I have no idea how formulas work in this game (would like to know though), but along the “simplified realism argument”, I think bonuses should be avoided, they seem to me a sign that the parameters used to reproduce reality are either not enough or are being inefficiently used. I rather have a couple more parameters than who knows how many random bonuses here and there. This keeping in mind that it would be nice to eventually show the relevant info not as a bunch of text like it’s now, but as a given set of icons and values that is as simple and standard as possible. It seems to me that with hack, pierce, crush, block (with a shield), parry (with a weapon) and dodge values, both for attack and defense, most things should be achieved (I don’t see any advantage to organic/metallic armor types, it’s just setting hack/pierce/crush defence values). Also, causing low damage should do very little, as a measure of actual penetration of defenses, to enhance differences, and penalising massing up the wrong kind of counter. Going through some of what has been said, and just as a first approximation: Spearmen: high pierce attack, very hard to dodge, low dodging. Archers: pierce attack, easy to block, basically impossible to parry, hard to dodge. Cavalry: low blocking, high dodging. Maces: high crush attack, hard to block. Axemen: high hack attack, hard to parry. Swordsmen: high hack attack, hard to dodge, high parrying, high blocking if they have a shield. Javelineers: high pierce attack, low crush defense, high dodging. Elephants: very high crush attack, high hack and crush defense, can’t be dodged, can’t dodge. Rams: extremely high crush attack, very high pierce defense, high crush defense, extremely easily dodged, can’t dodge. Buildings: extremely high pierce and hack defense, very high crush defense, can’t dodge. Slingers: high crush attack, low defenses. Slingers should be good against troops that don’t have crush defense, cheap, sling faster and farther than bows (bow damage should decrease more with distance), but have no armor, be unable to form close formations when slinging, should have a straight line of sight (slinging over friendly troops was too risky, iirc), and, most importantly, take a lot of time to train. Regarding their slow demise (they would be around up to the Middle Ages though), certain late agricultural technologies should drastically increase how much their training takes, since they came from a more agrarian background, and besides some early “biconical lead projectiles” tech to increase their pierce damage, they should not be improved much by other techs (around the middle to late timeframe of the game sling technology was historically at its peak, while others kept improving). Going back to unit’s values: archers’ arrows should be somewhat easy to block with shields, they would lose against cavalry, who should lose against spearmen, who should lose against swordsmen (here I wasn’t sure if they should win or lose against archers, I guess it should depend on how much their shield can block). I like the hard to block/parry/dodge dynamic of the mace/axe/sword, going along with the reality that crush damage transfers more through shield and armor, axes are hard to parry, and swords have faster attacks. I have not emulated all possibilities in my head though, all remains to be tested. Javeliners are good against elephants (from a distance), which along rams are good against buildings. It’s a matter of fine tuning to get what one wants, as Enrico Fermi told Freeman Dyson: “with four parameters I can fit an elephant, and with five I can make him wiggle his trunk". No bonuses should be needed!

I work on a similar mod, and I have started with the Mauryas. Some ideas I am testing. Temple can be build on neutral terrain, but does not project a sphere of influence Several new Temple technologies (not necessarily unique to the Maur): Aura around Temples that increase worker gathering and building rate; Temples grants structures +x% territory influence radius Historical reasoning: It is claimed that Ashoka spread Buddhism and interlocking of religion and economics / state Stables and Elephant Stables switch places Historical reasoning: According to sources I have read elephants accounted for about 20% of the empire's "cavalry"; enough to use some leeway to make them gameplay-wise more relevant, and make the Civ more unique Visha Kanya can be trained in the Temple Only trainable after researching a tech? Are one or both of them trainable in P2? If Archer is trainable in P2, P3 tech to increase poison damage Swordwoman has a damage bonus against heroes Still trainable in Palace? Give all Maur heroes an aura that buffs Visha Kanyas? Historical reasoning: According to the folk lore Visha Kanyas were young girls who were administered various poisons and antidotes to make their bodily fluids toxic, and then sent out to assassinate influential targets by being intimate with them    

Shameless self-adulation: All changes of the above mod will be in 0.29, btw. Better GUI: You have a better structure overview in the main menu, and it adds more hotkeys. You can also use it to "cheat" if you want to. Haven't tested it, but this seems to be good: This makes the differences between Civilians and Citizens more noticeable, as currently male Civilians and Citizens use the same voice:    

French =B

public Litterally, a must-have.