Advanced Civ

f1rpo · Aug 23, 2020

crullerdonut said:
In the other game that I recently played with AdvCiv, I won by a Domination victory. So, I definitely had to deal with the revolt system.

Thanks for looking into it.

After conquering a city, I kept moving my stack, only keeping 3-4 defenders on the city, but not enough to actually reduce the Occupation Timer in a reliable way. This way, the conquered cities stayed in resistance and didn't cause crippling Maintenance costs. (This is an interesting mechanic: you can selectively bring cities out of revolt in a way which eases the (possibly massive) increase in Maintenance costs.)

One can of course also delay the occupation countdown in BtS by leaving a conquered city without any garrison. Avoiding starvation could also be a reason for doing that. But I guess it's true that the possibility of revolts once occupation has ended creates an extra incentive to keep cities under occupation until a sufficient garrison (and maybe a Missionary) is available. Unlike in BtS, once peace has been made, revolts can occur even during occupation (with a reduced probability) – so that players can't keep a city in occupation indefinitely. Perhaps this rule change should also apply when at war. This would be kind of pointless if it weren't for the free units that the new owner receives after a revolt; maybe I hadn't thought of those units when I came up with the change. Revolt while in occupation can be pretty punishing (especially when at war), so reducing the probability may not be the best approach as it leaves a chance for freak accidents; better to decrease the foreign culture strength.

Converting the conquered cities has a dramatic effect on revolt chances, so it was critical to convert them.

The enemy state religion being present and yours absent (like in your game) is the worst case. I see that spreading your religion almost halves the foreign culture strength then. That's what BtS does too, but I had meant to reduce the impact of religion. It's not really working because I'm exponentiating foreign culture strength (to put it on a scale that is compatible with garrison strength) after applying the religion modifier. I'll probably change the order so that the religion modifier is applied last.

So in this case at least, the AdvCiv revolt system didn't really slow things down too much, if at all. It's important to pause after conquest and consolidate your new holdings, so I'd have to wait a little while before attacking a new target regardless. However, I'm not sure how much this rosy experience was tinted by having the Sistine Chapel.

When the game is in the bag, – and I think that's the case in the savegame where you're beginning your consolidation phase – then it should be OK to mostly ignore revolts. You've gotten to a 140% power ratio vs. your strongest rival without the Roman cities, so they're not essential for your war economy, and they're not going to flip either. Since you were able to produce culture cheaply, you were probably right to pause for a moment, but I don't see revolts greatly delaying victory either way. That said, on larger maps, where Domination is more difficult in any case, being able to snowball is going to be more important.

Leoreth · Aug 24, 2020

Oh yeah. There's a similar issue with stacking sound effect volume when multiple wars are declared simultaneously due to defensive pacts. Maybe not an issue in BtS, but in RFC defensive pacts are cascading (i.e. not immediately canceled when they are triggered) and that can get very loud when like 6+ civs are involved.

f1rpo · Aug 24, 2020

@Leoreth: I'm already suppressing the war-horns for DoW (and peace) involving vassals. But I guess DoW from (non-vassal) defensive pacts will also (always) play their sound on top of the original DoW. So I'll simply play the war-horns if and only if bPrimaryDoW is false in CvTeam::declareWar. The deal-canceled sound upon declaring war can also get loud; I've dealt with that in a recent commit.

keldath · Aug 28, 2020

HI,
noticed issues,
but im not 100% its advc related rather than doto, but, here goes:
resource popup for enemy cities -> i got a message alert that iron discovered for enemy city beyong my visible area.
city count is wrong on the score board - its with a basis of -1 -> unless you count the capital as 0 ?

also,
i have a very strange issue in the buttons for some techs in the pop up menu when prompted foe a new research and in the tech bar after tech is finished. i got a pink button there for pottery and a wrong icon for others like writing.
i dont recall ever changing this. doto doesnt change the related python.
any ideas ? its been in some of the latest doto and before.

Zacar · Aug 29, 2020

Background music fades away too quickly when zooming in, unlike in BtS / K-mod. The usual zoom distance I use has no background music as a result.

Edit: Eh, looks like you just recently covered this.

Edit2: -1 still has the issue, but the BtS value of 2200 works.

Edit3: It looks like you have removed the pending city growth alert. Is there some way I can restore it? It is tedious to go through all cities every turn to check if I should prevent the growth of a city for whatever reason.

Edit4: Looks like the within-borders forest/jungle tooltip leaks a bit information as it updates even within the fog, unlike borders themselves.

keldath · Sep 6, 2020

hey,
well i found 2 bugs i think.

minor bug:
1. when combat limit is reached for an attacking unit, AI . the message that is given is that a unit has withdrawn from battle.
a message of - max damage should be displayed instead i think.

major bug:
2. combatLimit for ai is wrong. when ai with a combat unit attacks a human unit.
the combat limit that is used does not match the combat percentage .
for example:
i gave a swrodsman the limit of 50% combat limit.
with WB i stacked a few of these vs a small warrior stack if mine.
declared war.
swordsman attacked the warrior stack until the combat limit was reached on all the warriors.
the problem is that the streangth of all the warriors stays on 1.2 while the warrior strength is 2.
the expected result is 1 out of 2.
if its vice versa, human vs ai, the limit is indeed 50% and sets the strength to 1 for the warriors.

i have reproduced this in the latest advc master branch.
i think this sort of imbalances battles in favor of human.
i tried its something maybe in in the bestdefender function and canattack, but i failed to see a root cause.
will keep investigating, but i hope you can fix,

edit: so far i saw that the limit is higher when the ai attacks a stack. still no idea why.

i found something strange, maybe kmod originated:
CvSelectionGroupAI::AI_getWeightedOdds
you added lots of nice comments there that helped me.
but, the int iodds - its send to bestdefender - but it does not return,
later on, you deduced attackchanceodds from it.
the definitions of this seems weird?

f1rpo · Sep 8, 2020

Back after a bit of a break.

@Zacar:

Zacar said:
Background music fades away too quickly when zooming in, unlike in BtS / K-mod. The usual zoom distance I use has no background music as a result.
Edit: Eh, looks like you just recently covered this.
Edit2: -1 still has the issue, but the BtS value of 2200 works.

Which field-of-view value do you use and how far do you zoom in (if I may ask)? The proper v0.98 release will have an option on the BUG menu, but I'd like to make the automatic setting work for as many players as possible.

Edit3: It looks like you have removed the pending city growth alert. Is there some way I can restore it? It is tedious to go through all cities every turn to check if I should prevent the growth of a city for whatever reason.

Uncommenting these 5 lines
github.com/f1rpo/AdvCiv/.../Assets/Python/Contrib/Civ4lerts.py#L423
should restore the original behavior. Currently, the alert only triggers when a city is about to shrink. Similarly, the pending happiness and health alerts are only triggered by negative changes. Those alerts can be restored by removing the "and willPass" from this line:
github.com/f1rpo/AdvCiv/.../Assets/Python/Contrib/Civ4lerts.py#L360
I guess pending (positive) growth is useful when a city is known to lose happiness/ health soon (or gain anger/ bad health).

Edit4: Looks like the within-borders forest/jungle tooltip leaks a bit information as it updates even within the fog, unlike borders themselves.

Thanks. That's easy to fix.

@keldath:

keldath said:
noticed issues, but im not 100% its advc related rather than doto, but, here goes:
resource popup for enemy cities -> i got a message alert that iron discovered for enemy city beyong my visible area.

Looks pretty waterproof:
github.com/f1rpo/AdvCiv/.../CvGameCoreDLL/CvTeam.cpp#L4531
Resources on unrevealed tiles are skipped. But I guess you're saying that the city mentioned in the message was unrevealed. But I don't see how that could happen either:
github.com/f1rpo/AdvCiv/.../CvGameCoreDLL/CvMap.cpp#L575
The call being

Code:

kMap.findCity(kLoopPlot.getX(), kLoopPlot.getY(), NO_PLAYER,
      NO_TEAM, false, false, NO_TEAM, NO_DIRECTION, NULL,
      /*eObserver=*/getID());

from the CvTeam that receives the tech.

city count is wrong on the score board - its with a basis of -1 -> unless you count the capital as 0 ?

Seems to show correctly the number of revealed cities on my end. :think:

I guess one could count capitals even if they're unrevealed, but I think that would be rather confusing.

any ideas ? its been in some of the latest doto and before.

Not off the top of my head. Have you fixed it already?

keldath said:
when combat limit is reached for an attacking unit, AI . the message that is given is that a unit has withdrawn from battle.
a message of - max damage should be displayed instead i think.

Sounds like BtS behavior; and, as far as XP is concerned, the attacking unit does withdraw when it reaches its damage limit. One could append "(Damage limit reached.)" to the withdrawal message, but it seems rather superfluous to me because e.g. a Catapult will only ever withdraw due to its damage limit (exception: Tactics promotion from a Great Warlord), and a Chariot will only ever withdraw through its withdrawal chance.

keldath said:
combatLimit for ai is wrong. [...]

Worked OK in a test (screenshot attached), but I hadn't taken into account your edit from yesterday. I'll look into it again tomorrow.

keldath · Sep 9, 2020

Back after a bit of a break.

12 days....i was beginning to worry

i hope your are well!

Looks pretty waterproof:

ok illl have a look.

I think that would be rather confusing.

yeah, well its none important anyway.

Not off the top of my head. Have you fixed it already?

no, this one eludes me, a real puzzle. i left it like this for a while now.

Sounds like BtS behavior

yes, easliy changable.

Worked OK in a test

i have been working none stop the past week on learning the code, the battle functions and much.
i still see some quirks, to reproduce, have 3 warriors some enemy unit stack 4-5 like swordsman with each on 50% limit.
let the ai attack the warriors till limit.
i only got the warriors to 1.2.
anyway,
ive been thinking, its funny - every turn , the unit is healed, and the ai comes back to atack the remaining new healed damage.
so the ai gets in a cycle of attacking turn after turn.
i thought about,, removing heal, while next to an enemy unit while at war - what do you think? where shoul the lines be?

another issue - see my PM to you, theres something in the weightedodds function that seems strange to me.
i added tons of code for my ranged combat, i really started to get the hang of it and im adding AI behavior all around and testing it on various situations)
for example , i wrote a func that tells tje ai to stop attacking a units if all the units on the stack are below the hp to a certain % , depending if theres an attacking unit (ranged units cannot move into enemy plots)
very fun.

glad your are back and pushing more updates

be well!

f1rpo · Sep 9, 2020

keldath said:
no, this one eludes me, a real puzzle. i left it like this for a while now.

Could be a problem with the tech-religion button. I'll write you a PM.

i still see some quirks, to reproduce, have 3 warriors some enemy unit stack 4-5 like swordsman with each on 50% limit.
let the ai attack the warriors till limit. i only got the warriors to 1.2.

I think you're not seeing the maximal damage because your units heal at the start of your turn. I had moved the Warriors in my test; in that case, they don't get to heal.

Code:

if (hasMoved()) {
   if (isAlwaysHeal()) // (March promotion)
       doHeal();
} else {
   if (isHurt())
       doHeal();
   if (!isCargo())
       changeFortifyTurns(1);
}

You could argue that a unit shouldn't heal right after getting attacked. If a unit gets attacked for two turns in a row, should it get to heal at some point in between those two attacks? If yes, then letting units heal at the start of their owner's turn (as it currently works) seems preferable. This way, if I have a 1.2/2 Warrior with a 5% fortification modifier, then I know that those stats will be used when an enemy unit attacks my Warrior; i.e. it won't heal before the attack, won't increase its fortification modifier, and so I won't have to do the math to figure out that healing will put the Warrior at 1.6/2. Well, either way, I can't really know my defensive odds because, if I hover for combat odds, my defensive modifiers aren't taken into account. That might be a neat little improvement: If Ctrl is held while hovering, show the odds of the selected unit (best defender if multiple selected) surviving an attack from the best attacker in the mouse-over tile.

:undecide:

Wouldn't really want to show a hint "CTRL for defensive odds" all the time when hovering for combat odds, and flipping all the colors around in the help text would be quite a bit of effort; see the attached screenshots of the current help text for illustration. There's a bug by the way (will fix): It doesn't make sense to attack with the 5-XP Axeman; it has the exact same odds as the 4-XP Axeman.

Also, moving healing to the start of the turn would give the defending side an additional advantage in some situations, e.g. when a damaged unit moves into a city on turn x, idles on turn x+1 and gets attacked in between turns x+1 and x+2.

As for not allowing units to heal in between attacks – game balance considerations aside –, the UI would then arguably have to mark units that have defended (could show the yellow dot on the flag – as if the unit no longer had full moves) so that the human player can tell which units will heal in between turns.

ive been thinking, its funny - every turn , the unit is healed, and the ai comes back to atack the remaining new healed damage. so the ai gets in a cycle of attacking turn after turn.
i thought about,, removing heal, while next to an enemy unit while at war - what do you think? where shoul the lines be?

I'm not a fan of the current healing rules, but this shortcoming of the AI doesn't seem like a convincing reason to change them. Your rule change would mean that units in a city under siege can't heal at all. Unless you exempt cities (and forts), but units don't tend to camp in the open anyway – except when besieging a city, and the heal rate in enemy territory is just 5 HP/turn (ENEMY_HEAL_RATE). Anyway, apart from CvUnit::doTurn (quoted above), CvUnit::healRate would be the place for changes to the healing rules:

Code:

void CvUnit::doHeal() {
   changeDamage(-healRate());
}

I guess the current AI behavior could be justified as keeping the enemy units (maximally) weakened until reinforcements with lethal damage arrive. This doesn't sound so bad when the enemy units defend a city; otherwise, reinforcements seem unlikely. In any case, the AI shouldn't attack when the defending unit is already very close to the damage limit and no lethal follow-up attack is available. The latter part is not quite simple to check.

12 days....i was beginning to worry
i hope your are well!

Sure, all good.

keldath · Sep 9, 2020

hey,

ech-religion button.

,
humm ok, interesting.

so, healing, well, i didnt think the implications of what would be if a heal would be on the end of the turn and in a city under siege as you said.
those are good arguments indeed..
i dont use combat odds much, or examine them deeply, im the rare plays ...i have.
maybe changing this heallng routine might cause some issues as you explained.
in the ranged attack im working on, i added liek a 10% lower combat limit on the average damage of the unit stack. meaning im telling the ai that if the units are harmed to a level of sort, thats enough , dont repeat the attack.
i have also thought of s simpler way maybe,
after the combat limit is reached, the damage done, is divided by 2. meaning it takes longer to kill unit for attackers with combat limit., i played with it a bit, seems cool.

There's a bug by the way

in the display of the calc?

one more -
the weightedodds , - takes into account adjacent plots?

f1rpo · Sep 9, 2020

keldath said:
i have also thought of s simpler way maybe,
after the combat limit is reached, the damage done, is divided by 2. meaning it takes longer to kill unit for attackers with combat limit., i played with it a bit, seems cool.

I see. A soft damage limit so to speak.

in the display of the calc?

In AI_getBestGroupAttacker. My code checks whether

Code:

kLoopUnit.AI_sacrificeValue(pPlot) >
kLoopUnit.AI_sacrificeValue(pPlot)

*giggle*
It only means that the "press ALT for best odds" hint is sometimes shown unnecessarily because the recommended attacker has equally good odds.

Elkad · Sep 9, 2020

I'm not in favor of removing/delaying healing at all.
A stack of March units with a proper (40hp) medic along heals 45hp/turn in enemy territory. If they spend 2 turns moving to the next city, they are all full.
A late-game city with a hospital and medic on defense is 70hp/turn.

Getting rid of that is a massive change.

Zacar · Sep 12, 2020

Ah, nice that I can just restore the pending alert via text edit.

My fov seems to be 36 and I zoom out one mousewheel tick when using the 2200 value.

f1rpo · Sep 12, 2020

Zacar said:
Ah, nice that I can just restore the pending alert via text edit.

If a second person needs the positive pending alerts back, I'll add a checkbox. (Or rather, I'll add adding the checkbox to my short list of BUG options to add.)

My fov seems to be 36 and I zoom out one mousewheel tick when using the 2200 value.

Pretty close. (Well, pretty much the BtS defaults, but displays have become larger.) I use 38 and zoom out almost as far as possible. Another extreme is:

Elkad said:
FoV 70, and zoomed out to straight down. (last click before the globe switch).

I'm attaching pictures of how all that looks on my display.

No way to tell in general what the default distance should be for a given FoV value. A BUG option (last attachment) is coming as soon as I can upload a v0.98 release candidate.

crullerdonut · Sep 13, 2020

f1rpo said:
If a second person needs the positive pending alerts back, I'll add a checkbox. (Or rather, I'll add adding the checkbox to my short list of BUG options to add.)

Yes, please. When it comes to the Civ4lerts section of the BUG menu, I've always just kept everything checked. It's nice to keep on top of all the cities, and the alerts help me with that.

Danwood · Sep 25, 2020

Made a new full game in marathon, and once again, this is the BEST civ experience i have had in my entire life (and i play since 2006). The only thing missing is the Mastery Victory mod, that would make it perfect!

f1rpo · Sep 28, 2020

Edit (15 Nov): This post is obsolete; I've added v0.98 to the resource database.
Edit (13 Oct): Release candidate #2 attached. (For the record, the first had 7 downloads.) Fixes several bugs, all in AdvCiv code, in particular:
• It seems that the Random Personalities option didn't have any effect in AdvCiv 0.97 (not sure about earlier versions). [advc.003q]
• Map scripts that aren't supported by the AdvCiv 0.98 starting position algorithm had crashed on some systems. [advc.027] Bug report; Big thanks to @keldath for working with me on this problem over several days. (I wasn't able to reproduce the crash at first.)
• The AI had sometimes razed major cities without due consideration (since v0.97, maybe earlier). [advc.ctr]
--

I haven't been posting much, in part, because of DSL troubles (seems to be properly repaired now at last). I'm attaching a release candidate for v0.98. A couple of balance changes could still be included. I still want to pick up the discussions again about culture victory, spy slots, meltdown chance and Divine Right, or at least explain why I haven't changed anything yet. Other than that, I'm trying to finally wrap this version up.

New since the pre-release versions (yet to be merged with this post for the full v0.98 release notes):

Spoiler :

Also, I've restructured some path finding code in the last few days. Just a heads-up; there should be no functional difference, but who can say for sure ...
Current version of the manual: (edit: link dead)

Danwood said:
Made a new full game in marathon, and once again, this is the BEST civ experience i have had in my entire life (and i play since 2006). The only thing missing is the Mastery Victory mod, that would make it perfect!

Thanks. Reassuring to know that Marathon works pretty well despite a lack of testing on my part. AI diplo memory decaying twice as fast on Marathon when it was supposed to decay half as fast (see the changelog above) is a pretty major jumble. I hope I'm not actually jumbling it now; easy to get confused about that formula.

I've gathered the MMod Git commits (by karadoc) for Mastery victory:

Spoiler :

Perhaps someone wants to do something with that; I probably won't. It would be easy enough to hide the victory condition behind some XML setting, but it looks like quite a bit of work to merge it in the first place (and maybe also to maintain it).

Elkad said:
I'm not in favor of removing/delaying healing at all.
A stack of March units with a proper (40hp) medic along heals 45hp/turn in enemy territory. If they spend 2 turns moving to the next city, they are all full.
A late-game city with a hospital and medic on defense is 70hp/turn.
Getting rid of that is a massive change.

That's for sure, also a big effort to implement such a change. So, no worries.

Speaking only hypothetically, further changes that disadvantage aggressors could be evened out by dialing down defensive advantages.

Danwood · Sep 28, 2020

Even without the Mastery Victory mod, I'll keep playing AdvCiv, as it's the most polished Civ experience I've found in many years.

About the relation decays, I've noticed that actually, but didn't think it was too strange to be an actual unintended issue/bug, glad to know it's fixed!

I'll try the next version and report back! Thank you for your awesome work!

PS: I really appreciate the care you put into the manual, from a fellow perfectionist :3

keldath · Sep 29, 2020

gogogo 098

i might be up for the challenge in getting Mastery. maybe after official 098 will go out so i can align my mod also.

btw,
PaltyPing wrote a pure Mastery victory in python which is module based also.
might want to check it Danwood.
but im not to keen on using gameplay changes based on python, ai and such.

cheers.

crullerdonut · Sep 29, 2020

f1rpo said:
When choosing random civs at game start, civs with multiple leaders have a slightly greater probability

Ohhh nice! Multiple people, including myself, have asked how the game chooses random leaders. I've never known if the game chooses by civilization first and then rolls again to determine the leader, OR if all leaders are weighted equally. If the first option, then leaders from multi-leader civilizations (Washington, Catherine, etc.) would be under-represented compared to single-leader civilizations (Pacal, etc.). If the second option, then multi-leader civilizations would be in play more commonly than single-leader civilizations. I've never been bothered by how the game does it, but I've always wanted to know how that works.

f1rpo said:
New option "Partial Score Graphs" on the "Advisors" tab of the BUG menu (disabled by default): If a civ's demographics were never visible but the civ has been known for at least 5 turns, then the score graph shows the score history since the turn on which the civ was met.

That sounds like a cool option, so I'll have to give it a shot playing with that option. Does the pre-contact score show up, once you gain enough :espionage:

to see all demographics?

f1rpo said:
Assigning multiple promotions at once; Stack Attack; and deals canceled upon declaring war
should no longer result in a loud sound.

You've single-handedly fixed a quality-of-life bug which has plagued this game since the beginning. That's something to be especially proud of, I'd say. Now, I don't have to worry about waking my neighbors when engaging in late-night Civ wars.

f1rpo said:
"Asoka" is spelled as "Ashoka".

By the way, during Sullla's Survivor series this year, Ashoka managed to make it all the way to the final. Multiple people including myself had been saying how to properly pronounce the name, and Sullla indeed made it a point to pronounce it with the sh sound. Of course, a few stray s-sound pronunciations slipped in, but after a decade of habit, that's just what happens. But hey, it's never too late to correct a historical inaccuracy, just like you did with the Pacal II thing. :thumbsup:

f1rpo said:
Civilized Jewelers no longer requires Mass Media, i.e. requires only Corporation as in K-Mod.

I'll have to keep that in mind, and maybe it'll be another one of those games when I have an unwanted Great Artist pop up around that time. Then I'll put that bum to work and make me some money :gold:

Thanks again for all your hard work, f1rpo.

Advanced Civ

plastics

Blue Period

plastics

LivE LonG AnD PrOsPeR

Warlord

LivE LonG AnD PrOsPeR

plastics

Attachments

LivE LonG AnD PrOsPeR

plastics

Attachments

LivE LonG AnD PrOsPeR

plastics

Emperor

Warlord

plastics

Attachments

Warlord

Warlord

plastics

Attachments

Warlord

LivE LonG AnD PrOsPeR

Warlord

Similar threads