I think I'll have to write a bit about my playstyle and what I like/dislike about Civ5 to qualify my choices (which still might see more tweaks, but you've definately understood my major changes correctly).
I haven't played a Civ game since the original 1991 Civilization game, so to me alot of things were new, great people for instance which I understand was part of some of the previous games as well, and the new things for Civ5 like hex and one unit per tile. Unlike many I actually enjoy 1upt alot, and were never too fond of stacks of doom even in the first game (although I believe that perhaps 3upt with diminishing benefits might be better, but that's discussion for another game). Hex tiles too I find is brilliant, and I love the maps in Civ5. I actually can't remember if RA's were in the first game or not, but I don't have anything against them by themselves, just a part of the "problem" as I see it with my playstyle.
I usually play pretty slow paced, enjoying developing my cities, teching, social policies (also one of the things I love in civ5), and making an army that is enough to protect myself, but rarely much more. Then whenever I encounter a Civ that either is starting war with me, inhabiting land I want, or just plain annoys me, I build an army that I feel will succeed and take a few chunks of his empire, slowly advancing throughout the game. Sometimes I spread a bit on the map (I really want to try getting some outposts near far off city states in some game), but I rarely build a large empire.
In the original Civ my most common problem (as I remember it, been two decades since I played it) was that I'd have a pretty small start I enjoyed then when finding other civs I realize I'm completely out teched and there is no hope. Or alternately the reverse, that I find the AI is still in stoneage while I have nukes (figuratively speaking... mostly)
I've played Civ5 since not too long after it was out, through all the various major fixes and reworkings that the game has had (seriously, the manual is pretty useless now heh), so became pretty apparant that the balance was pretty arbitrary although it seems alot better now after all those patches.
First few games I played at standard speed, and although I found the game somewhat fun, I noticed that I didn't even have proper time to explore the map, let alone get into any major wars, before I got social win, tech win or diplomatic win, and the victories generally felt hollow. I did play a bit lower difficulties then too though, I'm mostly playing King now (although if the AI will have trouble with my mod I might take it up another notch).
So I changed to marathon speed, and at first it felt alot more right to me, I had enough time to have units check most/all of the map, and got into some major engagements that were interesting.
But I started to notice a few things that were exactly same in standard and marathon speed. Since making units and buildings were slowed same as techrate in marathon, building an army compared to technology advancement was equal. And as the game progressed it felt like the technology was accelerating at an exponential rate. I literally couldn't get into any offensive wars since whenever I built an army it would get obsolete immediately, or more commonly I'd be thinking "oh I want to invade that civ, but first I want to get musketeers... oh wait, just 3 more turns for metallurgy then I can bulb rifling, better wait for that..."
Some places of the techtree are more obvious problems for me on that than others. I can barely remember building a caravel ever for instance since frigate is just a tech away, and other units I usually skip are spearmen, lancers and cavalry, the first airplane, and more. Again this is because of my own playstyle, not advocating that my way is more right than anyone elses, just how I prefer to play. I am certain that less OCD type players will be building army constantly and just sending wave after wave, I just don't like that way of playing. And I feel a bit of a problem with suspension of disbelief when my army is already obsolete when it reaches my enemy. It's like if WW1 started, and before it was over there were stealth fighters being built in America and the troops are wondering where the heck are our reinforcements.
Come to think of it, has anyone ever fought a war in Civ5 that felt like WW1 at all? feels like it's just a so very very tiny period on techtree where there is simple firearm units and cavalry side by side... but I digress.
This technology speed which really was same on standard and marathon prompted me to look for mods, and start playing around with my own to try adjust things. One of the first things I tried was just adjusting the tech costs (about same amounts as noted above, although a couple of games I had to abort since the Int techcost wrapped around), and what I found surprised me, that even changing tech radicly and exponential like this, it still felt like tech went really fast and although it removed the "one tech per 1-3 turns" in later times, it still had me doing them in like 10 turns which still felt like very fast to me leading to same problem I mentioned above. I've rather recently realized that the most major reason for this is probably the additive bonuses to techs from buildings, so like you mentioned, adjusting these is probably 90% of what is needed.
My secondary objective was keeping the game in pre-modern times. I don't so much enjoy modern/future civ gaming, so as personal preference I wanted to adjust the game so that I'd have enough time to Dominate the game before reaching modern (even with my slow war/expansion rate). And I didn't want to just disable the later techs, since if there is a "max tech" to reach, I'd likely be playing slowly again wanting to reach that before seriously starting to war, and whenever I reached max tech I knew that: A. no AI can have better stuff than me, and B. There is no more interesting things to tech. Both of which could lead me to boredom and abandoning the game.
So I decided on a more exponential model which also should help lessen tech gaps between civilizations, since even if you get a lead, your next tech costs exponentially as much as your low tech opponent, so it doesn't mean you're certain to keep the gap for long. The tech spread features also help protecting this, and I've considered toying a bit with those, or add tech diffusion mod to the mix, but I need more testing to decide.
So far in testing various iterations of my changes, I feel that I'm on to something that makes the game more enjoyable to me, and gives me back the "just one more turn" feeling that I didn't have so much with the unmodded game. It's not perfect, I still don't think I've ever built a spearman, and I kind of doubt I'll use caravel either, but it does help. I've had battles in various ages and seen how cavalry slowly gives away their superiority as pikemen arrive, and how city sieges evolve over time. And when I was looking around for mods that did something like this I did see many others that felt they had the same problem I did, so atleast some share my playstyle I think.
That said, I will need to tweak alot more... after nerfing the science buildings powerfully I saw a much more radical change than the tech cost increase, so I'll need to see how to balance it... it will take me a while.
Yes, it does. And if you're drastically shifting the balance to where that should be the primary thing they do, then you'll need to change quite a few other things, like the Flavor ratings on improvements.
Good point, I should look into that.
Fact is, the Great Scientist's free tech was never that great, mathematically. Until you reached the Renaissance (or later, in my mod), you'd be better off making an Academy instead. Players loved to use it for a quick tech slingshot, but that usually wasn't the best use in the long term; gaining a bunch of extra beakers every turn for hundreds of turns would quickly add up to more than the cost of a single tech. (Remember, Academy yields get multiplied by the University and such...)
I've read some of these things as well, but I've never been quite convinced. Maybe it's since I like building wonders, and bulbing a tech to make sure I can start building it before the AI feels very comforting to me, especially in the early game. At same time, my mod will probably be a bit flavored by the fact that I do generally build alot of wonders and it might be alot more unplayable at high difficulties.
I should probably lower the yield of the academy, but will test a bit and consider.
The AI, however, can't handle that concept. To it, it's always more or less equally valuable to start a Golden Age, place an Academy, or bubble a tech. If you remove one of the options from that list, it'll just randomly pick between the other two, and that's a problem.
Yeah, I need to do something to prevent AI's from using a major beneficial tile improvement for a golden age instead.
Well, the problem is that you really overdid it on the nerfing of RAs. The main reason RAs were so strong wasn't that they were inherently a good use of your money, it's that they were cheaper than the other things the AI would consider using his money on. The AI has no concept of "saving"; if he has enough money to buy something, the only question becomes Yes or No on whether to purchase that particular thing on that particular turn. Since RAs were substantially cheaper than buying units or buildings, or bribing city-states, the AI would invariably buy an RA because somewhere in the five to ten turns between when he had enough for an RA and when he had enough for anything else, the AI would pull the trigger on the purchase.
I have a few major problems with RA. First of all it will strengthen a continent with more civilizations on it compared to one with fewer ones (add into this the tech bleed as well and it becomes more profound, as people playing earth map in america's have noted as well, europe getting major tech lead). Secondly, it was starting to anger me that I would get into RA's with a civ, then 10 turns before completion it would start a war with me, wasting the benefit (although I've learned not to sign RA with close civ's now). Also, amazingly, I have actually signed RA's for 7500 gold in my modded games, when AI asks me for it, so although doubling it might make them less common, even 10 times cost doesn't make AI completely ignore them.
Also, with techs taking 40-60 turns to research on average throughout the game, RA's give a pretty huge benefit, and atleast with my playstyle I'd happily pay a few thousand gold for saving that many teching turns (and giving more chance to build a new wonder first).
Not saying I'm adamant on 10x cost though, it was mostly a bandaid to see where most the blood was coming from in the teching wound, and it turned out that science buildings are likely the worst, not the RA's.
The problem is that you've created a binary solution set. If the player doesn't take Piety, he can never grow large because there simply won't be enough Happiness for any other strategy to work, but if he DOES take Piety then there's no Rationalism. I've never absolutely had to use Avoid Growth in any games I play, because you quickly learn how to pace your expansion to the policies and buildings at hand. The first instinct for players coming from Civ4 is to expand as fast as possible and grab all the resources you can, which then forces you to use Avoid Growth to keep from crippling your empire with unhappiness. But the AI knows better; this is WHY it doesn't expand as quickly, its behavior has been set for a pace that won't cripple itself.
And this leads to the major issue I mentioned: right now, the AI sees Piety and Rationalism as being more or less equally valuable. But if you're changing the game such that you NEED Piety just to play, then the AI won't know how to deal with it.
You might be right, I'll need to do testing to try figure out how much of a problem this will be. The AI is bad no matter what, so not sure that giving binary happiness/research choice will affect that so greatly. Also, Honor can give pretty huge happiness benefits as well with the defensive structure and garrison bonuses.
You can take this sort of effect a bit further if you want a slower play style, but like I said, there are practical limits on how far you can push the AI before it can no longer adapt without heavy adjustment in other areas (like city growth). You might find that you have to adjust unit maintenance costs, gold production buildings, various Flavor entries, and so on as a consequence of your changes.
I won't be surprised if I will need to make further alterations to make the game still work right, but my main goal is really keeping game "as is", but stretching the tech periods out alot to give time to enjoy the game before it's over (I think 2000-3000 turn game would be about right for me).
Biggest AI problem is really that it is incapable of saying to itself "Ok, I have enough units, I should put cities to gold/research instead now since it's peacetime", but not much I can do about the AI really.
That'll depend heavily on the map type, though. On a pangaea, it's practically guaranteed that someone else will have a border right up next to yours, which invariably creates friction and leads to early wars. But try a Small Continents map, where it's very likely that a few civs will be on isolated landmasses. In that sort of game, a mod like this would cause the AI to quickly drown itself in units, with a combat unit on every single land hex and no one nearby to declare war on. I've seen this happen in vanilla, let alone a mod that deliberately encourages that sort of thing.
Also, the AI will handle a large army much differently than a human player will; a human can get the same amount of effect out of a much smaller army, which frees up a lot of gold for other uses and a lot of production time for Wonders and such. Hence the issues I mentioned.
I will admit I've been testing on pangea and continents only, not small continents which I can see might be problematic (and archipelago is obviously a problem no matter what for the silly AI). Although hopefully once the AI has sea technology it will spread out right and even things out a bit right?
