DESCRIPTION
What does this mod do?
It adresses a bunch of severe balance issues that are currently in the game, mostly those pertaining to men-at-arms, in terms of overall power and balance between different unit types both, death spirals, realm stability, building balance, AI building and warfare logic and development being insignificant tax/levy-wise.

NOTE : I'd recommend using the steam workshop, since I update the mod more frequently there, and merely drop its current version here from time to time.

Short Changelog:

(Full Version in a discussion thread on the steam workshop)

Men-at-Arms

All building bonuses to men-at-arms are now % based, and flat modifiers now only come from inventions. Duchy buildings that boost their effectiveness now simultaneously increase their upkeep, whereas the barony buildings that do the same, reduce the upkeep. Men-at-Arms AI logic changes ensure that they will hire the most effective retinues available to them, instead of a random mismatch of many types.

As a result, men-at-arms power no longer scales exponentially (you could get 1500%+ damage and toughness in vanilla quite easily), and consequently lategame meme-at-arms are no longer super soldiers that can each chop through hundreds of levies - and are now quite expensive to upkeep, rather than literally free. Unlike in vanilla, it is no longer optimal to hold only duchy capitals and no "mundane" baronies in your domain, despite the penalties.

Each culture now has a starting "tradition" invention that starts already unlocked and slightly improves a certain type of men-at-arms, which corresponds to that of the cultural unit, and helps distinguish them for what they were known for in warfare, even before they unlock their culturally unique unit. French cavalry is now the french cavalry, as are the scottish schilltron and the english longbowmen (extends to all cultures in the game, these are just examples), and fighting different realms might now require you to adapt your strategy.

Men at arms base stats have been rebalanced, leading to a bit more impactful early presence overall (they are much weaker mid and lategame even if optimized). Furthermore, different units of the same type now remain distinct all game long in stats and terrain advantages & disadvantages, since these differences are no longer overshadowed by flat building bonuses, so you shall fear the 4000 byzantine cataphracts!

Economy

Most importantly of all, AI will now actually build stuff. Furthermore, all buildings were overhauled and rebalanced from the ground up, and AI building logic was greatly improved.

It will now seek to construct buildings that are relevant to the builder's culture and its strengths (MAA unit type), rather than random buildings, do so without compromising the economy, and all the while also avoid constructing buildings with little synergy in the same holding, seeking to build a set of buildings as complementary in their effects as possible.

The impact of each point of development on tax has been quadrupled (+2% per point, from the pitful vanilla +0.5%), making development-oriented buildings actually worthwhile long-term, and tribal nations less overpowered early on compared to feudal ones. This results in more money, but this is accounted for.

Development no longer has any impact on levies (was +50% at 100), but buildings, particularly upgraded ones, provide more levies and garrisons. This is more realistic, since more developed realms didn't exactly have more levies - they had more money to build up the infrastructure to have more levies - and remember, the AI now builds stuff!

All vassal contracts were rebalanced. Top lieges get more gold in general, letting them develop many buildings in their own domain (which they now prioritize over constructing in a temple in random vassal's land), as well as hire more mercenaries and maintain a full retinue.

Knights

Knight limit bonuses from buildings were severely reduced (between 25% and 50% of what they used to be). They are now less killy (100->70 damage per Prowess), but more durable (10->30 Toughness per Prowess), and have a greatly reduced chance of dying in battle (around 40% of vanilla), though they are more likely to be captured, so you need money to ransom the good ones.

Only characters with at least average prowess (9 or more) are now eligible to become knights, and landed characters (except barons and vassal counts) now periodically attract knights as guests, the frequency of which scales up depending on their title rank, and their level of fame. This fixes the vanilla death spiral of warfare where your entire court and all your male children and vassals, even the scholar with 1 prowess, constantly die due to becoming knights because there are no other candidates and you forgot to forbid them from knighthood!

AI and Warfare

Warfare AI overhauled. It behaves smarter in big wars, pursues war objectives, and is much less likely to engage in a shamefur dispray in general.

Embarking and disembarking both take longer and sea travel is much slower (2x land speed, was 5x), meaning it is easier to catch an invasion force with the -30 advantage penalty, for the player and AI both. The army upkeep while at sea is 200% (was 125%), making naval travel more costly to begin with.

Levies now gather with speed 15 (down from 40), doing so at a pace equal to the vanilla naval speed. This means less compact nations will spend more time gathering troops, and will be vulnerable in the meantime.

Advantage is now +2.5% per point (up from 2%), and some sources of it are rebalanced to make more sense. This makes defensive terrain, buildings and commander skill and perks more impactful, and helps the AI defend itself against numerically superior enemies.

A sample of other changes

You can capture knights and commanders while fighting with/against raiding armies, and they can capture yours! Raiding can now be very profitable, and dealing with raids can be as well. Capturing prisoners is more common if the battle is a stackwipe.

Diplomatic range on the default setting reduced to 750 (what "restricted" is in vanilla"), and restricted range to 500. This helps prevent blobbing. Note : Iceland is cut off from the world on the latter option, but Iceland is irrelevant.

Added "Hard" and "Very Hard" difficulties, where all AI characters (excluding player vassals) get reduced build cost - and nothing else - which has no immediate impact, but massively improves the size and quality of lategame AI armies.

Renown gain per ordinary, untitled dynasty or house member reduced by 75%, reducing the renown income of countries with concubines/polygamy, which could in vanilla complete all the renown trees within 150 years of game start due to their sheer population.

Holy Orders are fixed and relevant! You can lease them extra holdings if you are their patron, strengthening them, as will the AI. They are also much costlier in piety to hire, which results in them not being in permanent use, and they can no longer ask for / be given cities - which they got no levies/taxes from, since they use a feudal-ish government form.

Fix for North Korea mode - holdings above your domain limit no longer provide you with any tax or levies, however the artificial percentage penalties to the holding limit are no longer there, either. You now need vassals.

[Mod Compatibility]
(Always place Sane Warfare last in the load order!)

Incompatible : Full Overhauls

Maybe (or Balance Breaking) : Mods that add new MAA types, make huge general balance changes or, to a lesser degree, add new buildings (modifying vanilla ones should be fine if you overwrite another mod with this one).

Should work just fine : Everything else (for ex. the Better AI and More Game Rules mods)

CHANGELOG
v4 - Released 2020-09-29
Keeping up with steam version.
v3 - Released 2020-09-22
Keeping up to date with steam.
v2 - Released 2020-09-20
AI was, in fact, building too much after the recentmost changes, and had a bit too much gold to spare still... - Scaled down the income from buildings to around 80-85%. - Mercenary hiring cost 40% -> 50% of the hiring price of the men-at-arms inside per 2 years. - Levies cost a bit more to keep raised compared to before (1.5 gold -> 2 gold per month per 1000). - Garrisons were too big on average for the AI to handle, and there is no way to adjust that particular behaviour, so they are now a bit smaller. However, all the men-at-arms and mercenaries the AI now makes with overwhelming vigour make sieges a breeze, so fort levels from buildings were significantly increased across the board.

REQUIREMENTS

Crusader Kings III

Full game

FORUM LINKS