Master of Orion III had great ideas.
There. I said it.
Master of Orion III was justifiably trashed by critics and reviled by the strategy gaming community, but it did some wonderful things – it did most of these things very poorly, or cut them down to just a fraction of their original promise – but it did them.
Not controversial enough? OK, try this one: MoO III’s near-destruction of the entire franchise may have been a lesser tragedy than the sloppy execution of its great concepts. While the former meant over a decade in cold sleep before Wargaming revived the franchise, the ramifications of the latter have been deeper and more destructive in the long run. Quicksilver’s catastrophic failure was taken as a sign that straying from the path worn by its predecessor, Master of Orion II, would end in ruin. The result has been a derivative and incrementally evolving genre that we’re still (mostly) living with today. A few games have dared to strike out on their own with varying degrees of success. But by and large, space 4X has stuck slavishly to a formula that, while charming and successful, thrives largely on fear of the unknown.
But I’m not here to bash the space 4X genre. Instead, my purpose is to examine just what MoO III got right – or came so close to getting right – and what ideas developers could glean from its legacy.
Empire Building & Automation
When MoO III launched, no mainstream strategy game better simulated how vastly, hugely, mind-bogglingly big space can be. Sure, newer titles dwarf its paltry 256 maximum star systems. But in 2003, that scope was well beyond most of the competition. Along with providing a bigger scale of empire building, Quicksilver aimed to give players broad and deep control over their territory. That, in theory, means lots of things to do, in great detail, spread over a huge empire. Sounds like micromanagement hell, doesn’t it?
Quicksilver tried to combat this by adding several mechanics to make this manageable, with automation as the linchpin. In most 4X games, players give specific orders: “Work this tile. Build this building.” That’s still present to a degree in MoO III, but in line with its big-picture gameplay, players are encouraged to set broader policies for their empire and let the governor AIs sort out the details. These policies include spending levels for different industries, labor policies, and military spending thresholds. The crown jewel, “Development Plans,” lets a player put planets into categories (like Core, Frontier, Large, Rich, etc.) and set priorities for each.
The idea of a game built around being an actual galactic emperor rather than a galactic workflow optimizer was pretty novel at the time. Even today, while many games claim to have “big picture” gameplay (almost as often as they claim to be inspired by Master of Orion II), few actually follow through with any real success. In theory, MoO III’s automation systems allow players to paint their empire in broad strokes, while letting the AI handle the nitty-gritty details like “How do I get the most minerals out of the ground while still keeping my people fed?” and “How many troops should I maintain in this garrison?”
Sounds great, doesn’t it? Which brings us to…
Why it Failed
The biggest problem with automation is one of transparency. At its core, MoO III’s automation isn’t the worst in gaming history. In fact, the Development Plans have a lot going for them. However, it makes some criminally stupid decisions, like building system colony ships (no FTL travel) in systems with no viable planets for colonization. But perhaps the worst part is that the classification of a planet is set and changed automatically by the AI, with no indication of why and no alert when it does. Players, however, can only manually choose from a small pool of player-defined classifications. So if you need extra food, and you want to change a high-fertility planet from Manufacturing to Farming classification… Well, it sucks to be you. On top of that, the game and documentation gives you almost no idea of how automation works or how to use it. This means players have no idea why the AI is doing the dumb things it does or how to get it to stop – other than by giving up on automation and handling everything themselves.
To make matters worse, shortly after launch, Quicksilver replaced most of the original developers with a fresh team, which had virtually no familiarity with the game or its code. (Famously, one new developer posted on the forums that the code was “spaghetti” to them.) This meant there was no one capable of maintaining and fixing the game, let alone communicating how it was supposed to work. As the game matured, players cracked some of the secrets of automation. With enough forum- and guide-delving, you can start to figure it out. But, commercially and critically, the damage was already done.
Those who do enable automation often complain that MoO III pretty much plays itself. Because so many of the decisions either require extraordinary levels of micromanagement or full automation, the player’s role as the “controlling power” behind their empire can quickly turn into repeatedly pressing “next turn” and engaging in combat (more on that later). The game doesn’t strike the balance between the hands-on and hands-off decisionmaking that gives players a sense of accomplishment and agency without bogging them down with micromanagement.
I believe, at its core, this is just as much a transparency issue as it is a gameplay flaw. In my experience, a player’s decisions do matter in MoO III, but the game does a terrible job of letting players see or feel the impact of those decisions. A little well-organized feedback can go a long way toward imparting a sense of accomplishment and pride. Unfortunately, MoO III’s interface goes out of its way to make that experience as alien and cumbersome as it can possibly manage.
Other Games That Tackle Automation
It would be a crime not to mention the crown champion of automation: Distant Worlds. DW is not a game for everyone, but it handles automation gloriously on all fronts. It’s frankly everything MoO III’s automation should be. It lets a player specify what game elements they want to control, what they want the automation to handle, and how they want the automation to handle it. Most importantly, it does a decent job of performing its tasks. (Those looking for smarter automation should check out Icemania’s AI mod.) While DW’s presentation of information is by no means user-friendly, everything you need is there. Players can feel and observe the effects of their actions, and are presented with meaningful decisions even if they don’t want to wade through the volumes of feedback and data the game provides.
A more recent entry that achieves a strong but more limited success in automation is Polaris Sector. The automation only extends to a planetary focus AI, but it does a good job of getting the most out of a planet while following the player’s instructions. It’s nothing revolutionary, but it cuts down on the tedium of managing a large empire.
Finally, Stellaris uses administrative sectors to push gameplay away from tedious micromanagement. All of a player’s colonies, apart from a handful of core worlds, are largely autonomous. This reduces the burden of managing large empires, although at present, many players feel they have too little control over their sectors or that they don’t make for a compelling mechanic.
Before we talk about MoO III’s economic model, let’s talk about its predecessor’s. MoO II’s vision of economy is very direct: pick up people and tell them what thing to make for you (food, industry, or science). Some buildings increase output of one of those functions. Ever since Master of Orion II hit shelves in 1996, this has been the go-to model for space 4X games and a growing number of their terrestrial counterparts that buck Civilization’s tile-based production system. And that’s fine. There’s nothing wrong with the system… But there’s also really nothing great about it, in my opinion.
MoO III takes its predecessor’s model and essentially turns it on its head. True to the title’s big-picture focus, its economics are managed at a higher level, with players setting DEAs, or Dominant Economic Activities, across entire planetary regions. DEAs are pretty much what they sound like: the region’s economic focus. Your empire’s output, therefore, is tied to how to develop your planets, rather than which box you shove your people into.
The DEA system has options for a few focuses you don’t usually see – such as government – in addition to the usual research, farming, etc. But the real win is that this systems rips out what used to be a tedious min-maxing chore (shuffling workers between jobs) and replaces it with a framework based more on long-term planning. Your DEAs aren’t a fluid asset like your workers. You can’t just flip an entire region of your planet from farmland to research laboratories with one click of a button. And unlike MoO II’s buildings, which you can spam without consequence, each DEA competes for real estate on your planet. The result is a system that’s much more about how you manage your empire than what color overalls your people are wearing. And that’s great.
Why it Failed
This system’s failure is mostly tied to MoO III’s overall problems with scope and automation, with its own twist: data oversaturation. DEAs are a great idea, but when you try to manage them, you get this:
Maybe it’s just me, but this is way too much technical data for me to easily comprehend. Sure, I can make sense of it, but repeating it hundreds of times across dozens of planets is too much work for too little gameplay. I admire that Quicksilver went for such a granular approach, simulating lots of real-life considerations. But this is a game where the player’s empire can easily span dozens of colonies spread across the galaxy and consist of multiple races with different capabilities and environmental preferences. Our poor brains can only take on so much at once, and fiddling with your DEAs is a taxing affair.
Paradoxically, MoO III’s focus on big-picture economics creates the very micromanagement nightmares it sought to avoid. The idea was that the game (and players) could handle this level of detail, with the player just giving it broader orders. Unfortunately, because of the sorry state of the automation systems, the player ends up taking the full brunt of that data blast straight to the brain.
Other Games That Tackle Economic Development
Galactic Civilizations’ tile-based economic system means players have to plan ahead on how to develop their planets, while mercifully abstracting most of the little details down to a couple of numbers. Polaris Sector features strategic-level planet management and good planetary automation. While its output model isn’t anywhere near as deep, it approximates the same effect without breaking your brain.
Endless Space scraps planet-level management entirely in favor of system-level development. The very first Master of Orion and the original Sword of the Stars took a similar approach, abstracting economic management to simple sliders.
Unlimited Power, Limited Grasp
The last empire-management mechanic that deserves mentioning is Imperial Focus, and a marvelous mechanic it is!
Here’s the idea: you, the “controlling power behind your civilization” (whatever that means), have unrivaled breadth and depth of influence over your empire. But, you’re limited in just how much you can do per turn. As Quicksilver repeated time after time, “You can do anything, but you can’t do everything.” Imperial Focus was touted prominently and repeatedly by the design team as “…the ultimate preciousresource [sic] that a player must manage.”
The entire game was clearly structured and designed around Imperial Focus and its impact on empire management choices. Micromanaging your worlds would take up Imperial Focus Points, forcing you to make choices about where to focus your development.
Now, this concept of limited time isn’t something new to gaming. RTS games do the same thing, except the limitation is how many actions you can make per second instead of how many points you have. Board games often limit each player’s actions per turn to one or several out of many possibilities. Simulation games also generally limit the number of concurrent tasks in some way, and RPGs are all about action economy. But amongst turn-based 4X video games, it’s virtually unheard of.
I can understand that some people don’t like this system in 4X games, or maybe don’t like the prospect of it. But it can do so much for gameplay. Action points of this sort are a currency that functions on a completely different axis than your usual food, minerals, money, etc. They can apply to a huge range of different mechanics, from specialist vs. generalist economies, to diplomatic interactions, to marshalling forces, to changing government types. Games become a dance of give-and-take between multiple mechanics and priorities, forcing the player to make compelling choices. If you think about it, this concept can cut down on micromanagement, as each player (AI included) has to prioritize optimizing workflow on Asteroid 10307 vs. setting a new research priority.
In addition, a system like this can be used to balance game mechanics. As opposed to MoO II’s Command Point system (a major weakness of that game in my opinion), an Imperial Focus mechanic means that larger fleets consume more points to order about. Their size is balanced out inherently without tacking on arbitrary fleet size caps or penalties. Leveraging that big fleet either drains your ability to manage the rest of your empire or requires its own specialized infrastructure. This system is slick, brilliant, and easily one of the biggest contributions MoO III made to the genre.
Why it Failed
MoO III’s Imperial Focus system failed because it doesn’t exist.
In April 2002, the team announced that they’d cut the system from the game entirely. Although this was almost a year before MoO III’s launch in March 2003, the system was part of the foundation of the entire game. Accommodating such a seismic shift would have required a fundamental rework of much of the game. As such, Imperial Focus was simply chopped off, without anaesthetic or even antiseptic. (For those interested, Bruce Geryk talks a lot about how the loss of Imperial Focus mangled the gameplay in his article for Wargame_[space], which is also available at Quarter to Three.)
As tragic as the loss was, I have to admit that Imperial Focus probably would not have worked for MoO III anyway. Combined with the sorry state of the game’s automation, Imperial Focus very easily could have piled a frustrating game mechanic on top of simple tedium. This would have locked empire optimization behind even more barriers. Hopefully, others will pick the idea up and do it justice, backed up by competent automation AI.
Other Games That Tackle Strategic Limitations
The Romance of the Three Kingdoms series has been using this system for ages. The upcoming Imperia looks to feature a similar mechanic with Edict and Admin systems. King of Dragon Pass, a game from 1999, had a similar restriction on actions per turn and it remains one of the hallmark features of that game. Finally, tons of games from other genres have this mechanic or something that essentially boils down to it.
Like a lot of teenagers, I enjoyed the hell out of reading Ender’s Game. One thing that really grabbed me was the role Ender played in the space engagements. He gave broad, high-level orders which would then be carried out by his commanders, as opposed to the unit-by-unit control we usually see in strategy games.
MoO III sought to capture that essence of being the fleet admiral, setting the policy and leaving your subordinates to carry out your commands in real time. Here’s an interview with Tom Hughes from 2001 on space combat. (Optional reading: an article on sector seats, a nifty supporting mechanic.) Players would organize their ships into individual Task Forces, each with a defined mission (like Short Range Attack, Long Range Attack, Carrier, and Indirect Fire) and with core, escort, and picket rings. As for executing ship combat, each side would orchestrate the combat from a higher level, delegating the specifics to individual commanders. As Hughes said, “Think of what Yamamoto or Nimitz was feeling, and that’s the kind of experience we’re shooting for on a higher level.” It was pretty exciting stuff for those who were looking for more depth in the space combat department without having to issue orders to dozens of units each turn – a common complaint with MoO II’s late game.
Adding to the combat system, MoO III has a robust ship editor. Whereas MoO II only lets you choose “yes/no?” and “which one?” for ship components like shields and engines, MoO III gives you many more options, like size of thrusters and shield generators (along with class). MoO III also includes even more weapon modules than its predecessor. For example, players can set up ships with long-range vs. short-range guns (and specific AI for each) and choose from an expanded list of mounts including the spectacular Spinal Mount. All these inflection points make design and combat a lot more engaging. And the task force system means players can easily leverage specialized ships or go with a more balanced, generalist approach.
Why it Failed
You know that awesome Yamamoto/Nimitz/Ender’s Game system I mentioned? Guess what happened to it…
That’s right! It joined Imperial Focus (and sector seats) on the cutting room floor late in development. Task forces still exist, and they’re good at what they do, i.e. letting the player group ships together into formations for easy management. But instead of providing a framework for orchestrating the battle, players are given the two standard choices: “move” and “shoot.”
And therein lies the problem: Quicksilver removed the original mechanic, but failed to replace it with anything that gave the player interesting decisions to make. Absent are the original design concepts of scouting and probing an enemy fleet, or proceeding cautiously or recklessly, or setting engagement ranges for your different weapon types – the orchestration part of the equation. Without these options, combat feels limited to clicking on enemy task forces in the order you want to see them explode, or hitting the retreat button. Granted, some people enjoy the ship combat, and I’ll admit that it’s not the worst ever. But just like with Imperial Focus, it’s very clear that something integral was removed too far into development and the team didn’t have the resources to design a full-fledged, compelling replacement.
Other Games That Tackle High-Level Space Combat
Distant Worlds has the same sense of giving broad orders to your ships and then letting the tactical AI sort out the details based on those orders. However, the interface is rough and the (few, basic) orders are tied to ship designs, so to change orders you have to refit your fleet. The same can be said of many grand strategy games such as Crusader Kings II and Europa Universalis IV, along with Stellaris, whose lack of granular control usually means giving broad orders with little say in combat on the tactical level.
Several recent titles – StarDrive II, Star Ruler II, and Polaris Sector – do ship design to a similarly detailed level, but with tile-based layout systems. Each also features real-time combat. Players in both StarDrive II and Polaris Sector command fleets on a ship-by-ship level, but Star Ruler’s fleet system also captures some of the taskforce feel. Galactic Civilizations III makes an effort to deliver engaging high-level combat, but with its oversimplified dynamics and few meaningful player decision points, it fails to provide anything of the sort.
In most space 4X games, ground combat is pretty much a sideshow. It’s designed to take as little time as possible so that the player can get back into the meat of the game, like empire management and space combat. In that respect, most ground combat systems are largely successful in their simple and abstracted approach. MoO II is a paragon of this method – battles consist of a couple of supremely cathartic seconds of tiny soldiers blasting each other with no player input.
In its inception, MoO III’s ground combat sought to provide a compelling, worthwhile mechanic in and of itself. While Quicksilver completely gutted some other mechanical systems, with ground combat, they merely amputated most of the limbs. The skeleton of the system, the Battle Plan, is still present – with its share of problems.
Battle Plans were an attempt by Quicksilver to give the player high-level control of a large-scale planetary invasion. They sought to do this without miring players in the minutiae or abstracting the affair to ludicrous levels. Most of winning the battle happens on the empire level – training your troops, building transports, picking targets, that sort of thing. Once your boots are on the ground, it’s time to light up the corn cob pipe and get to strategizing. Players set battle intensity, nuclear/chemical/biological weapon authorization, collateral damage tolerance, and a specific plan of attack or defense – things like Pronged Attack, Feint, Attrition, Ambush, etc.
So the stage is set and the pieces are in place. By our projections, this system should be a resounding success, right? Well, you know what they say about plans and contact with the enemy…
Why it Failed
The fatal flaw at the heart of MoO III’s ground combat is its utter lack of transparency, with a side order of self-destructive randomness. Players can initiate the third stage of their campaign on a high-gravity, heavy-atmosphere world, only to find that the combat is somehow now taking place in a zero-g vacuum. As far as I can tell, this is determined randomly. Not procedurally generated, but really, truly randomly. Procedural generation means creating data according to set structures. Random means rolling the dice and going with whatever turns up.
This randomness is detrimental to the game not only in terms of strategy – you can’t plan around randomness – but also in terms of immersion. The gameplay feels arbitrary, which crushes any chance of engaging the player strategically or tactically. Maybe there is reasoning behind behavior like this, but if it exists, it’s never communicated to the player.
The lack of transparency is amply apparent in the execution of the Battle Plans. Players are given no feedback on the impact of collateral damage and N/B/C weapons, and not the slightest insight into how maneuvers interact with the enemy’s. You won’t know how your Battle Plan stacked up, or if one is superior to the other and in what circumstances. You won’t even know if the Plans are making a difference with all the interference from the other (seemingly random) factors at play. I imagine that there’s an underlying logic to how the system works. However, again, it feels completely random, robbing the system of any semblance of strategy or tactics.
I spent hours searching for information on the effects of the different combat maneuvers. Despite the impressive official documents, I never gained the slightest speck of insight into the impact of my decisions – especially the choice of Battle Plan. That was complicated further by the fact that so much of the original system was clearly scrapped after those documents were written – so who knows how the system actually ended up working. And since the original developers were replaced early in the game’s post-launch lifecycle, no one on the inside can provide the community with an explanation.
Other Games That Tackle High-Level Ground Combat
Polaris Sector features lengthy ground campaigns and a high degree of transparency, although they have no tactical considerations beyond rock-paper-scissors. Galactic Civilizations III seemed like it wanted to go in this direction, but cut the system at the last minute – and with disastrous results.
Other notable elements
For all of MoO III’s unrealized ambitions, which might still merit study by future 4X designers, the game does do some things right. In fact, some of the systems I haven’t mentioned yet work really well and, as such, are less controversial. It would be a shame to entirely gloss over them, so I’ll briefly mention a few here.
The Antaran Xs are powerful secrets locked away by the resident abusive precursors, and they serve as a victory condition. To find them, players mount expeditions. Each X you find unlocks a powerful bonus for your empire. A player had to collect all five to claim the win, and Xs in possession of another empire could be acquired via trade, military conquest, or espionage. While the victory condition itself is a straightforward “gotta catch ‘em all” type, the bonuses they provide and the vectors for inter-empire interaction make the whole thing quite appealing.
The research system is more of a mixed bag, but has some interesting ideas. Rather than choose a research project to pursue, players divvied up their research efforts between six fields of study: Physical Science, Energy, Mathematics and Computers, Economics and Business, Social Arts and Sciences, and Biological Sciences. Advancing in a field gives you access to the technologies in the newly-unlocked level. It also borrows a concept from MoO I, in that each individual technology has only a certain percent chance to be available to any given empire on any given playthrough. This chance is influenced by how creative the empire’s species are. While not a resounding success, the more organic approach is novel, and far from the worst mechanic the genre – or MoO III – has produced.
Diplomacy with other species is a pretty standard affair. But it does have a nice twist: you can choose not only what to propose or demand, but also how you say it. So if you feel like sucking up to your powerful Trilarian neighbors, you can take a deferential tone with them. If you’re negotiating a trade with the Sakkra, you might decide that they’ll respect a firm approach.
It’s a neat mini-game but with its own share of flaws. There’s no feedback on what effect your tone produces. So at best, you have to play guess-check-revise, and at worst, you can’t sort out the signal from the noise. And some of the diplomacy dialogue the game pastes together is downright laughable.
The Orion Senate is another neat gameplay element. Along with serving as a victory condition (as in MoO II and many other games), it actually puts some important, high-impact policies up for vote. Members can also wield the Senate as a weapon against their enemies, isolating them politically and economically. While the Antarans monopolize the power at the start, real multilateral action through the Senate becomes possible as the other empires grow in power.
Finally, MoO III introduced a casus belli (cause for war) system. This has been a staple of Paradox grand strategy games for ages, but very few traditional 4X games use it. A common complaint is that the Paradox casus belli model is too granular and low-abstraction for 4X. That view is completely reasonable, but MoO III proved over a decade ago that you can have a very 4X-appropriate, high-abstraction casus belli. An empire has a casus belli for each other empire, which represents how its citizens feel about that enemy. If your population really likes your Silicoid neighbors, and you declare war on them, your people will get really unhappy. It’s slick and very 4X. Its biggest fault is that there’s very little a player can do to influence casus belli, but it’s a great blueprint for future games to consider.
Like research, this one splits the community. MoO III’s art director, Rantz Hoseley, was an unabashed hater of the rubber forehead aliens look of many MoO II aliens. As a result, MoO III’s style is much less cartoony, and much more alien. Here’s a few side-by-sides:
Some of the more out-there aliens rubbed a lot of people the wrong way. (Check out the Nommo, the Ethereal species, or… Whatever the hell this thing is.) But personally, I loved the really alien aliens. I like it when sci-fi presents me with things that are foreign or confounding in some way. And to me, they look a lot better than MoO II’s Rob-Liefeld-esque slabs of space-meat. The somber aesthetic supports the game’s overall tone, and the species’ personalities and characteristics are strong enough that I can identify with them.
I won’t pretend that all of Master of Orion III’s failings were on the execution end. A lot of its design decisions were at best controversial and overly-ambitious from the onset. As such, they bear some of the blame for nearly killing the franchise.
However, I firmly believe that as frail as the body was, the game’s DNA had a lot going for it. As a finished product, the game is a heaping mess, but as a design document, it’s a treasure trove. Too often, I’ve seen good ideas for 4X games dismissed on the basis of “MoO III did it.” My hope in writing this article is to push back against that stigma, and to put these innovations back on the table for discussion and consideration – and maybe even adoption. MoO III cast one hell of a cloud over the franchise and even the entire 4X genre. I’d love to see what kind of tapestry a studio could spin from such a magnificent mess of silver lining.