Breaking Down Age of Empires IV's Launch Problems
Assessing what went wrong, and the lessons therein
Age of Empires IV released in late 2021, and I was not a fan. I felt that the game shipped with a lackluster campaign, an unfinished multiplayer, and a host of technical problems.
I think it’s worth analyzing the launch window issues more rigorously, to glean some understanding about the RTS market. Age 4 is one of the biggest real-time strategy games to come out in recent memory. And, helpfully, it’s one that I actually have context on, having played the game extensively through the launch window. I think there’s a lot to learn from diving in and cataloging real user feedback. At the very least, it keeps me grounded in real world data.
(For what it’s worth, I revisited Age 4 a year and a half later, and I came away impressed by how far it had come.)
Limitations
Before I dive in, I should address some limitations. First, I think I can only compare feedback against other feedback; it’s hard to draw absolute conclusions about what people value. For instance, I didn’t see any complaints about the save and load capability in the campaigns - but that’s probably just because the feature exists and works as expected. If the developers had excluded it, I’m sure it would have been a top issue. (Even Bannermen, with its numerous other issues, got burned for this in its Steam reviews.)
The other major issue is data quality. I don’t have visibility into private channels, like phone tech support or bugs filed directly to the developers. The best I can say is that public complaints should generally, with some caveats, correlate with private complaints.
There are some limits to the data I do have, too. Neither Reddit nor Steam officially support ranked, time-based queries. The official Age of Empires forums do, though, so I catalogued the complaints there and cross-checked them against other boards (via a time windowed Google search) to catch anything that fell through the cracks. The official forums get good traffic based on view and like counts, so they should offer a decent window into the game.
Data
I catalogued the top 100 posts on the official forms (by views) into buckets, for the first two weeks after launch. Here’s what I found:
I broke negative feedback down into some high-level categories to help understand it a bit better, and I leveraged some of the context I have from playing the game during launch to try to better capture the spirit of the feedback. For instance, I grouped a bunch of issues into the “parity” category, which I’ll discuss down below.
I cross-referenced this against searches through Steam and Reddit, skimming through to see if I missed anything. I’ll add color where relevant down below.
Analysis
Technical Issues
The top category of complaints relates to technical issues, with about half relating to reliability problems like crashes or significant performance issues. This tracks with my experience, too: the game launched in a buggy state.
I know I’m not splitting the atom by pointing out that launching a solid technical product is table stakes for putting out a good game. But, well: I think it’s table stakes for putting out a good game.
One observation I had during the lead up to Age 4 was that Relic didn’t make much of an attempt to get this game into people’s hands ahead of launch. I participated in the Age Insider program and I was invited to both Closed Betas, participating in the former; I even gave feedback on the official forums. Anecdotally, it didn’t feel like a lot of players participated, and each Beta was really short (which is why I missed the second one).
I feel that Relic missed a large opportunity here. There was a lot of interest in the game leading up to its release. Driving broader Beta participation would have allowed the team to collect data on how the game ran on a wider variety of device configurations, and perhaps would have pre-empted some of the technical teething issues that emerged during the launch window. I’m sure there’s other ways of going about it, but given that the company did all the leg work to conduct two Closed Betas, I feel that it wouldn’t have been too much of a stretch to scale them out more than they did.
One category of technical issues that flew a little bit under my radar were the ones relating to masteries and achievements. I mean, it’s intuitive if you think about it: there’s a sub-group of people who really care a lot about racking up in-game achievements. Hell, even I made sure to get every single StarCraft II in-mission campaign achievement - and one of the reasons I emphasize “in-mission” is because I wasn’t a fan of the Lost Viking achievement, and I still remember it years later.
Purely anecdotally, from my own limited experience hunting achievements, they sometimes feel like icing on the cake - a bonus feature, or perhaps even an afterthought, at least from the perspective of some developers. But some players perceive them to be part and parcel of the core experience. And I think the cost of making those players happy is, relatively speaking, low-effort and high-reward - make the achievements actually work, ensure they make sense, and try to make them fun to collect.
Negative Feedback #1 - Single Player AI
The second largest bucket was a wide variety of negative user feedback. I tried to organize this based on my own understanding of the game and the state it launched in.
Let’s start with the single player AI. The post complaining about it is the fourth-most viewed topic during the launch window. In addition, the same basic complaint appears in many of the long lists of feature requests and general complaints that people put up on various forums. Many people took issue with the fact that the single-player AI wasn’t very good.
I think that’s important. I think, with most RTS games, the large majority of players can be found in the campaigns. (For StarCraft II, this was confirmed directly in an interview with a designer). The next biggest bucket are folks who play skirmishes and custom games (many of these people also play the campaigns). The competitive side, which tends to draw a lot of attention online, is usually significantly smaller than either of these groups. Thus, shipping a poor single-player AI is going to hurt the longevity of the game with a large and important group of players, and I think this contributed to Age 4’s steep dropoff in player counts after release.
My take here is that I think Relic bit off more than they could chew with Age of Empires 4’s engine. I think that way because these sorts of issues occur across the experience - from the single player AI, to the scripting hiccups in the campaigns, to the difficulty in clicking on things at launch. These problems felt too large to be simple lack of polish - I just don’t think the game engine was ready for prime time when Age 4 shipped.
For example, in one play test I ran, the AI never advanced to the Feudal Age. It put sixteen villagers on gold and, when faced with a men-at-arms rush, fought them hand-to-hand under the town center instead of garrisoning.
As a point of comparison, the Genie engine that powers Age of Empires II makes it easy to write single-player AIs - Ensemble even published a book on it. I read that book when I was younger, and I was able to write a decent single-player AI that followed basic build orders, built an army, all that kind of foundational stuff. This was at a time when I didn’t even understand the concept of boolean logic (which the book helpfully explains). The engine was just really solid and made it simple.
I get that it’s easy to be an armchair critic regarding low-level technical details. But, well, this sort of trade-off does come up in software development from time to time, and it’s not limited to games. I myself have worked on projects where I found myself patiently influencing stakeholders that getting the foundational technology right would enable us to move faster and deliver more value down the road. Making that case takes patience, work, and a lot of data - and you’ll always have people pushing back. I’m honestly sympathetic to the situation. But it does seem like a mistake that’s easy to make in software development, one that anyone is capable of making. The best I can do is observe the end product and try to learn from it.
Negative Feedback #2 - Parity
Parity issues come up a lot in the data, and I think that’s worth exploring a bit. I made a video on this back in 2017; the problem with franchising is that it’s something of a double edged sword. People are more likely to purchase a game because it carries a familiar name, but they also come in with higher expectations. They expect the new game to work similarly to the games that came before, sometimes in ways that may seem trivial or even unrealistic.
One complaint I saw repeatedly - sometimes even framed as a technical issue - was that depositing relics works differently than in Age of Empires II. That game allows you to shift-click a monk to a relic and back to a monastery; Age 4 only allowed queuing the deposit once the monk already had the relic, thus requiring an extra action. Definitely annoying, but not a big deal, especially in light of the fact that relics actually see broadened gameplay scope in Age 4. You could argue the Age 4 approach was even an intentional decision, to encourage players to see relics as a tool rather than just a collectible. But many people complained about it, because they had certain expectations from the previous games.
From my perspective, a number of issues fall into this bucket, like the lack of player color selection, or the inability to simultaneously pause and look at the map (you needed to bring up the menu to pause in single player). The replay system was significantly improved over previous games, but the delay in saving them due to cloud sync resulted in multiple bug reports complaining the feature didn’t work as expected. And so on and so on.
People have expectations going into any game, of course. But franchise games are held to a uniquely high bar - they need to move the gameplay forward in a meaningful way, while still maintaining everything people love about the franchise to begin with. That’s a big ask! I wonder if broader user testing could have helped catch disparities that fell through the cracks. Maybe consciously hand-holding players around experiences that have changed since previous games could work, too, like a tutorial focused on the relic’s new conversion ability. I’m curious now to go back and replay other franchise games to see how they approached changes to existing mechanics.
Negative Feedback #3 - User Interface
I was surprised to see so many complaints focus on the UI. I’ll pick this post as a representative example - it cites a wide variety of issues, but many of them boil down to fundamental issues with the interface, or gameplay features downstream from the interface like waypointing and multi-select.
Stuff like this reminds me that I’m not very representative of the average player. Maybe I’ve just played too many older RTS games, but I have low expectations of the UI. I’m used to it not being very good and, well, that’s that. I get the various concerns about zoom and pathing and the mini-map and shift-click and all that, but they just don’t bother me very much, and I think that makes me an outlier. Reading these posts made me thankful that I put together this article, because it made me realize a blind spot I have in how I think about these games.
Notably, I didn’t see many complaints regarding my biggest issue with the UI - the inaccuracy in click targeting. To me, that was a significant miss. But it turns out not to have bothered many people. I’d hazard a guess that other players play the game more slowly and with less emphasis on precise clicking, on average, than I do. It reminds me of a similar revelation that I had years ago, when I realized many players don’t use more than a handful of hotkeys to play real time strategy games.
I’d love to learn more on this subject. As a follow-up to this article, I plan to connect with UI designers to pick their brain, as well as to catch as many knowledge shares as I can on the topic. And if you’re working in this area, please feel free to reach out. I’d love to hear from you!
Balance
The third largest bucket of complaints were about competitive multiplayer balance. These seemed more popular on Reddit, but I don’t have rigorous data on that.
Age of Empires IV was not well-balanced at launch. But then, I don’t think any RTS game is, or can be, during the launch window. It’s hard to know what is broken without a good chunk of initial playtime. So I’m not sure how seriously I can take this kind of feedback. And, to be honest, I feel that Relic actually did a reasonable job in this area, releasing several balance patches within the first few months. Microsoft also sponsored at least one large tournament for the game early in its life cycle.
I’m actually surprised that they put in as much effort as they did; I feel that the developers should have placed higher priority on more pressing issues. I get the argument that it’s easier to cut a balance patch than it is to fix a bug (especially if it’s foundational to the engine). I also recognize that games companies have lots of people playing different roles. But I feel the amount of effort put towards esports and competitive multiplayer demonstrates a partial lack of focus on the part of Relic.
A good example is the spectator mode that launched with the game. It works great. As a competitive player and a fan of esports, I appreciated it. But we live in an age of ubiquitous live streaming and relatively cheap compute resources that make distributing replays simple and easy. Why spend time building a feature like this in a game where more basic functionality - like clicking on stuff - wasn’t working?
I think a lot of people who go off and develop a real-time strategy game will have fewer resources at their disposal than Relic did. A lesson I draw from this release is that a game’s prioritization scheme has to be clear in order to really nail the core experience. I’m sure it’s painful to release an RTS and have ranked play be unbalanced. But it’s also painful to release a game without a solid AI, or a buggy campaign, or a host of other issues. And sometimes you’ve got to live with the resources you have and the timeline those resources allow you to deliver on, and you’ve got to prioritize. And I feel that’s something that could have been done better in this case.
Final Thoughts
I learned a lot from reading through this feedback. I’m always happy when I catch a blind spot, see a new perspective, or realize I’ve been thinking about something the wrong way. I got a little bit of all of that from this exercise! I’ll probably do it again in the future.
Sometimes I worry that the only relevant audience to articles like this one is myself. But I like to think that this kind of stuff is fun to read about for anyone who’s a fan of the genre. I hope you found this to be an interesting read!
Until next time,
brownbear
If you’d like, you can follow me on Twitter and Facebook and check out my YouTube and Twitch channels.