How I Trained Age of Empires II

Hey folks, I’m brownbear. Over the past year, I spent some time grinding the Age of Empires II ladder. I wanted to write a little bit about this experience and how I approached my training.

Early Days

I started out in May 2020 as a solid beginner player struggling against the extreme AI. My initial goal was just to get competent enough at the game to put together a solid retrospective video, as a follow-up to my retrospective on Age of Empires III. At the time I had a skeleton program in my head of how I wanted to approach things, a sort of untested “theory of learning”. It went something like this:

  1. Play with structure (use build orders)
  2. Play with breadth and depth (go deep enough to learn a thing properly, but switch it up often to avoid becoming a one-trick)
  3. Play to improve, not to win (do what makes sense even if you suck at it)
  4. Play consistently (a little bit everyday beats a lot in one day)
  5. Play with segments (a couple hours of focused training beats a long, mindless grind)
  6. Review often (you don’t know what went wrong until you watch the replay)
  7. Watch and learn from the best (if you’re doing something different from the pros, either you figured out something they didn’t – unlikely – or you’re wrong)
  8. Set an ambitious goal (the harder the goal, the more it forces you to train correctly)

I developed this program from the learnings of several grinds I’ve done in my adult life – learning a new language, training StarCraft II, and distance running. This time, though, was the first time I’d put it all together into an actual list of guiding principles.

Early Successes

I had great early results with the program. I grew into a top 2% player after about six weeks; I wrote a bit about this last year. I want to expand a little bit more on some aspects of the program that I think will be misunderstood, to better explain why I set them up that way.

I think one common misconception relates to the suggestion to play with structure; specifically, that it makes you unable to adapt. Structure is just the notion of building your gameplay on top of a foundation that is known to be a solid or near-optimal way of playing the game. In real-time strategy, the foundational structure is build orders: common openers and transitioners into mid- and late-game tech that exploit particular game, civ or map characteristics.

I’m not claiming here to say that I know “the best way” of playing Age of Empires II. Heck, “the best way” changed during the course of my grind. 21-pop scouts and 22-pop men-at-arms gave way to 20-pop and 21-pop, respectively; FC builds more commonly incorporated a second lumber camp; pre-mill Drush became more common than a standard Drush; etc. The community’s understanding of the game and the particular quirks and nuances of the current meta and balance patch all influence “the best way” of playing the game.

Structure is just an attempt at centering your gameplay around the best practices currently known in the current meta; it’s the most efficient way of constructing a broader decision tree of how you proceed through a game. You have your middle, optimal path cutting down the center. Branching off of it you have your initial reactions to things that you scout, and each of those reactions has its own optimal path, its own reactions, etc. It looks something like this (with a lot more branches):

The decision tree is an abstract concept; it exists whether or not you’re aware of it. If you switch villagers from wood to gold in response to something you scout, like it or not, you’re traversing down a decision tree – you originally planned to go down a branch where you had more wood, but now you’re going down a branch where you have more gold, relative to your original strategy. Build orders don’t create the decision tree – it’s always there! – they just make it explicit in your head as you play the game.

This is what I think people don’t get when they claim that build orders make you inflexible or unable to adapt. Regardless of whether or not you’re following a build order, you traverse a decision tree in every game you play. With a build order, you do that from the starting point of a fairly optimal set of decisions; without a build order, you’re reverse engineering the tree from scratch. At the end of your training, you arrive at the same place. But one option takes you way longer!

One of the key benefits of structure is the ability to repeatedly compare a set of decisions across many games and see where they work and where they don’t. It’s usually inadvisable, for example, to directly engage a 21-pop men-at-arms opener in the first couple minutes of the Feudal Age when you open scouts. The reason for this is that the scouts opener offers a stronger economic foundation, whereas men-at-arms offer a better early power spike. It’s in the men-at-arms player’s interest to engage, and it’s in the scouts player’s interest to buy time and cut off reinforcements.

Build orders enable you to figure all this out fairly quickly – you’ll notice that one build always researches horse collar and double-bit axe at the beginning of the Feudal Age, while the other requires you to cut corners just to research the latter. You’ll notice that one build takes a little while to get going militarily, while the other positions three powerful military units in the opponent’s base shortly after reaching Feudal. You’ll put these and other things together and realize, hey, these builds work this way for that reason, which is why this one transitions that way and that one transitions another way, it’s why X reaction is common when you scout Y, etc. You gain a concrete understanding of how the game works, which you can then apply to all of your future games.

Streaming and Plateauing

Around the time I hit top 2%, I decided to start streaming my ladder games. I didn’t put too much thought into this, thinking of it as a sort of “free lunch”. If I’m going to grind ladder, why not stream the process? It’s free content!

Streaming taught me a couple things about training. First, I had setup a rule to play with segments, under the idea that a couple hours of focused training was better than a long, mindless grind. Streaming for a couple hours, while aligned with the letter of the rule, didn’t really align with the spirit of the rule. Reading and responding to chat, feeling pressured to put together content that was “entertaining”, and the day-to-day tech challenges of content creation all incurred a non-trivial drag on my learning. After awhile I wasn’t sure if I was mainly-streaming or mainly-training, and that was a problem.

Never half-ass two things. Whole-ass one thing.

Ron Swanson

I ended up plateauing at top 2%, and I gave up completely in mid-August. I had the convenient excuse of moving from Japan to the United States, but my continued commitment to distance running made it clear that I was capable of sticking with things during this transition. Looking back, I think the choice to stream my gameplay ended up being the deciding factor in why I called it quits for most of the Autumn.

It wasn’t just the mechanics of streaming that I struggled with. For a while, I wasn’t sure how to verbalize my issue. But when I came across this Reddit thread, it crystallized for me a lot of the problems I have with communication on the Internet.

The advice in this thread is not good. Hera’s advice is correct, and if you scroll down far enough, you’ll eventually find someone who understands:

His point was it’s a bad habit to get into. The idea of not waypoint scouting forces you to continually tab back to the scout increasing your multitasking ability.

But imagine reading this thread as a new player. How do you know who to believe? The signal-to-noise ratio is terrible. The highly upvoted guy that has spent months learning how to Dark Age scout, and still uses waypoints? Sweet Jesus. Anyone, regardless of experience, can learn the basics of Dark Age scouting in a short amount of time if they approach it the correct way. It’s leaning on crutches like waypoints that make it difficult and time-consuming.

Well-meaning people would sometimes trickle into my stream and shout out terrible advice. Here’s something that would happen often: I’d lose a game, and, while watching the replay, someone in the chat would mention that the civ match-up was imbalanced and bad for me. I’d reply that balance doesn’t matter too much at my level, and that correct decision-making, foundational skills and mechanics are what decide games. They’d then go on a tirade about how I’m top 2%, what am I talking about, stop it with the false modesty, of course balance matters, etc.

I know that this is bad advice. But hearing it over and over from different people was really stressful. It hurt my confidence, which bled into my gameplay, and it made me doubt how I was approaching my training. And if the person stepped over the line and I wanted to time them out, that made me feel worse. This person is just trying to help, I’d think. I used to think that way, too. I should help them understand it’s not a healthy way of thinking.

It’s the stress, really, that put me off from streaming, which then put me off from training the game at all. I thought to myself, if I’m going to play, I have to stream, and I don’t want to stream, so I’m not going to play.

The Return

I decided to get back into Age of Empires II in early December. My motivation was basically that I didn’t like how things had ended – I felt like I had done what I often do and quit something in the middle instead of seeing it through to the end. I wanted to get to a point where I felt proud of what I’d achieved and how far I’d come; that instead of just hitting some toy achievement of top X% within N weeks, I’d actually feel like an advanced, solid, good player.

I stuck with this for a solid eight months. And it turned out really well! At the end, I consistently maintained two accounts in the top 1%, playing random civ. I learned a bunch of new openers and build orders, I played aggressive all-ins, I played pure defensive macro, I took games off of solid 2k+ players. I even won the top two groups of Hera’s RBW3 community tournament! I did a sort of “greatest hits” of Age of Empires II (to the extent that a non-professional can do that), all on the back of less than a thousand lifetime 1v1s.

This second grind reinforced for me the importance of systems over goals. If you recall, my original program suggested setting an ambitious goal. Ambitious goals are a forcing function for efficient, effective training programs. I still think that makes sense, and I still think it’s important to have a north star vision of where you’re going in your training. But I tripped over my ambitious goal a few times during my grind.

I wanted to reach 2k. For the longest time in Age of Empires – stretching back years and years – 2k has been the mark of an advanced player. It was the point where you could say that you had really gotten good at the game. For me, there was something about the change in that first digit – going from a 1 to a 2 – that felt really impactful. I had been a 2k player in Age of Mythology, I had been a 2k player in Age of Empires III, and I wanted to be a 2k player in Age of Empires II.

That goal proved to be… unhelpful. My original program emphasized playing to improve, not to win. It’s rule #3 – one of the more important ones. The problem with having a rating goal is that it’s at odds with playing to improve. Playing to improve means that you’ll lose games that you could have won, for example if you pigeonholed yourself into an all-in that you’re good at instead of playing the match-up in the academically correct way based on the game state and map gen.

I would sometimes have big swings in ratings that I didn’t understand. One day I’d take games off of higher rated people easily, the next I’d lose games to lower-rated folks with a good match-up and a great map gen. The problem was all-ins: I like doing them, I’m naturally inclined to play aggressively, and they’re easier to execute than pure defensive macro. And with a rating goal, it was all too easy to ignore my prior advice of “making myself uncomfortable” and just panic all-in when I wasn’t sure what to do.

And that’s really what it was – panicking. I would play nervously and without confidence, driven by a fear that I wasn’t getting closer to my rating goal. Knowing that this was not effective, I would force myself to play correctly, and this would work for awhile; but given enough time, I’d still drift back to old habits.

I’d estimate that I lost around one in five practice games to this mentality. Not lost as in I lost the game, but lost as in I didn’t learn something meaningful because I just panicked and all-in’d. In these moments, I was too focused on achieving a particular rating goal and not interested enough in what that achievement was actually supposed to mean. Ultimately, I ended up as a solid 1900+ player when playing random civ with pure macro styles. To this day I think I could have reached 2k if I had spent a couple weeks only knight/siege pushing; but also to this day, I think it’s having that thought in the back of my mind that prevented me from becoming a solid 2000+ player.

I’ve started to wonder whether there’s a sort of decay built-in to all training programs, which can only be resolved by periodically refreshing your discipline. I got this idea from distance running, a program I’ve maintained continuously for almost two years, and probably not coincidentally, a program I’ve refreshed several times since its inception. I went from pure 80/20 to Run Less, Run Faster to daily outdoor running to doing lots of events to running everyday in the gym to double running some days to join my co-workers. The principles of the program haven’t changed too much, but the execution and mechanics have changed several times. Maybe this is the missing puzzle piece I’ve been looking for.

The End

I’m not done playing Age of Empires II. I love the game and I love playing it competitively. I’ll be back on the ladder, and back on Twitch, soon enough. But I would call this the end of my “first grind” of the game. I started out as a beginner and I ended up as a solid advanced player. I’d like to go further than that, but somehow I feel it’s going to look very different – it’ll be a “refresh”, like what I did with running. No matter what, though, I’ll never forget this first grind – getting up early, not wanting to play, not wanting to stream, but pushing myself to play, pushing myself to stream, pushing myself to watch the replays, and ending up having a really good time.

I’ll see you back on that ladder.

Many thanks to the regulars in my stream – squeaker, MagicFish, Delta, and several others – who helped keep me going during my grind. And don’t worry! My next grind will be different, but the stream schedule will remain the same. Hope to see you folks soon. 🙂

Also, many thanks to Hera for his amazing Age of Empires II content, particularly the build orders. It made the grind more fun and much smoother. Check out his stream if you’re interested in the game.

Finally, thanks for reading! If you enjoyed this article, I’d love it if you followed me on Twitter and Facebook and checked out my YouTube and Twitch channels to receive all future updates. I don’t check social media or user comments, so if you need to reach me, send me a DM in Discord.