Crazy Spatz's Alpha Centauri Mod

[Madsapper]

Combat Engineer: I'm going to try to find a good bulldozer model to improve it, but what it has now is at least acceptable.

While you're at it, could you change the name to Civil Engineer? Combat Engineers are more about mines and demolitions than building happy things like roads and farms..

 

[Spatzimaus]

Combat Engineers are more about mines and demolitions than building happy things like roads and farms..

I know. My Dad was one; his unit in Vietnam basically specialized in blasting roads through hills, making airstrips on short notice, and stuff along those lines right alongside the infantry. Also, he used a lot of Agent Orange, but I'm not planning to add chemical weapons into this mod.

The thing is, the CEs in this mod were made for exactly those roles; they're much better than normal Workers to accompany your troops through the front lines and lay roads/railroads as you go. That's exactly what I just used mine for in the game I played tonight; Workers wouldn't have been nearly mobile enough to keep up, but the use of CEs made a huge difference in my ability to keep moving. Of course, normal Workers are fine for the safer behind-the-lines stuff, and I'm trying to see if I can get the AI to differentiate the two a bit better (instead of upgrading every single Worker ASAP and bankrupting itself). The other thing I use them for heavily is laying (blasting) roads through hilly areas, since they won't have to waste a turn moving there before starting; in Civ4 I LOVED India's "Fast Worker" UU because of exactly that sort of thing.

So the question becomes, then: should they have a Combat rating? The Labor Mech and Golem, in the future eras, are worker units that have some attack ability, so I'd be okay with giving the CE a combat rating of, say, 20-24; just enough to defend themselves a bit, but dead meat when attacked by tanks.

The downside to doing this would be that these could no longer stack with combat units. If I do that, then I'd want to add a resource requirement to the unit (say, 1 Oil) so that late-era starts wouldn't use them. I might also just have to remove the upgrade chain from the Worker, but that has its own issues.

 

[Spatzimaus]

Okay, I've found a bad problem. A few versions ago the game would start doing something strange when you attacked a spore tower, where it'd move your attacking unit to some hex adjacent to the tower instead of moving into the tower's hex. Well, it's worse now; it's moving my AIR units into the hex of the psi units (ANY psi units) when I use a bomber that has the Logistics promotion. Since the unit is no longer in a city, it dies at the end of the turn.

I'm betting that this is related to the psi units' healing mechanism, since it involves moving the unit to the hex it's currently in (which redraws); somehow, it's doing that for the attacking unit instead, meaning I've flipped a reference somewhere.

I'm going to try tracking this down tomorrow, but it means that I'm almost definitely going to have to make a new version before I go on vacation on Wednesday. So if you find any other bugs, now's the time to tell me.

 

[Spatzimaus]

Found it. In 1.06 I made a change where Psi units would reset their strength to their base value at the end of a fight. Unfortunately, I mistyped; if the defender is a Psi unit, it resets the ATTACKER'S strength to the defender's value. Which, if the defender is dead, kills the attacker. This sort of thing is why I normally insist on playing a complete game after any changes are made before I upload them, but this time I rushed it because of my vacation schedule.

I'll be releasing a new version tomorrow, as soon as I test out one other bug fix. In the meantime I'm going to remove the download links from the front page so that no one grabs the buggy version.

 

[dezuman]

I just realized why every AI is drowning in money in my games: I play without City States. Tried out putting some in, and I had this Askia with like 150 GPT in medieval era that apparently used it all up on the < 10 CSs in that game...
And he didn't get a single effing allied state

Oh and BTW I wubbed the fast worker too

Edit: Does the AI NEED their normal over 9000 bonus against barbs? Feels a bit off now with cuddly neobarb worms all over ... Is it something about them not attacking "too strong" units AT ALL, even if they are pillaging the out of them?

If you start a game, in FireTuner it'll say something like:

 Map Script: Adding Oil resources to the Sea.	3
 Map Script: Adding Supplemental Oil resources to the Sea.	3
 Map Script: Adding Dilithium resources to the Sea.	8
 Map Script: Adding Supplemental Dilithium resources to the Sea.	8
 Map Script: Adding Omnicytes resources to the Sea.	9
 Map Script: Adding Supplemental Omnicytes resources to the Sea.	4

This is from the game I just started. Now, I'll translate from Spatzese to English:

Oil:
> Take the number of land-based Oil deposits, in UNITS, and divide by 8. Round this down, and then place N 4-unit sea deposits. That's the first 3.
> Take the number of civilizations, not including city-states and barbarians, divide by 3, and round UP, again placing N 4-unit sea deposits. That's the second 3 (I'm on a Standard map, 8 players).

Dilithium:
> Take 2/3rds the number of Whale deposits, 2/3rds the number of Pearls deposits, and 1/3 the number of Fish deposits. Add these and round down. Place that number of 5-unit Dilithium deposits (that's the first 8).
> Take the number of civilizations, not including city-states and barbs. Place that many 3-unit Dilithium deposits. (That's the second 8.)

Omnicytes:
> Take the number of land-based Omnicytes, IN UNITS, divide by 2, and round down. That's how many 2-unit deposits of Omnicytes to place in the sea. (That's the 9.)
> Take the number of civs, divide by 2, and place that many 1-unit sea Omnicytes. (That's the 4.)

The point is that for each of the three sea resources, there are two separate spawn methods; one depends on how many deposits of certain other resources were placed by the randomizer (and can therefore vary wildly based on your choice of maps), and the other depends solely on the number of civs and so won't vary with anything other than map size.

So in my current game, there'll be 16 Dilithium deposits for 8 players (half at 5 units, half at 3), although many of these will be in city-states' territories. If the CSs hooked up the resource correctly, this'd be plenty, since by the time you reach the late Digital at least four of the eight civs should be dead. Assuming four surviving civs, that'd average ~16 units per active player, which should be plenty to support your army but not enough to put Fusion Labs in anything beyond your core cities. By the time you hit units that require far more Dilithium (i.e., the Titans) you should be making Quantum Labs.

Comparing to other resources is hard. You should see a large number of Omnicyte deposits on the map, but they'll all be 1-3 units, because Omnicytes are a hybrid strategic/bonus resource, where the tile yields matter more than the strategic quantity. And Neutronium should be fairly scattered, since it counts as a luxury.

I can tweak the numbers a bit, like making the deposits 6s and 4s instead of 5 and 3. But that's the basic idea.

Oh, just me being ******** again. Was generating in the world editor (So no players, durr)
It did however seem like it NEVER generated in 4 corner maps...
And yeah, I had your mod loaded

 

[Darsnan]

Two games between yesterday and today, with the following observations/ notes:

1. How coming bombing (either with arty or aircraft) Workers/ Great People nets XP, but attacking them with regular units does not?

2. Concerning game generation: does the game place civs based on the order in the selection menu? Just a curiousity question on my part regarding how the game goes about placing civs.

3. I've specifically been playing againts Greece to see if I could replicate the SC/ CR observation. In my first game the Greeks spawned near me and chose the war victory condition very early (I knew this because they went to war with me very early in the game), so probably not a good litmus test there. The one good observation was that the Greeks did select the Collective Rule Social Policy, meaning they had four cities to combat me with!

4. With the new patch I'm sensing a new combat algorhythm whereby the AI will withdrawy its units on one turn, then re-commit them to the attack the next, sort of a "one step back, two steps forward" mechanic. This has the effect of massing the units with newly arriving reserves, so it gives the effect of coming at a player in waves. Is anyone else sensing this (or sometehing similar)?

5. Concerning "History repeating itself": I found it amusing the parallel between this and what the Aztecs must have been thinking when they offered up the tribute to the Spanish! Also, wouldn't it have been better to invest that gold in buying some combat units instead?

6. Concerning "Nethack Terminus Bug": FYI.

7. Concerning "Really Stealthy Bomber": is this related to the bug you documented above?

8. Concerning "Unreachable GG": the aircraft icon completely covered the GG icon, and the only way I could access it was via moving the AC to another city.

Anyways, have a good vacation!

D

 

[Spatzimaus]

I just realized why every AI is drowning in money in my games: I play without City States.

I played a game last night that had two continents. On the other continent were Greece and Siam, both of whom were bribing the heck out of anything that moved, and both had Patronage. It was HARD keeping them from winning diplomatically, especially once Siam conquered Greece and started picking up his city-states unopposed. I ended up having to go all Mongol-like and conquer city-states just to reduce the number of total votes available, even though it alienated pretty much every CS on the map.

The other thing is that the AIs that alienate every other major civ will also have a large gold surplus, since they won't be buying RAs. This also happens to the player when he gets too far ahead; last night I did a Renaissance start, got the Porcelain Tower and went Rationalism, so I would have had awesome RAs, but no AIs would deal with me after a little while because I'd pulled too far ahead.

Oh and BTW I wubbed the fast worker too

It's a good example of how a tiny change (+1 movement) can have huge repercussions in balance. That one extra movement point allowed him to move onto a forest or hill AND start construction on that turn, and it just made a huge difference in the long term.
I've tried to recreate that effect, first with the CE, and second the Planetary Transit System (+1 movement to all of your land units that started the turn in your territory). This is why quite a few of my custom effects are things that you might not otherwise want (-1 visibility to enemy units in your territory? A 5-turn warning when someone is building a Wonder?) but that end up being handy in enough situations.

Edit: Does the AI NEED their normal over 9000 bonus against barbs? Feels a bit off now with cuddly neobarb worms all over

It's a tough call. I personally think the various anti-barb bonuses (the base value, the Honor tree, etc.) are just a bit too strong given that I'm heavily leveraging the Barbarian faction for my mindworms, so I've been considering reducing them. That's why I added the Neural Amplifier, Citizen's Defense Force, etc. in the first place. The problem is, if I drop it too much then city-states won't be able to survive the Breakout.

Is it something about them not attacking "too strong" units AT ALL, even if they are pillaging the out of them?

Yes. If there's too big of a disparity, the units will back off and find something else to do. You also see this when you fortify an infantry unit in a good terrain; the AI will run up to it and then reverse course without fighting. This seems to be a recent change.

Unfortunately, this works against the Psi units a bit, because they don't get their +/-25% until the fight actually starts, so the worms don't realize that they'll be a bit stronger when it matters. This is one reason why I capped it at 25%; if I'd tried to be more SMAC-like and not cap it, then it'd really screw up the AI's decision process.

Oh, just me being ******** again. Was generating in the world editor (So no players, durr)

Never thought of that. If you're using world editor, then EVERY resource would be screwed up, because nearly every strategic has that sort of population-based supplementary distribution now. I think Horses are the only strategic left that doesn't; the other 7 land resources do, and all three sea resources. (Oil and Omnicytes are on both lists.)

It did however seem like it NEVER generated in 4 corner maps...

Four Corners shouldn't have had any problems; the only AssignStartingPlots routine it alters is the one that places city-states, so it should be compatible with my changes. I'll check that out later to be sure.

I'm still trying to track down why city-states won't place a work boat on the sea-based Dilithium deposits. No luck so far, and if I can't fix that, I'm going to have to change the Dilithium distribution method to compensate, because you're SUPPOSED to be getting a fairly large fraction of your dilithium (like half) from city-states. (I'm trying to encourage making CS allies, even if you're not going for a diplo win.)

 

[Spatzimaus]

1. How coming bombing (either with arty or aircraft) Workers/ Great People nets XP, but attacking them with regular units does not?

Because it's stupid.
Really, I'm not sure why; noncombat units have some defense against ranged attacks and aren't killed instantly, so maybe it's because of that, or maybe it's just because without that bonus the AI would be punished even more for not understanding the proper way to deal with these units.

2. Concerning game generation: does the game place civs based on the order in the selection menu?

Not exactly. What happens is that the game splits the map it's generated into Regions in AssignStartingPlots:ChooseLocations(). These regions are ordered, lowest to highest, in terms of fertility, so that the tundra and desert-type starts are first on the list. It'll balance out these regions a bit to make sure they're all equally desirable, though.

Then, when it comes time to assign starting locations in AssignStartingPlots:BalanceAndAssign(), it'll give first priority to those civs that have Biases (Iriquois start in woods, Arabia near desert, etc.), and assigns them to the first available region that matches that bias (if any). Once it's done with the civs that have biases, it goes to the rest.

If you've disabled start biases, then the civs are all randomly shuffled before they're assigned. If you've left biases turned on (the default), then it'll only shuffle within each type; all coastal civs are put together, shuffled, and assigned; then, River biases, then Region biases, then everyone else.

4. With the new patch I'm sensing a new combat algorhythm whereby the AI will withdrawy its units on one turn, then re-commit them to the attack the next, sort of a "one step back, two steps forward" mechanic.

Yes, I've noticed this. The AI is much less willing to make an attack if it'd leave his units vulnerable to counterattack, and less willing to face bad odds in general. I'm still testing to see if this is really a good thing, but so far it's made the AI's warfare more interesting; as you noted, when he attacks, it's en masse, and so far more likely to do damage.

5. Also, wouldn't it have been better to invest that gold in buying some combat units instead?

There are a lot of bits like this; the AI is hard-coded to keep a certain reserve in various situations, to keep from getting crippled by the things it doesn't look forward towards.

6. Concerning "Nethack Terminus Bug": FYI.

Nice catch; I just misspelled "title" there (ttile) in two places. Fixed in the next version.

7. Concerning "Really Stealthy Bomber": is this related to the bug you documented above?

If you'd just attacked a Psi unit, then yes.

What was happening was this: aircraft have a ranged combat rating but not a regular non-ranged one. You'd attack a Mind Worm with your bomber, and at the end of the fight, instead of the worm's strength being reset back to 50, the BOMBER was being given a strength rating of 50. Since it now was seen by the engine to have a "melee" attack, the next time you tried to use it it'd move from the city to the destination hex, attack with this new melee attack, and then park itself there for the rest of the turn. At the end of the turn it'd die for being outside of a city.

I'll put a new version out in a couple hours that fixes this and the other bugs I introduced. There was also a bug in UnitPanel, in that it was trying to get the Help text for units, but the game doesn't HAVE help text for the Great Scientist, Merchant, Artist, Engineer, and General, so I had to add some.

8. Concerning "Unreachable GG": the aircraft icon completely covered the GG icon, and the only way I could access it was via moving the AC to another city.

I know. You can also select the units through the Unit List menu in the upper-left, and sometimes if you click on a unit flag it'll move that one to the back of the stack (like it's supposed to do). I've run into this issue in a few different ways, including naval units in cities being unselectable behind workers and such.

Really, whatever routine handles unit flags needs to be overhauled. Instead of upper-left for combat units and upper-right for noncombats, it needs to move the noncombats to something in between. I might fix that eventually, but chances are I won't have time to look at it before my vacation.

-------------
Also, this reminds me of a quirk I wanted to mention: you know how the Vertol, Nessus, Former, and Gravship (the four all-terrain units) are technically naval units? It had an unintended site effect: they can stack with land units. Yes, yes, the Stacks of Doom are BACK! Well, a 2-unit stack, generally a Modern Armor or Skimmer screening a Vertol from enemy fire, a Nessus Worm screening a Chiron Locust raider, or a Gravship screening your land units. And honestly, I'm okay with all three of those situations...

 

[Strategist83]

Very impressed with your mod, Spatz. Thank you.

edit: Also, you got me started playing SMAC/SMAX which would probably have never happened hadn't I been introduced through your mod.

 

[Spatzimaus]

Okay, status update:

I've fixed the bugs we've talked about above, plus a couple other minor ones I've found along the way.

What I'm trying to do now is tweak the Victory Progress screen to show both the space race (even if you somehow turned it off) and the Transcendence race, as a prelude to decoupling the spaceship construction from the space race. It's not working; as the game starts to load and it reads the DOM text, the screen goes black and it won't initialize the game. Obviously, I've got a problem in the Lua/XML for the Victory Progress screen, but I'm having a really hard time pinning it down.

I'm going to try working on it for another hour or so. If I can get it working, then that'll be in 1.07, otherwise I'll just disable its VFS and push it to 1.08, and release the 1.07 fix this afternoon. (Since the files WILL be included in the package, if any of you want to try figuring out what's wrong...)

 

[Spatzimaus]

Emergency patch time!

Here's the fixes for the issues we've talked about, v.1.07, which ONLY affected the Content mod. The links on the first post have been re-enabled, but the Balance mod remains 1.06 as nothing changed in it. If you're one of the ~30 people who downloaded the 1.06 Content mod, make sure you get the 1.07 ASAP.

v.1.07
CONTENT:
* Fixed a significant bug involving air units attacking Psi units.
* I THINK I've fixed the bug where attacking a spore tower would cause your unit to end up in a hex adjacent to the one it was attacking. Tell me if you see that happen any more.
* Added Help text for the five existing Great Person types, to go with the new Unit Panel changes.
* Reduced the number of units that spawn at the Breakout by 25%, which should take it a bit closer to what it was before.
* The Nethack Terminus' espionage reports will now have the correct title.
* The All-Terrain promotion wasn't letting the units cross rivers without ending their turn. I've fixed this.
* Because of the above fix, there's no longer any need for Vertols, Nessus Worms, and Formers to have the Amphibious promotion, since everything it does is already handled by All-Terrain. My concern is that AI units will select Amphibious, then, even though they don't benefit further.
* Similarly, there's no need for Vertols to have the Hovering Unit promotion or the "Ignore Terrain Costs" promotion, since that's all covered by All-Terrain. Of course, a vertol that had been upgraded from a gunship will still have some of these.
* The Former could use its Terraform action on hills regardless of terrain type, which was causing problems when it was a grasslands or plains hills. So now, the Terraform action only works on flat land; if you want to terraform a tundra hill, use the Raise/Lower Hills action to make it flat first.
* The Plant Forest and Plant Jungle can no longer be performed on Forest or Jungle hexes until you reach the transcendence era. (i.e., never.) These actions will still appear in the UI, but will be ghosted out; I might clean that up later. The stupid part is, I'd tried this exact method before (setting the auto-chop's tech prereq to something unattainable) and it didn't work, so something's changed.
* Increased the Science flavor of the Apollo Program. I'm hoping that this'll help the AI try to build it ASAP, even if it's not going for the space race victory.
* In the vanilla game, the four spaceship part projects had no flavor ratings. I'd previously given them a FLAVOR_WONDER of 50, but I've now also given them a FLAVOR_SCIENCE of 50. The units themselves also had improved flavors in the last patch; if this doesn't work to make the AIs build these more, I'm going to have to decouple the space race altogether, which'll probably mean losing the "launch" graphic.
* Fixed a typo that was preventing the Perimeter Defense from getting the production modifier of the Power policy.
* Re-enabled the selection sounds on the new units.

Now, I didn't get the Victory Progress update working, so that's been disabled for now. I'll see what I can do with it for the next version.

 

[dezuman]

Quick question: What happens to the new 1-hex range attacks under the effect of the Hunter-Seeker algorithm? Having 0 range ought to suck pretty hard

 

[Spatzimaus]

Quick question: What happens to the new 1-hex range attacks under the effect of the Hunter-Seeker algorithm? Having 0 range ought to suck pretty hard

It probably would. But remember that the Command Nexus, which unlocks at the same tier as the HSA, adds 1 to your units' ranges (but not cities' ranges, which unfortunately I can't change). The two cancel out. Or more specifically, they cancel out when you're in enemy territory; when you're in your own territory you still get the CN bonus but not the HSA penalty, and your Vertols will have a range of 2. Similarly, the Command Nexus' other benefit (+10% when in enemy territory) generally cancels out the Citizens' Defense Force's first bonus (+10% when in friendly territory).

Also, there are a few other possibilities:
1> As a Siege unit the Vertol has access to the Range promotion, not that I expect that to get taken often. The Nessus doesn't have this option.
2> While these units now have a ranged attack, I did NOT give them the "ONLY_DEFENSIVE" promotion, the one that prohibits archers from attacking with their melee attack. So, even if its range is dropped to zero, Vertols can still make a 50-strength melee attack against units in the field; they just can't use it against a city. For the Nessus Worm, his melee attack is actually STRONGER than the ranged attack (same for the Bolo), except that he'll take damage when he uses it (but with his full regeneration AND titan healing that's usually not a problem), so again, you only lose out on the ability to attack cities.

As long as the AIs are keeping pace with the player, technologically, it should all work out where one national wonder counters the other. If the AIs aren't keeping up (and at the moment, they're not doing so well), then I need to fix that first, but even so these units can still make melee attacks.

 

[dezuman]

I know this isn't really on topic, but how complicated would it be to implement civ4-style tile improvement "growth"? It was only used by cottages normally, but even I managed to apply it to a few other improvements xD
It just feels more worth it to pillage if they can't just rebuild it at full yield in a few turns ... And that ones "core" areas will be more worthwhile to pillage than the fringe!

 

[Spatzimaus]

I know this isn't really on topic, but how complicated would it be to implement civ4-style tile improvement "growth"?

The XML stubs are already there for "upgrading" improvements, but I don't think they're hooked to anything.

With Lua, things like this are easily possible, although there's no simple way to store what turn each improvement was created on. You'd have to either add some extra save/load logic (still doable) or make it random. For instance, way back when (like back in March-ish), I was still having problems with my Plant Forest and Plant Jungle actions, due to the limitations of SerialEventImprovementCreated. One concept I had was to add the Tree Farm (name taken from SMAC, of course), a new Improvement that had a 25% chance each turn of "evolving" into a full-blown Forest or Jungle. No need to save turn numbers, just a simple end-of-turn event (like the one I now use to plant forests and jungles).

The biggest problem, as always, was the AI. I needed him NOT to plant a Tree Farm and then immediately try to replace it with a Trading Post or something during the turns where it was "growing". This is why the current ones instantly grow into complete forests or jungles as soon as they're finished; I was even tempted to auto-place lumbermills and trading posts on them, just to make sure the AI didn't immediately chop them. And that's also the problem with adding Civ4-style cottages; the AI needs to understand that the simple little +2 gold improvement it sees now is going to eventually be a +5, and that it should destroy it to place a non-growing +3 one right now. But I'd be surprised if the current AI is capable of that, despite its obvious Civ4 roots.

If you want pillaging to be more crippling, then there are other ways to do that. Take Fallout as an example (the feature, not the game); if you nuke a hex, it doesn't just pillage the improvement, it also places a Fallout feature that adds -3 to each yield. So to truly fix the tile and return it to its original yields, the worker doesn't just need to repair the improvement, he also needs to clear the fallout, and that takes time. All you need to do, then, is add a couple new fallout-type features with varying penalty levels. Add a Lua trigger, to where pillaging an improvement also places one of these negative features on the hex (although you can't do this if the hex had a different feature, like forests or jungles). Done.

 

[dezuman]

For some reason I haven't used a nuke in Civ5 for ages ... Might be that I never ever get any strategics
Might as well try it out ... Now that I know to put in a few CSs as AI money sinks so they can't just buy an army for their over 9000 gold reserve (Well, more like 100k...)

And it's not like the Deity AI ever liked you in the first place, I've only had 2 neighbors and a few distant peeps ever NOT declare war on me at some point ... Although they do a shitton of peace trolling
Edit: The first one was Darius with 3 cities and constant wars going on, but the other was ASKIA with quite an empire ... might be that the personality randomization made him soft

 

[Spatzimaus]

For some reason I haven't used a nuke in Civ5 for ages ... Might be that I never ever get any strategics

Just use a few Graviton Missiles. They drop fallout on the target hex.

My original plan for the Gravy weapons (and yes, plural, there were originally going to be others, based on the SMAC terraforming missiles) was to have them be sort of stalemate-breaking weapons:
> Any units in the target hex or in adjacent hexes would take damage, including those inside cities. This part works, but currently it'll only trigger if you target a unit, not a city. I'm working on that.
> If it impacted on a unit in the field, then the target square would be devastated (terrain type set to Tundra, improvements destroyed). I changed that later on to just add fallout to the hex, although I should also have it pillage the improvement.
> If it impacts on a city, then the city would have a chance of having its Walls or Castle destroyed. (Arsenals and Military Bases also, possibly. The future-era ones should never be destroyed.) The idea was that these could be used to weaken a fortress city in preparation for your other units. But right now, I can't trap a city impact, although someone pointed out a possible way to fix that.

And it's not like the Deity AI ever liked you in the first place,

That's one of the things I dislike by going that high. King or Immortal is bad enough with the everyone-on-you dogpile, but on Deity it's just ridiculous. So my goal has been, through these changes, to make the game balanced enough that you can get a good game experience on King-ish difficulties. Unfortunately, I'm still seeing runaway players, where it's not too long before the player is sweeping every Wonder, and then it's only a matter of time. I've traced this back to a few factors:

1> The space race. The player almost always wins it, because the AIs simply aren't trying; this gives him a free tech, a free policy, and a long golden age. It looks like I'm going to have to decouple the spaceship from the victory types to force the AI to not pick this, although there's one more thing I can try with the economic AI tables.

2> The probabilistic approach the AI uses for construction is really hurting him. I need to make certain "core" buildings have much higher Flavor ratings, similar to how the Factory is a 100. This is especially pronounced on the growth-related structures.

3> There needs to be money sinks in the games at at least two points: the late Nuclear/early Digital, and mid-Fusion. I've been trying to accomplish this through increased building costs, but ideally, these would be new Route types, similar to how the advent of Railroads in the Industrial suddenly drops your income from +60ish to almost zero. So if I were to add Superhighways (or Rolling Roads if you want to follow the lore) in the first case and Magtubes in the second, and have them cost, say, 3 and 5gpt per hex... that'd pretty much kill your profit margins. Obviously they'd have to have some benefit, besides just greater movement; I'd want the highways to increase the trade route income and the magtubes to add even more production (probably back to the original +50% that railroads used to give).

Actually, what'd be even better is to bump railroads to 3gpt, add something between roads and railroads at 2gpt, and go 4 and 5 for the new ones. This'd make the Trade Unions policy (-33% route maintenance costs) far more valuable...
Right now there are route movement increases at Machinery and Monopole Magnets, but those can simply be replaced with new route types and smaller increases to the old types. I'm just not sure how well the engine can handle this sort of thing; if I've got a rail link and start replacing it with magtubes, can I make sure it doesn't drop the rail production boost while it's still half-and-half?

4> Tech costs need to scale up a bit more steeply, especially in the early Digital Era where you're suddenly growing more quickly through the combination of Recycling Centers, Medical Labs, Children's Creches, and certain +food Policies. Right now, you gain techs faster than you can build the buildings at each tech, and that sort of timing heavily favors the human.
I'm also re-assessing the Rationalism tree; the research bonuses it gives start to make a huge difference, and while Piety is far better for Happiness and Culture, there are certain stages where Happiness just isn't an issue. I may have to drop a few of its percentages, just like I dropped the research yields of the University and such. So instead of "+15% when happy", it'd be +10%. That sort of thing.

5> At certain stages, military attacks become too easy. In the early game, when you've got swords and horsemen, seizing a city is HARD. Even when it's riflemen and cannons, it's hard, because the city strengths scale up quickly. At the start of the game, the Palace gives every unit the Home Field Advantage promotion: +10% when in friendly territory, and another +10% when ATTACKING in friendly territory. This has made a big difference... until the mid-Industrial, when suddenly it's Artillery, Infantry, Tanks, and then Bombers. City strengths mostly stop increasing, but unit power goes through the roof.

Likewise, in the Digital Era, the Citizens' Defense Force gives +10% when in friendly territory, and +1 healing in friendly territory, and that's helped quite a bit. But I need more of these, specifically at two points where the balance shifts heavily in favor of offense: the early Industrial, and the mid-Fusion. So imagine a new National Wonder in the late Renaissance that, when built, gives all of your units +10% when in friendly territory, and, say, an additional +10% vs. ranged attacks. Or maybe just a flat +10% on all defense. (I've already repurposed the Neural Amplifier for the later era, so I only really need the one.) That sort of thing is not enough to make the game completely defensive, but enough to slow the blitzkrieg attacks a bit.
I'd also like, somehow, to boost city attack ranges by 1 through that national wonder. But that's impossible for now.

 

[Darsnan]

5> At certain stages, military attacks become too easy.... when suddenly it's Artillery, Infantry, Tanks, and then Bombers. City strengths mostly stop increasing, but unit power goes through the roof.

I agree whole-heartedly with this. From the very beginning of an Industrial era game I ( a human player) plan my city locations around defensive-offensive manueverings so that I can quickly bring to bear superior forces with either intent. The AI has no concept of this when laying out their intial city placement: advantage humans!
Also, when the core of your military can travel 4-5 tiles (and Panzers travel 6!) this ability to "reach out and touch someone" is really amplified in land warfare, which a human can take advantage of moreso than an AI. Then of course there are aircraft, and nuclear weapons. One of the weaknesses in this later-era environment is the 1UPT, as the AI just can't retreat wounded/ vulnerable units far enough from serious threat. And of course there is the matter of weaker but necessary support units such as anti-aircraft and artillery: they can't effectively stand up to the plethora of military options a human can employ against them (i.e. a human can assess the tactical situation and adapt his attack accordingly, whereas the AI is programmed to play a certain way regardless of the situation at hand).
Finally, there are the misteps the AIs seem to make where they commit units which really shouldn't be at the forefront of an attack (Artillery immediately comes to mind), possibly because the AI can't retreat them behind the wall of units it has moving forward and so has to move them forward, or because the algorhythms judging the situation are weighted to push these units forward.

Also, this reminds me of a quirk I wanted to mention: you know how the Vertol, Nessus, Former, and Gravship (the four all-terrain units) are technically naval units? It had an unintended site effect: they can stack with land units. Yes, yes, the Stacks of Doom are BACK! Well, a 2-unit stack, generally a Modern Armor or Skimmer screening a Vertol from enemy fire, a Nessus Worm screening a Chiron Locust raider, or a Gravship screening your land units. And honestly, I'm okay with all three of those situations...

Question: have you considered making artillery (and possibly air defense units) stackable with other military units such as Infantry/ APCs? This would make it harder for humans to selectively cull an AIs attack force/ defensive network, as the human would have to first eliminate the defending Infantry unit before engaging the Artillery unit being protected.

D

 

[Spatzimaus]

The AI has no concept of this when laying out their intial city placement: advantage humans!

This is one reason I'm trying to figure out how to raise the city attack ranges to 3; if the AI's cities can interlock their fire a bit better, it makes it much harder for a human player to sneak an army in through the gaps between cities.

Also, when the core of your military can travel 4-5 tiles (and Panzers travel 6!) this ability to "reach out and touch someone" is really amplified in land warfare, which a human can take advantage of moreso than an AI.

Actually, it's generally the reverse as I see it. High mobility often favors the AI, because it means his units won't just stall out movement on a badly-placed hill or forest. A Human knows to move his units in the right order to get them all into the correct positions, the AI handles each unit separately. In the early eras, the combination of low mobility and 1UPT leads to the strange formations you often see where the artillery is advancing in front of the infantry.

You especially see this in my future eras; one of the reasons nearly every unit gets the "all terrain 1 MP" ability is that this allows the AI's army to stay a threat regardless of the terrain. If you invade, he can reinforce from his other areas very quickly, and that generally means he'll kill some of your units thanks to the Home Field Advantage ability. The point of my earlier post is that I just want to make that bonus a bit more pronounced, by adding something in the Renaissance that boosts it further.

One of the weaknesses in this later-era environment is the 1UPT, as the AI just can't retreat wounded/ vulnerable units far enough from serious threat.

Part of the reason I added the orbital weapons is that I wanted to make it where in the long term, NO ONE could retreat wounded or vulnerable units far enough. Cities would be the only safe place, except for Titans and the Mobile Shield (which I'm still working on; it's going to have an anti-Orbital. nuke-immune promotion, and what I'm trying to do in Lua is have it give that promotion to adjacent units temporarily, instead of its current Great General + Medic combo). Basically, I'm trying to limit the player in the same way the AI is already limited.

And of course there is the matter of weaker but necessary support units such as anti-aircraft and artillery: they can't effectively stand up to the plethora of military options a human can employ against them

This is the entire reason for the Plasma Artillery, combining two overly specialized support units into a single unit, and by doing so remove a lot of the AI limitations. I think it's worked very well; the AI has no problem with building them (or upgrading from the previous generations), and it makes things MUCH harder for the human when he knows that his bombers will get ripped apart by multiple interceptors. It's especially helpful for city-states, since they won't have a lot of units at any one time.

This sort of merging-to-help-the-AI is something I'm obviously fond of, given what I did to the Stealth Ship (destroyer + submarine) and Leviathan (carrier + battleship), but it also explains a few of my other changes, like why Skimmers and Assault Powersuits have interception abilities. By spreading these abilities around more, the AI is less crippled by a bad decision in the units it has built; you'll no longer be able to think, hey, if I use my tanks to kill those two SAMs, my bombers will have free run of his airspace! This sort of thing helps the AI tremendously.

the algorhythms judging the situation are weighted to push these units forward.

It's not that, from what I can tell. I mentioned this above, but I think it's just the movement limitations. The AI loops over all units and says "okay, now what do I do with this guy?", moves the unit, and then goes to the next one. If it picked the artillery unit to move before the infantry unit, then you can easily get a situation where the artillery ended its movement on a hill and the infantry stalls out right behind it with movement points remaining and nowhere to go.

This gets a lot better once things start being more mobile, especially in the future eras, so it seems to imply it's a movement limitation and not an AI prioritization issue. Easy way to tell would be to change the 1UPT limit to allow two units instead. (It's in GlobalDefines.) I'm not going to do it in my mod, but it's a simple mod that anyone can do. If you still see the artillery moving ahead of the infantry with that change, then there clearly is a prioritization going on, but if they advance together, then it was just what I mentioned.

Question: have you considered making artillery (and possibly air defense units) stackable with other military units such as Infantry/ APCs?

To be clear, the only reason those units I mentioned can stack is that they're considered sea units, and the game allows each hex to have one combat land unit, one noncombat land unit, one sea unit, and an unlimited number of air units. It's just that my all-terrain fix shifted those four unit types from the first category to the third.

Normally, you can't change the 1UPT limits by unit type; it's all-or-nothing. So the only way I could make artillery stack would be to make it either have no combat rating (which'd make it automatically die when attacked, like a Worker) or make it be a naval unit that can cross oceans easily. Neither of those is really a good idea; while the first option doesn't sound too bad when they're escorted, how well do you think the AI would manage that?

 

[Darsnan]

This is one reason I'm trying to figure out how to raise the city attack ranges to 3; if the AI's cities can interlock their fire a bit better, it makes it much harder for a human player to sneak an army in through the gaps between cities. .

That sounds like a really great idea!

This is the entire reason for the Plasma Artillery, combining two overly specialized support units into a single unit, and by doing so remove a lot of the AI limitations. I think it's worked very well; the AI has no problem with building them (or upgrading from the previous generations), and it makes things MUCH harder for the human when he knows that his bombers will get ripped apart by multiple interceptors. It's especially helpful for city-states, since they won't have a lot of units at any one time.

I think the reason I brought up the case of the Artillery is that in my Industrial era starts (Standard Size/Continents/ Emporer) the game is pretty much over by the time Plasma Artillery becomes prevalent enough to have any influence on the game.
Hmmm, maybe its time to up the difficulty? Or maybe play on larger maps (which should produce some duels with bigger empires)?

D