all knights, all of which are on a galleon sailing to a war that's over now (English "sneak" attacked me with the famous "drop off 1 attack unit next to a human city" move, so I took their fur town away from them).

But I could smack down the French and try for leaders (made easier now because I have replaceable parts -> artillery). Or maybe I'll get a SGL at some point (none thus far in the game).

-Arrian

[Harovan]

Originally posted by Arrian
(English "sneak" attacked me with the famous "drop off 1 attack unit next to a human city" move, so I took their fur town away from them).

Reminds me the old days of Colonization, when some AI civ dropped a rifleman right next to my city and tried to extort me, declaring war when I told him to bugger off. I attacked the rifleman then. Since in Colonization units didn't get killed, but only disarmed, I suddenly had a veteran Statesman just to capture. A great asset, because they were so incredibly rare in the early game and very expensive to train in Europe, and very useful in the race for more founding fathers.

[Alex]

thanks for the clear explanation, alexman.

[Arrian]

Never played Colonization, SR. I get the feeling I would have liked it.

-Arrian

[planetfall]

First of all, Alexman thank you for the analysis. Some questions about your initial report:

For corruption purposes, distance is always an integer number, and is given by:

code:
d = max(x,y) + min(x,y)/2
where
max(x,y) denotes the maximum between x and y,
min(x,y) denotes the minimum between x and y,
x is the distance in the NW/SE direction,
y is the distance in the NE/SW direction,
and the integer division is rounded down.

So X == distance at -45 degrees from vertical
and Y == distance at +45 degrees from vertical

If formula was exactly as you posted, it would seem best to have Palace at north pole to get negative number for distance. Isn't actual formula more like:

d= max(int(x), int(y) ) + min(int(x), int(y))/2

In the special case of communal governments, the distance for all cities is taken as:
code:
d = (MaxD) / 4
where MaxD = (MapW+MapH)/4, MapW is the width of the map, and MapH is the height of the map, as given in the editor.

The distance, d, is used to get the adjusted distance, da, as follows:
code:
da = 0.5^Ni * min(Gd * t * d, MaxD)

Is MaxD the same for all gov'ts? If not what is the value for noncommunal govt?

Finally, what is your recommendation for city layout to maximize production under non-communal govt?

An example of a good use of planning to minimize corruption/waste and of not so good planning would be helpful for understanding the impact of the formulas.

Again, thanks.

PF

[alexman]

Originally posted by planetfall
If formula was exactly as you posted, it would seem best to have Palace at north pole to get negative number for distance.

There is no such thing as negative distance. My x and y are distances, not coordinates. Also, I think you're confusing int with abs.

Is MaxD the same for all gov'ts? If not what is the value for noncommunal govt?

MaxD is just a number that depends on map size. It's half the average of height and width, and doesn't depend on government.

Finally, what is your recommendation for city layout to maximize production under non-communal govt?

See Arrian's example posted above for a good city layout, or visit the AU and strategy forums.

[planetfall]

Wow, are you fast.

I took min/max as functions not distances, thanks for clarifying.

I thought MaxD was the same for all, but wanted to clarify as it initially sounded like that part of the definition was only for communal govts.

I saw Arrians example, but didnot have time to load and view yet.

I find AU confusing. Which AU would you recommend?

Thanks again.

PF

[Arrian]

I think alexman just means reading some of the posts you will find in AU "AAR" or "DAR" threads (AAR = after-action report, DAR = during action report. They're essentially the same thing, but DARs have become the standard - people report in after they've played up to a certain point in the game, either measured by the year or by acheiving some specified goal).

I haven't read the threads on the ongoing AU game, since I plan to play it soon, but I'd imagine there is a fair amount of analysis in there. Possibly corruption related, since the 1.15 patch and the corruption system it contains is still fairly new to most players.

-Arrian

[planetfall]

Arrian,

Thanks for the definitions. Guess my problem was conceptually I was thinking of AU like War Academy and the model is different. It seems like AU is just a series of games and tread comments on them.

What I miss from AU is organization around a theme. I think AU is currently playing a C3C game, but I don't understand how this is different than looking at treads from GOTM.

Thus, I guess from your definitions I should read the DAR threads.

PF

[Arrian]

Well, I'm not all that familiar with GOTM, but I would guess that AU is less competitive - there is no scoring or anything like that. The express purpose is to play, have fun, and learn. Hence the detailed AAR/DARs.

There is also an AU mod. The AU mod is aimed at promoting stategic choice & improving the AI without making drastic changes to the stock rules. Some people play AU games with the mod, some play with stock rules.

-Arrian

[Tall Stranger]

Actually, AU is built around a theme. For example, the current one is "Power of Seafaring." Others have included one-city challenges, the power of communism, and total eternal war scenarios.

AU differs from GOTM in that there is no one "winner" of the game. Rather, the idea is to learn something from every game played so as to improve your understanding of/ skill at the game. The DARs and AARs are a crucial part of that. Hearing how someone decided to take a completely different tactic from you can be very educational. Additionally, since this isn't a competition, you have the freedom and incentive to try out new approaches, since there's no "penalty" for losing.

It's a lot of fun, and a great way to improve your game. Give it a shot.

[planetfall]

Tall Stranger,

You have valid points. I dropped GOTM and QSC because of the time involved and the funky scoring system. Civfan has more of a competition tone to the forums and poly has more of a cooperative tone. That's why I am on both forums. I haven't tried AU yet. I plan to complete my current revision of my opening game first and then will take a look at AU for refining.

PF

[Arrian]

Is anyone interested in the industrial era save, to see how the FP area is doing when more built up?

...

That would be interesting, although if you also had a spare leader as well, so we can rush the Palace at different locations to see the effect, it would be perfect!

Well I played on and on and on and I finally got that damned leader. I generated a SGL upon discovering Steel. I didn't have time last night to post the save, but I will early next week.

Trying for leaders was an excercise in frustration, I must say. I waited until my trade deals lapsed (I have this thing about breaking deals with the AI. I don't do it to them, only them to me) and attacked France. I figured with Infantry, Artillery (a lot of it) and Cavalry, I'd roll right over them, and have lots of victories for my 4 elite knights & 3 elite ancient cavs against 1hp frenchy units.

It didn't exactly go that way. A combination of the new-and-improved Muskteer, some bad luck (particularly with my elites hitting those redlined units... they did terribly!), and the sheer size of the French army (they built Knights Templar and at one point put together a formidable stack of musketeers and crusaders) meant the war was longer and more frustrating than I expected. I've broken them, but it wasn't a cakewalk.

Plus, the very turn I invaded France, England decided to move on the fur city I had captured from them the last time Lizzie messed with me. I was never in danger of losing the city (knights attacking infantry in a walled town), but I didn't have the forces to go on the attack, since I needed all the Cav I had to deal with the French. Speaking of the French, they dropped off a landing group next to the aforementioned fur city, and with everything going on, I forgot to wake up my Cav and kill it off. The result was them cutting my only fur supply, which knocked 1/2 my empire out of WLTKD.

All in all, minor irritations, but it was adding up. Then, despite two simultaneous wars, how do I get the leader? SGL. Figures.

Anyway, I'll post the game when I get back from skiing, and hopefully people will find it interesting to move the palace around and see the results.

-Arrian

[Arrian]

Here is the save. The SGL is in Salamanca.

-Arrian

[alexman]

I took a quick look.
Nice looking empire!
It certainly doesn't seem like it's worth burning a leader to relocate the Palace at this point.

Moving the palace to the north, where I would suspect it's going to be better in the long run (after you have developed the northern cities better), increases corruption. The problem is that you have more population to the south, so moving your palace to the north increases the city rank for a lot your population.

I think a Palace where your FP is now, and a FP to the north would be ideal for this empire.

[E_T]

We also recently got some inside information (a look at some of the actual corruption code), which helped find the last few missing pieces of the corruption puzzle.

Alexman, You are truely the Corruption Nazi!!!

But my question is, with this new information as to the missing peices of the previous corruption puzzle, does this have any effect on the older (Vanilla & PTW) version calculation models?? In other works, does that other corruption thread need to be updated?

Also, I know in PTW, there is a slight difference between a PBEM game and an SP game. Have you been able to quantify that yet? And has anyone assertained if the same applies to C3C??

E_T

[alexman]

Off the top of my head, things that have changed in C3C:

1. The FP does not affect rank.
2. The FP gives a higher OCN increase than before.
3. Communist gets a higher OCN increase than before.
4. Courthouses etc reduce maximum corruption
5. Specialists

The PTW corruption thread appears to have different numbers than this one, but it's mostly because I normalized the government coefficients by Despotism corruption and the trade coefficient by unconnected cities, whereas here they're normalized by Monarchy/connected cities.

I have heard you say that corruption is different in PBEM and SP before, but I have not seen it myself, and I have certainly not attempted to quantify the difference. If it happens, I doubt it's a difference in the actual corruption model, as there is no distinction in the code between SP and PBEM. My guess would be that the PBEM games pick up the wrong OCN (for a different size map).

[Arnelos]

Once again, great work on this.

Man, what on EARTH would we do without you strat forum guys on our site to help us out with stuff like this?

[robinguo]

The sum of the distance and rank components of corruption can be greater than 100%. The game sets an upper limit for the total corruption in a city to be equal to 90% minus 10% for each city improvement and minus 70% for each Wonder in the city with the “reduces corruption?ability:

Great works! Alexman!!
But may I ask a question about the upper limit,
does the "upper limit" actually worked in any situation?
is it means we can build courthouse and police station in
every city , and reduce the final corruption to 70%?

[player1]

Originally posted by robinguo But may I ask a question about the upper limit,
does the "upper limit" actually worked in any situation?
is it means we can build courthouse and police station in
every city , and reduce the final corruption to 70%?

exactly

[Nor Me]

I've just done a quick test for it and I think that all the OCN increases apart from those from courthouses, police satations and WLTKD are affected by difficulty level That would make the formula something like:

Nopt = max(OCN * (L/100 * (1 + c + Gr + Gp*Nwe) + 0.25*Ni), 1)

So courthouses really are more effective at higher difficulty levels but the commercial trait is not.

I'm not sure where all the rounding happens in the parts we haven't seen the code for but it can be significant. In PTW, the the OCN increase for the commercial trait was rounded down so if the OCN modified by difficulty level was just below 12, the increase would only be 2 and the corruption reducing ability of commercial would about a third less effective than it should be.

Another thing, there's no mention of WTLKD reducing disance waste. I've not tested it but I've seen it in the code.

[alexman]

Originally posted by Nor Me
I've just done a quick test for it and I think that all the OCN increases apart from those from courthouses, police satations and WLTKD are affected by difficulty level That would make the formula something like:

Nopt = max(OCN * (L/100 * (1 + c + Gr + Gp*Nwe) + 0.25*Ni), 1)

If you are certain of this, I'll update the initial post. Thanks. I thought I tested for it, but I could have made a mistake.

Another thing, there's no mention of WTLKD reducing disance waste. I've not tested it but I've seen it in the code.

Whoops, you're right of course.

[Nor Me]

I'm not certain yet, but it certainly doesn't work in the way your formula suggests. I know that in my test, a courthouse reduced corruption in one city by a factor of 6, for a commercial civ with an FP in Democracy when I modded regent to 1% of the OCN. That would seem to suggest that all these are affected by difficulty and courthouses aren't.

As it was a standard map, it might be that the modified OCN was 1 before the courthous which increased it by the expected 20/4.

I did tests with PTW that gave exact values and was surprised by your formula given the values I'd seen. Notably, the commercial rounding didn't seem to have any other explanation. But then I was surprised that the modified OCN was an integer so they can't have been that precise.

As I didn't look at exact numbers, there's a slight possibility that it's all one big rounding error for such an extreme value.

[fishleaf]

I suspect the exactness of the formula ALEXMAN posted.

[vulture]

It's not exactly exact - there are some rounding issues that don't appear in the formula. One that I remember from vanilla civ3 (don't know if it was changed in PtW or C3C) was that distance corruption, expressed as a percentage, was always an even integer. You could have 44% distance corruption, or 46%, but nothing in between, for example (possibly that was an artefact of some other curious rule, for all I know). And of course the distances are produced with a fairly odd formula as well, rather than the actual geometric distance between the cities.

But the formula doesn't need to be exact - what value does it have in the game to calculate corruption exactly, independent of the computer. The actual uses of the formula - determining where to put your FP, where to build courthouses and police stations, how to space cities - all these come from the overall form of the equations, which I'm confident that Alexman has nailed pretty accurately; they're not adjusted by the precise numerical details at the rounding-off-error level.

[alexman]

For the record, Nor Me is right that courthouses are more effective at higher difficulty levels but the commercial trait, better governments, and FP are not.

I already updated the formula when he posted his correction, and in the meantime I have verified that it is correct.

I will not even attempt to address rounding issues. It's too tedious to test, and as vulture says, we don't have much to gain from an exact formula.

[Aqualung71]

I'd like to weigh in with support of Alexman, Vulture and Nor Me here. I've taken a keen interest in the corruption calculations and have digested Alexman's calculations and formulas and applied them rigorously to "in-game" situations by manually calculating theoretical corruption percentages on a city-by-city basis, and I can confirm categorically that the formulas work very well without exception within quite small margins for error that can easily be explained by rounding differences.

I suspect that the adjustment proposed by Nor Me is probably correct, and I have no doubt that either way, the formulas give us a substantially correct estimate of corruption in all cases, with insignificant and very acceptable rounding errors within the context of the game. Trying to go any further and making the formulas absolutely precise serves no useful purpose in my view.

Again, for the exceptional work Alexman has put into unravelling the corruption mystery and kudos to Nor Me for spotting the minor amendment......as Vulture says, there is no value in trying to take it further at this point.

Great stuff guys!

[Rommel2D]

Originally posted by alexman
I will not even attempt to address rounding issues. It's too tedious to test, and as vulture says, we don't have much to gain from an exact formula.

Sure, elect a guy Emperor and he starts to neglect the kingdom. ;-)

[alexman]

Emperor?
Hey, I got promoted 15 posts ago, cool!