Nomenclature and Processes in Real Time Strategy Games: Part I

Hey folks,

Today I’ll be talking about nomenclature and processes in real time strategy games. Over the course of two blog posts, I’ll be discussing the following:

  • Part I: Formalizing concepts through the establishment of precise and widely-known nomenclature
  • Part II: Formalizing previously ad-hoc activities with well-understood processes (“standard operating procedure”)

I’d like to focus on two things: why this stuff matters and how it can influence the genre in the days ahead.

“Nomenclature and processes” simply refers to taking ad-hoc, informal ways of naming and doing things and transforming them into well-understood and precise concepts and procedures. This can happen as part of the natural development of the genre or it can be done intentionally.

Here’s a simple example to illustrate: how does a player determine why a game was lost? There are lots of possible answers to this question, but as the genre has developed, players have learned that watching the replay is typically the optimal first step. Here’s a few reasons why:

  • Replays are documented proof of a player’s missteps, allowing them to see conclusively where they could have done better.
  • Replays can be shared with others, allowing feedback to focus on a specific instance rather than vague, difficult-to-answer generalities.
  • Watching a replay reinforces a mindset of playing a game whose result is based on the player’s performance and skill within that game, rather than a reflection on their character.

This simple process has become widely accepted, leading it to become integrated into existing workflows.

2016-09-01_defeatScreen.png

An important takeaway is that new players benefit from this insight without needing to play hundreds of games and attempting the myriad ways of learning from a loss before finding the optimal one. They don’t need to be aware that the underlying design decisions – do replays exist, are they automatically recorded, how easy is it to watch a replay once a game ends – are based on insights gleaned from years of maturation within the genre and were, at one point in time, not cut and dry conclusions.

In more general terms, this formalization of names and processes enables people to 1) identify and categorize problems (“why did I lose this game”), 2) leverage well-understood processes to solve those problems (“watch the replay and identify your mistakes”), and 3) establish benchmarks to evaluate process efficacy (league/MMR/win rate over time in same situation/tournament results/etc).

Consider long-distance running. There are efficient, effective training processes that can be reliably repeated by most people. There are best practices that have practical basis (e.g. preventing injury, improving times, etc). There are volumes of literature on known stumbling blocks and how to deal with them. It would be unusual to train for a marathon by “doing your own thing” and not consulting these established resources; most people understand that there is a well-trodden, more-or-less optimal path to doing it. Centuries of competitive track and field have built a knowledge base and underlying infrastructure that everyone can benefit from.

Real time strategy games had not been formalized in this way when they first became popular. (Even now, it is still very much early days). The formalization of the genre over time, driven partly by the growth of electronic sports, has had a profound impact on the entire player base – even casual players – and how individual players approach these games.

One of the most profound changes has been the adoption of precise nomenclature that enables players to speak accurately about the game. Strategies in the early days were often bucketed into massive categories, like “rush” (focusing on early military units and attacking quickly) or “boom” (focusing on economy and building lots of workers). As the genre matured, this broad terminology was replaced by more accurate nomenclature, with individual terms getting broken down into more precise pieces.

Let’s think about the difference between a “rush” and a “timing attack”.

  • A rush focuses on early military units and attacking quickly.
    • Which units?
    • How do we define early?
    • What is the goal of our attack?
  • A timing attack seeks to attack the opponent at a specific time with a specific set of units in order to achieve a specific goal, hoping to exploit a specific timing weakness in the opponent’s build (e.g. prior to an upgrade) or a specific timing strength in your own build (e.g. very fast max supply).

Lots of important things happen when we start using the more precise terminology. The strategy is more repeatable, making it more consistent and refinable. This also makes it easier for the opponent to wrap their head around it and develop counter-strategies, or for the development team to identify a balance problem. Players who fail to execute the strategy can compare their execution against the ideal and see where they went wrong. The clean and efficient allocation of resources to hit a specific timing highlights the trade-offs players are making when executing this strategy, allowing them to better understand their situation in the game.

The most important aspect of specific nomenclature is that it it forces the use of explicit language by pointing out what’s assumed rather than just letting it be. Stating that you plan to do a marine rush sounds like a perfectly fine idea. Stating it as a timing attack forces you to think through each aspect of your plan and consider what it will accomplish. This will lead you to realize that you probably need to proxy it for it to be effective; this will lead you to realize that it’s a pretty all-in strategy; this may make you consider what kind of strategies are effective counters and how to prepare for them, whether this is a style of play that you enjoy or whether it’s even effective in the first place.

Not all rushes are timing attacks. Three-rax-reaper is a good example of a rush that doesn’t seek to exploit a specific timing window. Neither can all timing attacks be defined as rushes. The term has a precise meaning which reduces confusion and structures thinking in a helpful way.

By comparison, discussing problems in a very general way tends to make them intractable. Consider this interview with The_Sheriff, an expert player in Age of Kings, conducted in the early 2000s (ignore the date on the article), where he talks about the development of the metagame:

Xiphoid – In many of your games, assuming no flush, you sport some pretty impressive early castle times in the 14 and 15-minute range. Does this help your booming or defending of castle attacks? Many players have asked if it’s better to do one of these lightning-fast castles as opposed to 17 and 18-minute ones. Your ideas on this?

The_Sheriff – I prefer the quick castle strategy because its less vulnerable to a feudal rush. If I can get to castle age at the same time my opponent starts to attack, I can set up 2 more town centers that will give cover to my entire population. If you wait until 17 or 18 minutes, then you have to fend off an attack for 2 or 3 more minutes with just a single town center.

The lack of specificity makes it hard to think about the strategies described and identify their trade-offs. What does less vulnerable mean – what are we vulnerable to and what are we not vulnerable to? Why are the time ranges so broad? The problem – defending against a feudal rush – is not concisely stated, so it’s difficult to solve it. When we structure our thinking using precise concepts, it forces us to see that there are gaps in what we are saying; it forces us to state our assumptions.

One effect of this is the democratizing of the multiplayer scene. Formal nomenclature comes about to describe concepts that top players already knew about but needed words to express. The best Age players at the time probably knew exactly what The_Sheriff meant because they had done these strategies thousands of times (and probably talked about it with him directly). New players don’t have that kind of experience under their belt and can’t think back to past games to comprehend new concepts or understand the nuances of things they already know. They need words to wrap their heads around these ideas.

Here are some specific examples:

  • “stutter-stepping” teaches the player attack move if they didn’t know about it. It forces the player to think about unit positioning, unit reload time, and the nuances of how units behave when they’re not being controlled
  • “map awareness” teaches the player that they need to be aware of what’s happening on the map and consider their army’s positioning and home base’s defenses relative to this. It subtly rejects the notion of a “surprise” attack, putting the responsibility on the player to scout.
  • “opener” teaches the player several important concepts – that build orders exist; that their mid-game is connected to what they do in the early-game; and that what they do in the early-game is not their entire strategy, it’s just an opener

The crucial point is that the mere existence of the word structures how the player thinks about the game. It forces assumptions to be stated, describes concepts precisely rather than vaguely, ensures repeatability, enables incremental improvements, allows for the fluid development of the metagame, etc.

I think anyone who’s learned a second language can attest to the experience of understanding an idea better in the new language than in the old, or the feeling of needing paragraphs to explain what the other language can express in a single word. Language helps structure our thinking – it provides tools for processing ideas.

Sometimes, the mere existence of terminology can create a paradigm shift in the way a player approaches the game. Consider Day9’s Mental Checklist. It is a profoundly powerful way of playing real time strategy games – a repeatable and practical approach to managing tasks in an efficient manner with consistent results. It is also very different from how most people think about real time strategy games. It’s bottom-up (do this process as fast as possible until the game ends) rather than top-down (consider what’s happening in the game at any given time and decide where to spend attention).

As a result, the term includes within it a perspective that influences how players think about the game. It places emphasis on real-time; most people naturally place emphasis on strategy. It can feel uncomfortable to practice the mental checklist for an entire game if you’ve never done it before.

Historically, experienced players inched their way toward this approach across thousands of games. For them, it didn’t feel like a paradigm shift, just natural progression and improvement. Formalizing the nomenclature opens up the possibility for players to attempt it instantly – this is jarring.

I think this is one of the reasons players get frustrated with StarCraft II. I think we’re at a point in the development of this genre that many concepts have become formalized, but the process of applying those concepts has not. Previously, game understanding and mechanics would develop together – players would play a lot, blindly stumble into the mechanics of “map awareness”, then retroactively put a label on it. The proliferation of formalized concepts has changed that; players’ game understanding will often outpace their mechanics. This can be difficult to accept. Realizing that one needs “map awareness” to stop drops – but not understanding how to develop map awareness – can be more frustrating than just not knowing anything at all.

A key area where the genre can grow is enabling players to take concepts – e.g. “split” against banelings – and learn how to develop them. Much like someone playing pick-up basketball after work might want to run some drills to work on their shooting, so too might the average StarCraft player. This means developing processes – more on this in part two.

Recognizing the use of formal vocabulary can also cause us to notice where formal vocabulary is not used, perhaps because it doesn’t exist yet.

For instance, one interaction in StarCraft II that players complain about is when the opponent does something that’s mechanically easy but strategically powerful, like waypointing a Liberator to a mineral line and sieging on it. There isn’t a good word to describe this, or at least I don’t know of one. Without words to structure their thinking, players have a hard time thinking about this issue and moving past it. One effective way to approach this problem is to realize that every race can do things like this. The problem is stating what this is, which would allow players to quickly come up with examples from their own race to make themselves feel better.

Here’s another example: too often, professional players do a poor job explaining why the best Korean professionals seem to consistently be head and shoulders above the best foreign professionals. Not once in TRUE’s near unbeaten sweep of Dreamhack Montreal did someone concisely explain why the GSL Code S player was so dominant.

This is troublesome for several reasons. It makes it hard for professionals to repeat their successes and identify inefficient aspects of their own training process. It makes it difficult for aspiring professionals to climb up the ranks. And finally, it frustrates viewers by making it difficult to appreciate just how good professional players are.

I think part of the problem is a lack of vocabulary. To be clear: the difference is almost certainly in the process by which players train, and it is certainly physically possible to string together words into sentences to describe this. But as I noted above, the right words can communicate a lot of very important ideas very quickly.

As an example of where we’ve seen this before, the concept of a “team house” communicates the idea that a dedicated learning environment surrounded by people who work together toward the same goal is more effective than training in isolation. “Let’s live and train together” sounds strange; the “team house” concept expresses much more than that, yet with much less explanation required. This enables foreign players to take an effective practice and import it into their own countries.

Using the right vocabulary – or creating new vocabulary – and answering this problem succinctly and accurately would enable the sport to become more professionalized; it would force the articulation of testable, repeatable hypotheses that can be iterated on to produce better professional players in the future.

Professional athletes base their training regimens on science. They can often draw a dotted line from competition results back to specific training practices. Few would claim that the difference between an Olympic athlete and the guy who swims everyday is “swimming more” or “having a coach”. Yet when it comes to the difference between Koreans and foreigners, these are the types of vague explanations that are typically offered. In order for foreigners to catch up to the Koreans, they need to first articulate what the difference is. Specifically, they need to articulate it in a precise, testable, and repeatable way.

(Here’s one example of a former professional player, Artosis, attempting to explain it. Note the terminology “mapped out” – I think the map metaphor is a key piece of vocabulary needed to explain the difference.)

Next time I’ll be writing about the development of formalized processes. Stay tuned!

One thought on “Nomenclature and Processes in Real Time Strategy Games: Part I

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s