Learning from the most popular Civ4-mods: Fall From Heaven to Master of Mana

Background (skip this if you’ve played Master of Mana)

As a child I played to death the early RPGs and MMOs – Ultima series, Diablo series, Everquest/Ultima Online/etc. Years later I was at NCsoft, working on MMOs, and I was lucky enough to meet some of the famous developers behind those games – the Garriott brothers, Dallas Snell, etc. One day I spotted a screenshot on Scott Jennings’ (Lum the Mad) blog of what looked like Civilization, but showed a civ-leader styled as Satan. I would have moved on, but Scott was saying something positive about this, and I had huge respect for his opinions on good and bad game design. So I looked into it, and discovered Fall From Heaven2 (FFH2) lead by Kael.


I started with FFH2, which in turn was adopted and extended in a more “radical” form as WildMana (WM). These mods were often made by hobbyists with little or no programming experience or training – just a lot of game-design talent and motivation. They were stuffed with great ideas, and riddled with bugs. Crashes, corruption, massive loopholes and imbalance. 30 hour games that corrupted on turn 200, no matter which save you loaded from, wiping all progress. But the daring and innovation was so great it eclipsed all that, and they were enormously popular.

Eventually, one of the modders (Sephi) picked up the ideas in WM, and started again: cutting away the chaff, massively rebalancing the gameplay, and trying to unify and simplify the million-and-one design ideas. This became Master of Mana – the final, greatest, most popular Civ4 mod. MoM was never finished; the hobbyist author moved on (a great loss; I wish Firaxis had hired him); but one of the releases was open-sourced and picked up by a group of volunteers who keep it alive (albeit unfinished).

Possibly the most daring thing Sephi did with MoM was this: for each major point-release (and some minor ones), he rewrote the core balance and game-rules, almost from the ground up. Hundreds of Civ4 owners would descend upon it, playtest the crap out of it, and discover the flaws, the loopholes, and the places with too much micromanagement or too little. He’d watch, playtest himself enormously, and then re-think everything, and 6 months later drop something wholly new and unexpected on the community.

You cannot talk about Master of Mana as a game: you have to talk about MoM 1.0, MoM 1.2, MoM 2-beta, MoM 1.42-Xtended-v5 … etc. Each one represents a different variation on Civilization.

(Latest up-to-date downloads for community-maintained Master of Mana are in this forum thread on CivFanatics)

Lessons learned from Master of Mana

UPDATE: here’s an overview of some of the core gameplay changes and features in MoM compared to its precursors.

(I’ll say “from MoM”, but really I mean from the whole chain of mods, from FFH onwards. Sephi gets namechecked for driving the final stages, but credit is owed to the many contributors who made each mod great – browse the CivFanatics Civ4-modding forums to find them)

I learnt four big lessons:

  1. Show the player EVERYTHING that affects their decisions
  2. How to talk about game-balance in Civ games without getting lost in the details
  3. Be more ambitious in how much you change and add to the game-rules
  4. Scalability of power / effectiveness of in-game units, buildings, techs, etc

Show the player, don’t lie cheat and steal

Pet peeve: if you feel that your game design is improved by hiding information from the player that is known to them (i.e. part of the rules, or that they’ve discovered in-game) … you are a weak game designer and you need to up your game.

In a complex, data-heavy game this is absolutely critical. The best patched versions of Civ4 were those that added displays telling you what the AI players thought of you, how big their military was compared to yours, etc. You need this information, and you can estimate it with a calculator if you’re bored of life.

But these are games: they are supposed to be fun. Making players get out a calculator because you want it to be “harder” for them is almost never fun. Show them the odds of combat, preview the possible outcomes and percent chances; this makes yoru game better.

MoM mods took this to the extreme, but in a very good UX way. Instead of overwhelming the player with data, only the core data was displayed, and whenever you wanted to know “why?”, you hovered the mouse over it and a popup would give you all the factors that went into the calculation. For instance:

Screen Shot 2016-05-21 at 16.11.22

Game balance in Civs: vocabulary

This is what you need to know – some of which I hadn’t picked up even after 10+ years of playing Civ games (my bad!):


Civ games have “ages” or “eras” of the technology/research tree. These are mostly cosmetic in gameplay turns – although they often have small effects, such as when Wonders are invalidated the moment you reach the Classical Era (by reseaching any of the Classical-era technologies).

But the units in the game can be ranked more concretely: into Tiers. The TL;DR is:

A civ with units in one Tier will always win any war against a civ with units in the lower Tiers, even with a military as much as 2 times smaller, because of the power-scaling of later units.

In practice, a typical categorization of units might be:

  • Tier 1: the “1 attack, 1 defence, 1 movement” units you start the game with, and anything they can reasonably win (or survive) battles against
  • Tier 2: the “3 attack, 0 defence”, “3 defence, 2 movement” units that require a number of researches and are simply “better at everything” in game-changing ways
  • Tier 3: units that will never lose to T1, and rarely to T2; these are your Riflemen fighting Archers
  • Tier 4: elite units with game-changing special powers; rare units that have a per-player limit (because they’d be too overpowered otherwise); units that single-handedly obliterate previous tiers
  • Tier 5: super-elite units with game-breaking powers: for when the game should have finished, but no clear winner has emerged, and we’re running out of game-content (no more tech-tree left to research, no more map to settle and build cities on).


Civilization-1 had a Tier-2 unit available from start of game in Tier-1 (the Chariot). Severely imbalancing – if you saw one at start, you had to quit game and restart. It needed heavy penalties against acquiring it too soon, forcing the hand of the game-designers into ugly decisions.

Tech-tree subtrees

The primary “sub-trees” from Civ4 that persisted through all the revisions of the mods were:

  • Recon – Fast-moving units which specialize in “exploration” at the cost of combat. Either much weaker than same-Tier units, or heavily penalized in non-exploration combat to prevent them being used to rush across the map and wipe everyone else out. “Exploration combat” means they generally get bonuses to survivability – e.g. retreat/surviving a loss – instead of being made more powerful to match same-Tier units.
  • Melee – Standard city-attacking units that are your “armies”. Should be good at winning battles against everything. If there’s a unit that is good for building armies and winning wars, and it’s not a Melee unit, something is wrong and needs rebalancing
  • Ranged – Like Melee, but heavier on defence and promotions don’t increase their attack. Or: useless on defence and heavy on attack. Push players into simple A/B/C strategic choice of “attack, defend, or compromise?”.
  • Arcane/Disciple – the two new combat-paradigms that FFH introduced to the game: rather than Rock/Paper/Scissors, it’s “choose one of three styles of gameplay/micro-management”
  • Hero – slightly underpowered units that grow almost exponentially more powerful if you can keep them alive long enough. Massively restricted in number (each player can typically only have 1-3 per game).
  • Economy – things that increase your ability to grow the size of your civilization

And some pseudo-types, pseudo-subtrees:

  • Beast – powerful AI-controlled units that can only be gained by capturing them in battle. Roam the world passively, reducing player-expansion in early stages of game, and obsoleted in later stages of game. Accessing them requires specific sub-choices in the late-stage Recon tree
  • Religionthe more game-changing aspects of Religion that are optional to core gameplay. You can have Disciple units and techs and not get involved in the full Religion side; or you can you go full-throttle Religion and convert your units to the religion, getting extra promotions and penalties for all your units, cities, etc.
  • Guilds – ditto to Religion.

Scaling of in-game techs/buildings/units/etc

Civ4 (and Civ5, and Civ3) shipped as unfinished, incomplete games; this apparently started as an accident (with Civ3) but shamefully for Firaxis became a sales strategy for Civ4/5/6. Ship the game with only the first 50% of content implemented, then ship the missing parts as “DLC” that you charge almost the full price of the game for. In Civ4/5 you cannot realistically finish the game, in Civ1’s terms: you get to the late Renaissance era, and then the techs and units and gameplay runs out. Civ4 didn’t even have Spy units or Diplomacy – the main part of Civ1’s mid/late game – at all!

It has long been a difficult question: what should a Civilization game turn into once players mature out of the early game? Once you are settled, with a decent economy, a substantial number of cities, access to a wide variety of resources, etc – what next?

The FFH/MoM mods taught me a lot about how to answer this. I will write more about this in later posts, but a brief overview:

  • The player makes long-term commitments in perhaps four different ways:
    • The next 4-5 Technologies to research
    • The balance of Military versus Economy things to spend city-production on
    • The output of cities (heavily dependent upon location of each city)
    • The size of their empire
  • In classic Civilization games, the tech-tree and the Military-vs-Economy are used too much, and the other levers too little. Civ 1 gets an excuse in that they only had 0.6 MB of RAM to play with; the sequels do not.
  • Long-term commitments are generally mutually exclusive
  • Long-term commitments can be substantially weaker than their peers, if they enable a leap-frog later in the game
    • Leap-frogging should generally be 4-5 Techs in the future, or similar number of game turns
    • Slower leap-frogging should come with game-changing new abilities, because civs that chose this path will have fallen far behind as a result
    • Faster leap-frogs need to be evoltionary rather than revolutionary, see below
  • Short-term commitments should change the balance of the player’s options, but not the impact
    • Fast leap-frogs fit here: they should enable you to cancel a long-term commitment, or let you start to rebalance your army, etc
  • Highly predictable long-term commitments represent a baseline
    • Every player will always have the option of achieving these
    • Alternative options can be measured against these by the game-designers (players will do so!)
    • If a long-term radical commitment falls through, the predictable commitments represent the only sensible “Plan B”
    • …without these baseline options, players who fall behind, or make gambles, or guess poorly … simply lose the game. This leads to long, boring, drawn-out death-throes … or games that are too rapidly and easily dominated by one player. This is boring.
  • Semi-random events that significantly or radically alter gameplay provide an excellent way to rebalance power if some players have got too far ahead
    • MoM mods added the ability for two players to summon new AI players to the game, bringing benefits to themselves, and destruction to everyone else
      • In classic MoM-style good game-design, the two summons were asymmetric
      • One of the summons changes the game for everyone, but is also destructive to the player; the other is less radical, but also less destructive to the player
    • Random events that come with meaningful choices are great fun.