Some thoughts about the 'early game' phase of RTS

This article appeared originally in my game design blog. You can see the (somewhat prettier) version there if you are interested.

The 'early game' is a critical part of any RTS. Starting as soon as you load into the match, it extends... Well, I talk about that a bit below. Anyway, it's the first and often most-constrained part of an RTS match and as such it's desperately important to get correct.

It's not just a first impression to your game, it's a first impression that gets made again every time someone loads into many single player missions, any skirmish match, or any ranked/ladder/tournament game. It's experienced again, and again, and again, and if it provides a sub-par experience it's going to be really, glaringly obvious.

So, it's something that's going to need to be done correctly in order to give players a good experience. It's like... finding sand in your food. Once, it's going to maybe be a little annoying, ruin your meal. But if you find sand in your meal every time you eat, it won't be long before you choose a different restaurant.

But, uh, anyway... What does 'correctly' even look like or mean in this context? Surely, there's a ton of different ways to make an RTS and therefore a ton of archetypes for a good intro experience to a match? As a cop-out, yes, that's true. But I think there are some general rules that can be gleaned from a variety of games and situations that can serve to improve almost any 'flavor' of strategy game.

To me, the early game needs to guide the player to do 3 things: figure out what their enemy is doing (scouting or intel), commit to a general strategy (build order), and set themselves up for success going forward (player guidance). Also, ideally, the early game needs to be fairly 'stable' or 'robust' to make it somewhat difficult to knock out a player or team during this time.

Let's take a look, shall we?

First off, Let's Define Our Term

"A beginning is a very delicate time"

- Frank Herbert, Dune

I always like to start with a definition, even in a more off the cuff piece like this one is supposed to be. It's a good jumping off point; it's a good way to figure out the parameters I'm working within and to help me keep in mind what might and might not apply.

The early game is the beginning of an RTS match, of course: it encompasses the first moments/minutes of gameplay. The end of the 'early game' phase and the beginning of the 'mid game' is a more nebulous thing, however. Does the early game end when the player starts to loosen their economic constraints? That's what my gut is telling me: once the player begins to feel some lessening of economic pressure in the game, and is able to attack and defend, or to expand their income and expand their army. This of course will vary across different games with varying styles of gameplay: in Company of Heroes 2, for instance, the player's economy doesn't really 'take off' the way it might in something like Age of Empires, and so there's not really a 'lessening of economic pressure' in the sense I'm using above.

Command and Conquer games also all have kind of truncated early games, and several points which might signal the end of that phase. This is, in part, due to the fact that C&C starts players off with a significant cash reserve. Also, in C&C there's no limiting factor on the number of troops you can produce at any given time, since 'supply' or 'houses' are not existent as they are in games like StarCraft or Age of Empires. So there's no real limit to the rate at which the army is able to expand. This, in turn, leads to a lot of situations with hugely disparate army sizes. Later C&C games help address this with their 'rock/paper/scissors' damage system, but in C&C95/remaster it's often deadly.

Perhaps the 4th Refinery (in standard play) or the construction of an Airstrip or War Factory might approximately end the 'early game' in C&C Remaster/Tiberian Dawn. Or, when (if) the player has enough spare buildings that losing one of them isn't completely crippling (e.g. second War Factory and 5/6 refineries). But with a comparatively old RTS the context might be a little weird, since a lot of the 'late game' stuff isn't really competitively viable or used. the MLRS or Nod SSM are almost never seen in PVP matches, for instance, and even the Obelisk of Light isn't used in more than one game in 20, if that often.

But, I suppose I'm digressing a little bit: the specific mechanics of when the 'early game' becomes the 'mid game' in any given RTS isn't really the point i'm trying to make. And believe it or not, I actually kind of do have a point this time. I'm trying to advocate for a better early game experience, and describe what that might look like. But I'm trying to work my way around to the point in a logical manner and let's be honest it can be kind of easy for me to take some rabbit trails along the way...

Going to try to focus up and get to the point.

Next Up: What's the Point?

Now that we have a rough working definition of what the early game is, and we kind of know our trajectory the next step for me is to define the purpose or function of this phase of the game. To remind of you what I said above, the early game needs to guide the player to do 3 things: figure out what their enemy is doing, commit to a general strategy, and set themselves up for success going forward. Also, ideally, the early game needs to be fairly 'stable' - you should be able to lose in the first couple minutes of a match, but it's too often that this is kind of too easy to do. Losing in the early game can kind of feel crappy, especially if it's really possible to be stopped or rushed before you are even really able to tell what your opponent is doing (to say nothing of stopping what they're doing).

When I talk about 'guiding the player' I mean that the structure of play needs to set up expectations for the player and not confuse them. Let me try an example for clarification: in C&C Tiberian Dawn, the 'tech tree' does little to guide the player's experience. There's no considered flow between the buildings, not really (in my humble opinion).

Another example might be Act of Aggression, where for the Chimera faction at least, the tech tree didn't really 'guide' the player. I, personally, found it kind of difficult to grok what exactly was the game expected me to do next, which to me at least is a really un-optimal (and frustrating) experience. Warparty was a similar experience for me, where buildings in the build menu were hidden until they were available, making it impossible in-game to know what to do next, what to upgrade, that sort of thing.

This, in terms of advancing up the tech tree at least, feels really bad to a new player and can sour the experience. Even for a more experienced player I'd assume it could be of help: it's the difference between knowing your house so well you can navigate it even in the pitch dark, and being able to navigate your house in the dark but having light to see by. Either way, having the 'light' of seeing all of your options makes for a smoother experience.

Information Gathering

Two of the biggest problems with the 'early game' are: getting information, and allocating resources in the right place at the right time . Or, to use some more common vernacular from players of the genre, it's scouting and your build order. Or, to abstract it even more, it's to answer the questions: what is my strategy, and what is my opponent's strategy?

A lot of players will go into a match in an RTS knowing what they're going to plan to do. Some players just have a 'favorite' way to play: they know how to get a decent sized army out quickly and they'll just smash their force against the enemy while he's still building up. Or, they've managed to work out the execution of a 'cheese' strategy (I intend to spend a bit of time talking about cheese strategies, don't worry). Or, they like to build up a big base and a big force and will try to hole up as much as possible and 'turtle' their way into the end game. Et cetera.

Many competitive players will have 'breakpoints' in their strategy. They know they can generally learn what their opponent is going to be trying to do by such and such a point into a game, and based on that knowledge they can expand and 'macro' to a larger, higher tech army, or push in and disrupt their opponent's plans. In other words, these players have a general strategy (probably based on the particular map they're playing on) for their early game until they can gather information about their opponent's plans, then will adapt into one or another direction in order to handle what their foe is up to.

This is, of course, the ideal scenario. And it's why in high level RTS multiplayer games almost across the board, the gathering and prevention of intel is such a big thing. I mean, really, a fundamental thing, foundational. More synonyms, probably.

Dedicated scouts help the player understand and remember that they need to be keeping tabs on their opponent

One of the worst feelings in RTS to me at least, is feeling like I lost without getting the opportunity to even find out what my opponent was doing (and react to it) before it hit me. It's a delicate balance, of course: rushing and cheese, as much as some players are loathe to admit, are actually a needed part of the RTS landscape.

It is of course the timing of the various early attack options, and the balance of the difficulty of pulling them off versus the difficulty of scouting and defending them, that makes all the difference. If a Protoss player in StarCraft 2 can sneak a Probe into a Terran player's base and pull off a successful cannon rush that ends the game or puts the Terran player desperately behind 75% of the time, there's going to be a lot of that going on. Conversely, if it was only possible to succeed at such an attack 10% of the time, you'd never see anyone do it. The balance point of that strategy is somewhere that makes it risky and challenging to execute, and simultaneously rewarding enough when done with the proper timing and precision and skill. Bonus points are awarded, of course, if there are good partial success and failure states for each player.

It's good, necessary even, to have the ability for a player to succeed through early aggression. But, having games feel like a dice roll whether you're just going to lose about 2 minutes in, that's really bad. This is one reason why things like included scout units in Age of Empires games is a phenomenal thing. In Age of Empires 2, the initial scout is an indespensible part of gameplay for a variety of reasons.

Another great example of early scouting is in Red Alert 3, where units like the Attack Dog and Burst Drone are specialized to be able to get around the map quickly and act as scouts. They're not available from the start of the match, which to me is a bit of a disadvantage to Age of Empires 2's dedicated scout unit, but are generally able to get some scouting done or contest areas of the map in advance of many threats or strategies. Red Alert 3 has several other fairly early units that can be used for scouting: the Soviet Terror Drone and Sickle make good scout troops as well, and the Allies have a very early support power that makes it pretty easy to keep tabs on their enemy.

In particular though, a dedicated scout unit really helps the player to keep in mind that they can and should be out on the map seeing what their enemy is up to. Personally, I see this as better than the WarCraft/StarCraft model of committing a worker, and potentially putting yourself at an economy disadvantage, in order to get early intel. As a bonus, a dedicated early scout unit can provide some needed interaction and interest during the often slow ramp-up time at the start of the match, when the player is waiting for stuff to build, or accumulate, et cetera.

On the 'Stability' and Duration of the Early Game

So, what do I mean by 'stability' in terms of an early game?

Well, first and foremost, as I said above, the early game in many Command and Conquer titles is really volatile. In Tiberian Dawn, it's not super uncommon to lose like 2 or 3 minutes into a match due to flame trooper/hand of nod rush, or masses of Attack Bikes or Buggies, et cetera. Classically, it was also very common to see APC/Engineer rushes, due to the multiple benefits of capturing an enemy building to sell right away and the tantalizing chance of scoring a ConYard steal and preventing your enemy from building any more structures. This is a prime example of an early game that is not stable. It's really easy to derail a player in a variety of ways, and it's not easy to defend yourself from any of them, and it's also quite difficult to scout these early strategies until they're already hitting you.

Legacy of the Void's economic changes over Heart of the Swarm and Wings of Liberty prove to be an interesting case study in terms of RTS early games. Legacy of the Void, I believe, doubled the number of starting workers from 'classic' StarCraft. There were a host of other changes as well, but this in particular was of interest to me as a designer. Starting the game with more income potential truncated the early game dramatically.

Since the earliest stage of the game is when players have the hardest time reacting to unexpected threats, truncating the early game allows for players to have a more robust defense and offense earlier than previously. It also allowed some rushes and cheese to happen earlier, but in general players have more options and more money to react to their opponent sooner than before, allowing for more robust trading of blows without being crippled. Among other things, it shortened the areas of greatest weakness and vulnerability (while at the same time allowing for larger attacks to happen sooner).

There's also the matter of the 'QuickScript' mod in Grey Goo. Community leader XCet of the Grey Goo Fan Discord "Grey Goo Hangout" created a mod that, through the simple act of refunding the cost of the player's first refinery, cuts what is normally a roughly 7-minute windup in the early game in half and dramatically changes the shape of a competitive Grey Goo match for the better.

In short, I feel like a relatively quick economic ramp-up is one way to achieve some stability, especially in conjunction with other efforts.

On the flip side, Company of Heroes 2 has a relatively stable early game. Losing a squad is pretty bad, especially early in the game, but it's also fairly uncommon if you're paying attention. Also, a lost squad or two is probably recoverable due to the fact that the game's manpower resource is accumulated more slowly the more units a player has under their command. In COH2, you're mostly losing the negligible cost of a model in a squad and travel time as the squad retreats back to your base/starting area, then heals up, then walks back to where you need them to be.

The other thing that Company of Heroes has going for it in terms of 'stability' is that there's a lot of stuff to focus on instead of just the other player.

As with everything in terms of optimal RTS design, the length of the 'early game' is a matter of somewhat delicate precision. As Command and Conquer Tiberian Dawn/Remaster shows us (I covered this above) if it's too short, and there too many options that can be pulled out before they can be scouted or defended, it can really hurt a game. No one wants that 'dice roll' on match start to see whether the game will get interesting.

Thinking of 'stability' in the early game, the upcoming RTS 'Immortal: Gates of Pyre' has capturable turrets on the map in various places. Notably, each player starts with one of these turrets in their base, already owned by the player at match start. These turrets serve a role similar to the Age of Empires Town Center, helping keep the player's base safe during the early game when unit counts are still low. WarCraft 3, with Ziggurat turrets, Night Elf Ancients having attacks, garrisonable Orc Burrows and building spikes, et cetera, has similar protection for bases in the early and mid-game, though mostly these defenses don't really stem off a dedicated force in a meaningful way, that I've usually seen.

My general rule is that the economy should ramp up relatively quickly, while the player's ability to get crippling damage into their enemy's stuff should ramp up more slowly, and be balanced against the ability to get scouting information on what your opponent is planning. Of course, I repeat that Company of Heroes 2 is an example of a much more protracted early game that plays out in a generally positive manner due to the player's ability to keep their investments from being destroyed.

Refining the Early Game, A Summary

"The beginning is the most important
part of the work"

- Plato

Above, I try to lay out some of the various strengths and weaknesses of the 'early game' phase of RTS. This is really important to strategy games in general since they're one of the only genres of competitive game with such a long and uncertain ramp-up. In a shooter, you don't have a 5-to-7 minute window before you're really ready to take on your opponent, and being shot doesn't (usually, anyway?) reduce your ability to participate in the game for the rest of the round.

Length: it feels like a short early game in terms of economic expansion, and a slightly longer game in terms of crippling offensive options, allows for the most robust counter-play between players while still feeling 'fair.' Couple this with a built-in scouting option similar to what is seen in the Age of Empires games, and the players are well-equipped for an early game that can have encounters and exciting play, but is less 'unstable' in terms of how easy it is to be knocked out, increasing game equilibrium.

Fairly robust defensive options early can help as well. Consider the Age of Empires Town Center, which serves as effective army deterrent in the early game but doesn't scale well into the mid and late game (where walls, trebuchets, huge army sizes, castles etc come into play). And lastly, in general, I am much in favor of relatively straightforward in terms of offered buildings/options/unlocks, that present the player with a 'pathway' into the mid-game and do their best to limit false choices and player confusion.

If I have additional thoughts later, I might revisit the topic. For now, I worry I've gotten overly wordy. Thanks for reading, and I hope to see you on the battlefield!

This article appeared originally in my game design blog. You can see the (somewhat prettier) version there if you are interested.