v88 Jet's test games

Jet

No, no, please. Please.
Joined
Mar 16, 2006
Messages
2,419
I'm new to Civ 5 and started with some short test games to learn the game.

Then I played China, then Babylon, hoping for science victories, and both collapsed around 1000 AD. Babylon looked like it would be very strong, China too but less so.

The one other thing I noticed was that in some games there was no sign of the Huns spawning or having any effect, but I didn't look closely.

But about the collapses, when I saw that recaptured cities lost half their population and most of their buildings, I wasn't enthusiastic about playing with the expectation that a collapse would necessarily happen, and I would just have to deal with it. If you know the city will be trashed anyway, what buildings do you build? The play style didn't seem like it would be much fun.

There was not enough information in-game to understand what I could do differently to avoid the collapses, but they both had large negative economic stability at 1000 AD (-100, -150), so I read some of the stability code. Here's a summary of the economic stability code.
The economic stability score is the sum of two parts, EconomyBaseScore and EconomyModifier.

The calculations refer to
GNP = (total food) + (total production) + (gross gold) + (number of trade routes * 10).

EconomyBaseScore is recalculated from scratch every turn. It replaces last turn's EconomyBaseScore. It equals
(GNP * 5 / population)
+ (net gold, limited to range from -10 to 10)
- (2 * number of unstable civs you know)
- (number of turns your civ has been alive)/8, penalty limited to -35 (280 turns)
+ 20 if you're in a golden age
+ 10.

EconomyModifier is accumulated every turn. It is added to last turn's cumulative EconomyModifier.
It uses gnpChange, this turn's change in GNP as a fraction of the current GNP.
gnpChange = ((current GNP) - (previous GNP)) / (current GNP).
E.g. if gnpChange is 0.1, last turn GNP was 90% of the current GNP.
EconomyModifier = (gnpChange - 0.055) * 40.

Two things I noticed:

1. Of interest to players, economic stability includes a large penalty that accumulates every turn independently of anything the civ does. It looks like the penalty should be -0.055 * 40 = -2.2 per turn, although from the economic stability values I saw (e.g -150 at turn 300), I'm not sure how the penalty could be that high - I'm not sure how the growth terms could be large enough to cancel it. Nonetheless I think it's larger than anything the civ can do to counteract it. Over time, every civ is going accumulate a large economic stability penalty because of that fixed per-turn penalty.

2. Of interest to developers and very careful players, EconomyModifier seems to have the flaw that it rewards unsteady growth. For example you could double your GNP in 13 turns by growing 5.5% per turn, and accumulate a net EconomyModifier of 0. Or you could do it by growing 100% in one turn and 0% in 12 other turns, and accumulate a net EconomyModifier of 40 - (2.2 * 13) = 11.4. It follows from the same reason an elongated rectangle has a larger perimeter than a square of the same area. (Perimeter is analogous to a civ's accumulated EconomyModifier and area is analogous to its GNP.) I'm not sure how it works when some turns reduce GNP rather than increasing it, though.

I wasn't sure how to tweak the economic stability code to make it possible to play the long games I was trying to play - not while preserving any kind of challenge in economic stability, in the spirit of how the economic stability code was written.

I tried an Iroqouis game and got 3 units to start, which seemed like too little. It was fun to kill disembarked settlers before they could found cities, though.

I tried a Russia game and collapsed around 1000 AD again. I didn't read the logic for collapses closely, so I don't know exactly why it collapsed, but I saw the civ had problems with happiness (no problem, I can play to control that) and again economic stability.

I think to really change economic stability intelligently I'd have to understand GNP growth concretely: in some autorun games, log GNP, number of cities, and population, for every civ, every turn; then analyze it. For now, I think I'll try reducing the fixed per-turn penalty in EconomyModifier to 20% of its current value, and the GNP growth term in EconomyModifier to 40% of its current value. So the per-turn penalty will be both less consequential, and easier to counteract.
 
Last edited:
I played Russia to a Science victory in 1931 on Prince. It was medium-easy. I'm new to Civ 5, but Russia has such good land and bonuses.

Scores in 1931
867 Russia
696 India
516 Japan
503 Arabia
429 Germany, took Scandinavia
300s Native America, Aztecs, England, Spain
200s France, Austria, USA, Turkey

My score wasn't from killing units. I didn't have any wars or significant diplomatic pressure. Everyone was friendly to me except when Germany and Austria got upset about my buying tiles, but by then I was too strong to attack. But the role play, getting points by antagonizing other civs while also pursuing a science victory, sounds like an interesting challenge.

russia.jpg


I continued playing to 1943 when I researched Globalization, but I never saw the United Nations. I'm guessing Diplomatic victory is not enabled?

For this game I hacked GNP stability, as I suggested, to 20% of the original penalty and 40% of the original human effect. That might have made it too easy, but I couldn't tell. It's hard to relate the score to what you do in the game. I logged my GNP, but lost the log file.

Around 1700 my stability scores stopped changing: they were locked at -15, -5, 90, 86. I had no idea why. I noticed that someone else reported the same thing in an earlier version: "One thing I noticed is I was incredibly paranoid about the stability. (Also I'm almost positive it broke during the Ottomans game because the values were stuck at a pernament shaky state but the values itself were never changing for some reason throughout)."
 
I noticed that the game says you can't raze city-states because they're original capitals. I wondered if that meant that you had to capture every city-state to win Domination, so I played Greece to Domination on Settler (turn 137, before the Byzantium spawn), and I learned the answer is no, you don't have to capture city-states. I'm still a little unsure of exactly how "controlling every civ's original capital" is defined, given that rises and falls can happen in different ways, and capitals can even be razed.

Anyway, city-states work really well with RFC, I like it. However, I wish there was a better way to dominate a city-state militarily. If it's on a low-stability tile you have to either accept the stability penalty of owning a city on the tile, pay the city to make it your ally, or ignore the city and walk around it. You can't raze it or give it to a vassal like in Civ 4. Maybe if something like the Gunboat Diplomacy tenet was available to everyone.
 
  • Like
Reactions: PiR
Interesting, a CS territory could prevent a lot of actions indeed
@Rhye is it intended that the player cannot raze a city state?

@Jet I thought you could give cities you own? If you take the CS for yourself you still cannot give it?
 
Yeah! You can gift cities, and capturing a city-state and then giving it away might be better than just ignoring the city. But that would be less beneficial than making it an ally, giving it to a Civ 4 vassal, or in some cases, razing it.
 
  • Like
Reactions: PiR
Interesting, a CS territory could prevent a lot of actions indeed
@Rhye is it intended that the player cannot raze a city state?

@Jet I thought you could give cities you own? If you take the CS for yourself you still cannot give it?

It is from civ5 rules. They are treated the same as capitals. I don't know how to change that rule, except for special settings that I made for the Huns
 
  • Like
Reactions: PiR
Jet, since you are new to civ5 and you come directly from civ4 RFC, what are the main qualitative differences in your experience? Pls also try normal civ5 games. I would like to confirm that we are in the right direction to create some approach that is in the same ratio between RFC and standard Civ4.
 
It is from civ5 rules. They are treated the same as capitals. I don't know how to change that rule, except for special settings that I made for the Huns
I see, thanks for explaining.

Jet, since you are new to civ5 and you come directly from civ4 RFC, what are the main qualitative differences in your experience? Pls also try normal civ5 games. I would like to confirm that we are in the right direction to create some approach that is in the same ratio between RFC and standard Civ4.
Yeah, I've only played a few dozen turns of normal Civ 5. I don't have an analysis of the difference, just my subjective experience playing v88 for a week or two.
* long-term stability is a problem
* Civ 5 is noticeably simpler than Civ 4, and in some ways, less flexible
* I'm enjoying the novel features of Civ 5 like the Civ5 economy, the Civ5 versions of culture, religion, and espionage; city-states; and tourism. And 1UPT, although so far I'm indifferent to it.
* Diplomacy seems to be more opaque, and I don't like that. It looks like an earlier version of Civ 5 showed you the AIs' like/dislike factors toward you, but they removed it?
 
Last edited:
I tried to play Korea on King and collapsed around 1600 with -80 economic stability. I think there's a bug that always computes the human player's GNP growth as zero:

Code:
NewGameTurn()

    for the human player only,
        UpdateOnlyBaseStability()
            EconomyBaseScore() -- sets saved GNP to current GNP

    once every 5 turns,
        PrintStabilityOverview()
            for each player
                UpdateStability()
                    EconomyTotalScore()
                        -- This relies on EconomyModifier being called before the side effect
                        -- in EconomyBaseScore happens.  But for the human player,
                        -- NewGameTurn already called EconomyBaseScore once, so
                        -- EconomyModifier always calculates the human's GNP growth as zero.
                        EconomyModifier()
                        EconomyBaseScore() -- sets saved GNP to current GNP

A quick fix is to add a boolean parameter to EconomyBaseScore that controls whether it updates the saved GNP. UpdateStability calls EconomyBaseScore with that parameter set to true, and UpdateOnlyBaseStability calls it with the parameter set to false.

I thought the gnp-growth adjustment, for example -2.2 stability if your GNP growth is zero percent, was applied every turn, but now I think I understand that it only happens once every 5 turns, because PrintStabilityOverview() is where it happenes and that runs only once every 5 turns. Growing your GNP by 5.5% (the requirement to offset the penalty) every 5 turns rather than every turn would double your GNP every 80 turns rather than every 13 (*if* your growth is perfectly consistent from turn-to-turn; in Civ 5 it won't really be). That's less harsh than what I thought. But of course if the game thinks your GNP growth is always zero, that will add up a lot over time, and I think that's what I've been seeing.
 
Last edited:
  • Like
Reactions: PiR
Thank you Jet, I will look into that.

You haven't answered my question though! :lol:

Not about the existing bugs, not about how is Civ5 compared to Civ4, but my question was (specifically for you as a "freshman") if RnF is to Civ5 as RFC is to Civ4. And if not, to which extent.
 
You haven't answered my question though! :lol: Not about the existing bugs, not about how is Civ5 compared to Civ4, but my question was (specifically for you as a "freshman") if RnF is to Civ5 as RFC is to Civ4. And if not, to which extent.

It might be a while before I play much vanilla Civ 5 - for now I prefer to have fun playing your mod, to be honest - but a couple impressions so far:

1. RFC5-Civ5 are closer than RFC4-Civ4 because
1.1. Civ5 has more expansion limits than Civ4 - too many, in fact, so that the AIs don't want to colonize.
1.2. So far, RFC5 appears to have less impactful scripting - such as of wars, armies, and even plagues - than RFC4.

I imagine the RFC5 unit caps must make a noticeable difference, but I haven't played enough vanilla Civ5 to see.

2. I think the modern-age experience is a lot less refined in RFC5 compared to Civ5. It's not just smaller, it's not as carefully designed. Some obvious examples of that are The United Nations being absent (if I'm correct that it's absent), the Internet being absent but the Great Firewall still being present, and Solar Plants coming uselessly late. But I think there are subtler things that are also less refined than in Civ5, such as the tradeoffs between different research paths.

3. Heavy Plough is very noticeable for civs with grassland that don't start with it. For those civs, the introduction of the Heavy Plough tech in RFC5 makes a clear difference.
Because of that, and (I think) because of fewer resource tiles, and no ancient ruins, and more crowding, the RFC5 map feels more divergent from vanilla Civ 5 maps compared to Civ 4.
 
Last edited:
Good points!
Yes, late game is much smaller and for this reason I need much more insights on how to make it meaningful.
For example I was esecially careful to make sure that the unit upgrade paths would still work, but I need more tests to confirm that and more potential issues.
United Nations are not absent I think? I will check again.
 
I've played a few more test games, with the bug fix to economic stability that I described. It mostly eliminated economic stabiity as a factor I needed to think about.

India tends to be an AI with a high RP score. I won an India RP victory on Deity in 1852 as a side effect while playing for a science victory.

Germany can be good too. I won an RP victory as Germany on Immortal with 10 conquered cities in 1868, and in that India game, I saw the AI do the same with 12 conquered cities.

I'm not sure which other AI civs tend to have the best RP scores: maybe Arabia, Japan, or (with cities to the north of Mexico) The Aztecs. But I haven't seen any of them get close to 1000.
 
Aztec RP victory 1867, Deity

aztec.jpg


I think it's 9 victory points per kill, so 73 units killed.

Stability was 168 / 19 / 80 / -70 but I believe I had the bug where my stability values froze at some point. ... I wonder if that bug might be caused by saving and reloading in a certain way or at the wrong time, interrupting something so that it doesn't complete. It's just a hunch based on things I noticed half-consciously in games where it happened.

The AI barb scout units in North America don't attack (maybe they'll grab a worker?), they just stand there waiting for you to farm them for kills. Maybe if they were a warrior unit type instead, they would work OK.

My start was a little lucky: I got conquerors late, and they were distracted by barbs or Quebec or something and never came near me. I was turtling with no units exploring, I wonder if that helped. But I think the way I did it might also have worked with earlier conquerors, if not on Deity then on a lower level. I haven't actually done it though.
Spoiler opening moves :
* settle on the stone and on the silver 2N of it
* adopt tradition 0 / liberty 5 / rationalism 0. Don't use the third settler until after Construction
* drop an academy between the cities
* Tech AH -> Wheel -> Construction.
* There's good stuff besides just units (caravans, road tile, floating gardens) on the way to Composite Bowmen.
* Construction comes in about 1350, maybe 5 turns after the conquerors show up. Depends on luck.

After getting through conquerors, it was a straightforward game. I expanded, killed a lot of barbs (easy), teched to hospitals and hydro plants, killed the Maya at some point, switched over to military techs, then attacked the French who had the other half of North America. I believe the US never spawned, too much French culture or something.

The Civ 5 RFC Aztecs feel like less of an underdog than they did in Civ 4 RFC.
 
Last edited:
Top Bottom