Before I get into the response, I'd like to note that I now have the effects for the Dream Twister (all other civs get -5 Happiness), Planetary Datalinks (gain any tech known by 4 other civs), and Telepathic Matrix (gain any tech known by any other civ) working, with notifications as necessary (although I haven't fixed it so that the Dream Twister is listed separately in the Happiness summary at the top). I'm hoping to have the Nethack Terminus working tonight, although not in the form I'd originally intended; now, I think it's going to be a National Wonder that gives you a "they're almost done!" notification popup whenever any AI has completed more than 75% of a Wonder, and I'll probably switch the tech locations of Nethack Terminus (Digital Sentience) with Network Backbone (Optical Computers) so that you'll get this effect in the Digital age. Since the wonder also gives you a free tech, I don't think you can complain about it being underpowered.
I'm hoping to have a new version by Friday; there's a good chance I could have the production-to-culture conversion and happiness boost from Empaths working in at least limited form (where they wouldn't update the numbers on the screen but would kick in at the end of the turn), but those might have to get bumped to the next version.
1. Considering that you have 4 complete eras that occur after the completion of this project (and a ton of SPs to boot!), then I do think it is overpowered for where it is in the tech tree.
Terminology issue: while SMAC used "SP" to mean Special Project, I generally use the "Wonder" label. "SP" to me means Social Policy, i.e. the Policy branches needed for a Cultural victory. I'm not sure what the number of Wonders or their locations has to do with the spaceship's construction. You don't win by building the spaceship any more, so how would it stop you from getting those later Wonders?
Have you thought about moving it up the tech tree a little ways, maybe to the next era, or even after that?
Yes, I've thought about that. Most of the technologies needed to build an actual spaceship capable of going to Alpha Centauri would be a bit beyond the Nuclear-era technologies. For instance, Fusion Power is one of the first techs in the Fusion Era, so it seemed strange to make it a current-technology project. Thematically, it really would fit better as an early Fusion Era project.
The problem with this is Centauri Ecology, and its descendants. These should be unavailable until after you've reached Alpha Centauri for obvious reasons, but I need them to be early enough for some of the other techs to make sense (like all the bioengineering stuff in the Fusion Era). So I really didn't have any choice, and in fact moved three of the four spaceship pieces to EARLIER techs than they'd been at previously.
2. The only two SPs which are in parallel (Sydney Opera House and the UN) to the spaceship components are on a research branch which doesn't have any military benefits. Question: in your games typically what does the AI do in regards to research? Do they beeline for military advantage?
Depends on the civ. The AI seems to do a good job of considering the Flavor ratings of each tech; the only really strange one I've found is the Naval flavor. Civs that favor that (like England) focus WAY too much on naval-associated wonders and techs. For instance, if I start a Transcendence Era game, and one of the AIs is England, I can be sure that the very first thing to be built in London would be the Maritime Control Center. I've actually tweaked some of the Flavor ratings recently to reduce this effect.
And I'd be careful classifying that branch as non-military; while it has few units in the Digital Era (just the Mind Worms and Golem), those techs have all of the production, happiness, culture, and food boosts, along with most of the Wonders. (In the Fusion Era, you've got the Ranger and Troll units up there as well.) Yes, this puts you at a military disadvantage when facing a warmonger, but that's actually part of why the Digital-era combat units have such low power ratings; using Nuclear Era units (Modern Armor, Mechanized Infantry, Stealth Bombers, Jet Fighters, and Paratroopers), a defensive player should be able to hold off attackers in that era while building the Wonders and such. It's not until you reach the Fusion era that you really need some firepower to survive.
In my games I have seen the AIs complete the Sydney Opera House and UN before me, however that doesn't necesarily mean they focused on this branch of research while neglecting the other branches. I don't know the algorhythms regarding how the AIs determine research priorities,
What happens is this, from what I understand:
The AI takes the Flavor ratings of its various options. It adds a random number (say, from 1-5 for the techs, whose Flavors all total 10) to each, and then weights it by that particular AI leader's priorities. It then does a random draw on the sum of the various options. The result of this is that while all options are possibilities, those that more closely match that leader's preferences will be far more likely.
So it doesn't matter what actual items are at each tech, just what Flavor ratings were assigned to that tech. If you set the Flavor to -999 it'll never build/research it, no matter what benefit it gives, whereas a Flavor over 100 means top priority.
There are two problems with this.
1> Most military-oriented techs will have two or three Flavor ratings that add to 10 (Rocketry is 3 Air, 3 Ranged, and 4 Spaceship), while nonmilitary techs will be a single 10-point Flavor (Globalization is Diplomacy 10, Ecology is Production 10). And the way the math seems to work heavily biases towards those single-Flavor items. (See my previous comment on the Maritime Control Center.) So if any AI has an above-average Diplomacy preference, he'll go for the UN at all costs.
So the default values encourage the AI to go towards the Globalization tech line first.
I can easily tweak the Flavor ratings of some of these techs, and I'll probably do so in the next version.
2> The total Flavor ratings of Wonders goes up a bit as the game progresses; Ancient wonders, like the Pyramids, total 35, and the mid-range ones like the Sistine Chapel total 50. Eiffel Tower, Pentagon, and Statue of Liberty all total 60 (30/30), so it's still reasonable in the Industrial and early Modern.
But then you hit the last three normal-game Wonders: Cristo Redentor, United Nations, and Sydney Opera House, all of which total 100. (CR and SOH are 70 culture/30 wonder, UN is 70 diplomacy/30 wonder). A Flavor of 100 basically means "build as soon as you possibly can"; the Factory, for instance, is Production 100, so the AI will build factories ASAP as long as it has Coal.
(For reference, all of my new Wonders have total flavors of 50, with no progressive increase. I probably SHOULD bump them up a little, to 60/70/80ish.)
So I'll lower the Flavors on those three wonders to 30/30, which should cause the AI to not go so overboard on them. Like so many other things I've found, it's yet another case of how the developers decided to code the Modern Era as a completely unbalanced "endgame", which conflicts with any mod trying to extend the tree.
bulking up their military then weathering the storm while they focus on nabbing the SOH and UN, then pocketing the Spaceship at their leisure.
Well, if the bonuses for building the spaceship are large enough, then you won't want to wait until "at their leisure". And there's another factor, the tech prereqs:
Centauri Ecology is only unlocked through the spaceship. The techs along the top row of the tree all depend on it directly, but there's also a diagonal progression; Retroviral Engineering (row 3) requires Centauri Empathy, and that tech gives you the Genejack Factory, the future equivalent of the Factory's production boost. Subatomic Alloys (row 4) in turn requires Retroviral Engineering, and that's the tech that unlocks Neutronium, a key military resource. Advanced Spaceflight (row 6) in turn requires Subatomic Alloys, Matter Compression (row 8) requires that, and Quantum Power (row 9) requires that.
So if you don't build a spaceship, you simply can't enter the Nanotech Era, and only five Fusion Era techs don't lead back to Centauri Ecology. (Granted, they're very military-oriented techs.) You can get around this a bit through the Planetary Datalinks (which now works!), but that'll only kick in when 4 other civs already have the tech you're missing. And it won't ever give you Centauri Ecology, which among other things unlocks the Omnicytes strategic resource. This can be crippling to a military.
Tried playing a couple games this evening using your mods, however tech times are now over 60 turns starting in the Nuclear era,
I'm assuming you mean that that was the time for a game that starts in the Nuclear Era. I agree that it's a problem, and hopefully my next round of changes will help that a bit. For one thing, I'm trying to code into the .lua a patch that makes your first few techs cheaper than normal if you start in an advanced era. So for an Industrial/Nuclear start, maybe you get 4 times the normal beaker production on the first tech you research, 3 times on the second, 2.5 on the third, and so on until after 10 or so techs you're paying the normal amounts. With your numbers, that'd mean a 15-turn first tech. (To compensate a bit, I still intend to reduce the effects of research buildings significantly, so it'd probably be more like 20 turns when I'm done.) It might not display correctly, so it might still say that it'd take 60-80 turns, but you'd actually get the tech in 20ish and I can probably add a popup notification explaining this whenever you get a new tech. Something like "For this next technology you will get 400% of the usual progress."
The problem is just that the game is balanced around you having a complete empire by that point in the game, and no one really starts in the Modern Era in the core game because you're only a few turns away from winning by Diplomacy or Space. So there's no need to balance it for the stable-state long term, because you'll never reach that. The default was to just cut all of the tech costs by a large percentage, to compensate for your much smaller empire (which'd never grow to full size before the game ends). But this just isn't sustainable in later eras, because it doesn't scale; if the default is for all techs to cost 20% of their normal values for an Industrial start, you'd still only be paying 20% once you reached the Fusion Era (by which point you WOULD have a full-sized empire), and you'd be getting everything in one turn of research.
Question: do you believe that Firaxis just gave every faction Agriculture to avoid the issue of Factions possibly not researching this tech?
Basically, yes. In Civ4 you had civs with different starting techs, and while the smart move (unless you were beelining for religions) was to pick up the "essentials" (Agriculture, Mining, Fishing, etc.), it was just too easy for an AI to bypass at least one of those and go a little ways down a tech line. SMAC had been even worse for this; its tech tree was so convoluted that it was really easy for a civ to not have many of the low-tier techs. Counting Alien Crossfire, there were 8 techs in the lowest tier of the SMAC tree, and you could go a LONG way down the tree without picking up any of them.
Civ5 is different. The tech tree is a lot more cross-connected, to where you just can't progress quite so far along a single line (with a few notable exceptions), and there are no dead-end techs at all (which is a lot more SMAC-like than Civ4-like). This helps the AI, in that it won't made stupid decisions nearly so often since there'll be less need to look three or four techs down the road.
This actually is related to one of the basic design principles behind my mod. If you look at the tech tree you see a LOT of diagonal connections, as I mentioned in the earlier discussion on Centauri Ecology. This tends to encourage researching a single tier at a time and not moving on until you've completed the column, instead of beelining down a single row to get a good long-term tech. That's something an AI can handle pretty well, which reduces the Human Advantage.
And that's also why I tried to have every technology give two or three things, instead of the core game's tendency to have techs that only offer a single building or unit. That way, the AI won't be penalized so badly for researching a technology whose benefit isn't really what it needs; if each tech is giving you a Wonder, a unit, and a bonus to some type of Improvement, then the AI is never really harmed by taking that tech over others.
2. From your playtesting to date what do you see as the trends in regards to what the AI does well, and what the AI does poorly?
The biggest problem with the AI is just a question of adaptation. Take tile improvement; a human player would know that the priorities are:
1> Put the appropriate Improvement on every luxury or strategic you don't already have.
2> Connect every city with a road to the capital as soon as possible, and later on a railroad.
3> Improving bonus resources and redundant luxuries/strategics.
4> Everything else.
The AI DOES have some basic logic in it to determine how many workers he needs, and then sets priorities to "Need Workers", "Want Workers", or "Enough Workers". (One bug in the core game is that those first two are disabled for any games starting after about the Renaissance Era, so an AI in a Modern Era start in the core game would almost never produce any Workers. This was especially noticeable for city-states.) But beyond that, the AI is stupid; he'd rather move his Worker one tile and place a Farm than spend two turns moving to that new Strategic and hooking it up, even though the strategic would benefit him far more on the long run.
This is endemic to the AI. Each unit seems to react as an individual unit, based on what's around it, instead of being directed by some higher authority. For instance, I'd invade the Mongol Empire, and while I'm sitting on one side of the empire annihilating his cities, I can see his cavalry on the far side of the empire fighting an allied City-State that was no real threat. A smart player would realize that the City-State was no threat and would move those units to where they could save the empire, but the AI doesn't do that.