Author: Studio Ganzheit

Since there’s no new development screen to share, I’m just posting a picture of my computer instead. Looking at the photo now, my keyboard seems a bit dirty.
I spent the entire week revising the build. While the graphics are still missing, I’d like to hear some feedback—whether the game is terrible or not—and work on improvements. However, it’s not easy to find people to ask for testing. Maybe it’s because Metroidvania is such a niche genre, or maybe it’s because my game just isn’t appealing enough. Even the harshest feedback would be welcome; I just want to hear something so I can figure out what to fix. But getting even one meaningful comment feels incredibly difficult.
I’m moving forward, doing what I can, but I don’t want to end up in a situation where it’s too late to make changes. At the same time, I can’t just sit around waiting. Once I start working on the environment, fixing things will take much more effort, so I’d like to get things organized to some extent before moving on—but maybe that’s just my greed. I tell myself not to stress over things beyond my control, but I can’t quite let go of my regrets.
For now, I’m experimenting on my own—adding and removing things here and there—but I’ve noticed that most of the things I add end up being better off removed. I believe things will naturally flow in the right direction if I keep doing what I can, but I still wish I knew what I’m capable of doing right now.
I’ve thought about gathering testers through a community, but based on my previous experience, it was exhausting and not very helpful. Looking back, I think it was probably because the game wasn’t good enough. And now, with this temporary build still lacking backgrounds, it feels even more awkward to ask for testing. I can somehow manage making the game, no matter how grueling or demanding it is, but everything else feels incredibly difficult for me.
Talking with ChatGPT made me think that asking for help on Reddit might not be a bad idea. The people in the Metroidvania subreddit seemed even more hardcore than I am.

My perspective has changed. I’ve started to think that expecting feedback from enthusiasts to solve my concerns might have been a fleeting illusion caused by yesterday’s emotional swings. Hitting the bullseye with a handful of precise feedback is a fantasy; the reality is more about shaping and identifying the boundaries through numerous bits of feedback. For now, I’ve decided to focus on improving things during the current testing phase and prioritize quality over quantity when creating the background.

I’ve been working non-stop for the past few days. While I couldn’t fully address some parts that still bother me, I’m planning to wrap things up at this point and run another test with friends. This time, I might ask a friend’s son or some other friends to try it out as well. Looking back, it feels like the game has changed a lot compared to before, but I’m not sure how it will be received. Once I start working on the environment, I should finally be able to share some screenshots.

Using Unreal Navigation always makes me a bit uneasy. It works fine most of the time, but then one day, navigation suddenly stops working. The only solution I’ve found is to delete all the nav mesh boundary volumes and recreate them. The frustrating part is, just when I forget this trick, it stops working again and catches me off guard.

Once again, as I’m trying to wrap things up, I’ve encountered a bug I’ve never experienced before. A highly critical bug that’s never occurred from the prologue until now, with no clues as to its cause, has started showing up sporadically in the shipping build. I spent time yesterday reviewing what I worked on, but I couldn’t pinpoint the issue. Out of desperation, I switched the shipping build packing to a full build and rebuilt the code. For now, it doesn’t seem to occur, but it still feels shaky and uncertain. Truly, there are no exceptions.

Somehow, I managed to finish the playtest build and handed it over to a friend. Wow, as always, it was incredibly, incredibly tough. I ran into a few frustrating bugs along the way, but I didn’t panic as much as I used to. Now it’s time to start decorating. Progress is going well, even if it’s a bit behind schedule, but the country is in chaos.

The dev log is a few days late. While testing, I made significant adjustments to the level design and added a scanning feature. Whether it’s useful or not remains to be seen, but the scan works completely differently compared to the prologue. I also changed how doors operate, making the environments much larger while trying to speed up the overall game tempo.

After a week of crunching, it’s finally in a state where exploring doesn’t cause major issues. Next, I’ll be tweaking quests and enemy placements before running another round of tests with friends. It looks like I’ll need to dedicate all of November to wrap this up. Internally, a lot has been reworked, but visually, there’s still not much to show, so I haven’t been able to post dev logs on social media.

Solo development feels like an intensive crash course in everything I hadn’t learned before—whether it’s technical skills or life lessons. It’s incredibly exhausting, but I’m realizing a lot without crossing the line into self-blame.

Temporary things I made inevitably need to be properly redone. Well, that’s why they’re called temporary.

I’ve started testing, but after beginning a major rework of the level design, the effectiveness of the test has been cut in half. I think it’s because of my regrets over the prologue’s level design, which I never had a chance to fix—I couldn’t just let it go this time. I’m not sure if players will actually like the changes, but I had to make them. I initially planned to finish the build by the end of December, but it looks like I’ll miss that deadline. I should have started the background work in November, but I haven’t even begun yet. Nothing is going as planned. I’m trying to focus only on what I can do and just work on making a solid build, but it’s not easy. The levels I designed earlier this year had too much vertical overlap compared to later sections, making it hard to distinguish them on the world map. So I started spreading them out, which naturally led to the need for more connecting spaces. I couldn’t ignore how these areas would play, so I had to address them too. As a result, there’s nothing to share on social media this week, so I’m just posting a development update on the blog.

The first feedback from the first test feels like a bit of a turning point. It’s not that everything is decided at this point, but usually, when the initial feedback is something like “it’s okay” or “not bad,” it means that the game needs significant effort to bring it up to a good level. For the prologue, the feedback was something like “frustrating but playable,” and despite putting in a lot of effort, the final outcome wasn’t great. This time, however, I received “it was incredibly fun,” so I’m hoping it turns out better than last time. Hopefully. It was truly uplifting feedback during a tough time.

There’s no way I can stop making the game I want to make, and there’s no other opportunity to do so, and time is running out. Based on my experience so far, working alone seems to suit me best. Still, there are times when it’s so tough that I almost want to quit, but ironically, what keeps me going isn’t people—it’s the game itself. Fame and money are important, and it’d be nice to have them, but right now, having the opportunity to find more people who enjoy the game I loved creating feels more valuable to me. I need to make it properly first and then try to find them.

Gameplay refinement is almost done. I still see things I’d like to fix, but if I don’t wrap it up here, it’ll never end. It seems better to make adjustments as I proceed with testing. This week, there’s no video to upload to the blog, so I’m just posting an image.

The background work is slowly getting underway as well. I was a bit concerned when Megascans announced they were going paid, but lately, I’ve been using AI to create materials and textures, so it’s been a long time since I actually used Megascans. I also needed some HDRIs, but it turns out there are plenty of free ones available. The development environment has truly improved.

After receiving feedback from the last video to add camera shake, I realized that the camera shake that used to work had stopped functioning. I had used Unreal’s feature, and I confirmed it was working in a build from two months ago, but somehow, it’s no longer working. I couldn’t find similar cases or clues after searching everywhere, so I ended up just creating a simple version with only the necessary functionality. It feels like a foolish situation, but it seemed like the fastest route. I just hope it doesn’t cause any unexpected bugs.

Day by day, it feels like the progress is small, hard, and slow, but when I look at the checklist of everything accomplished so far, I sometimes feel like I’ve done a tremendous amount of work. If this effort gets recognized by the world, it would be truly rewarding. Que será, será.

Refining the gameplay this week. While there are still things that need improvement, I really plan to wrap up refining by the end of this week and start testing. For the background work, I’m aiming for a slight cyberpunk garnish over the actual COEX for the exhibition hall and outdoor areas, while for deeper places like non-preservation zones, I want something more industrial, similar to the manga BLAME! with a hard-edged aesthetic. Manually placing all entities is challenging, so I’ll be using a program to place most of them. Just feeding it prompts, voice prompts, photos, illustrations, or even a suitable kitbash set, and having it fill in background spaces accordingly would be ideal.

In the past, when I got stuck in programming, I used to turn to Unreal forums and API documentation, but the forums often lacked answers, or the answers provided were too risky or simply incorrect, mixed with the right ones, leaving things feeling uncertain. The API documentation also felt cumbersome after moving to Unreal 5, as searching for 4.27 documentation became awkward. The Unreal Assistant in ChatGPT seems quite decent in comparison. There are still hallucinations, of course, but personally, it feels better than the forums. Compared to when I first started solo development, though it has only been three years now, the development environment has changed immensely thanks to AI. Character voices are also done with AI, making modifications and additions quick and easy (though the quality falls short of actual recordings, but given the almost non-existent cost, it’s a blessing for indie developers). The advertising images and videos used in the game backgrounds are also created using AI. Back in the prologue, I only had placeholders for signs and couldn’t muster the resources to create the actual images. Some sound effects are also made with AI, and simple FX textures can be generated reasonably well. I recently watched a video of GTA 4 reimagined with AI, and it feels like I’m manually doing that myself. Ultimately, I hope it becomes possible to work quickly with low-poly to establish the big picture, and then have AI filters in the engine to upgrade the quality and adjust the tone. That way, I could make one game every one or two months, like writing or serializing a novel or manga. I have so many stories I want to tell, and I really hope it becomes possible.

Refining gameplay since last week, but progress is slower than expected. The list of tasks to fix grows a bit faster than it can be reduced. Behind the psychological shield of “posting for the sake of documentation” or “to find glitches,” a weekly dev log gets uploaded that hardly anyone sees. Yet, it’s still disheartening when posting seems to actually reduce the already few followers on social media—even behind that shield, it’s hard not to feel affected. After enduring a series of devastating failures and doubts—to the point of questioning whether this work is pursued simply because it’s the only thing I know, or if it’s a matter of being dragged along in self-deception—I thought I had come to understand that acceptance by the world doesn’t correlate with effort, but it’s still difficult. The saying “effort never betrays” should be changed to “effort promises nothing.” Efforts continue to avoid getting too caught up, whether it’s in social media or in the thought of bringing the world inside my head into reality, or in effort itself.

A bit of a tangent, but about one of my “shields”—the idea of “posting to find glitches”: while creating, I often can’t notice any issues, but after capturing gameplay videos for the dev log and uploading them, glitches suddenly become apparent a day later. I think it’s because during testing, my focus is too narrow, fixed only on what I intend to check. Even right after uploading, I still see only that, but rewatching a day later without any specific expectations makes the glitches stand out. Since I rarely have opportunities to show the game to others or get advice during development, this kind of self-review is incredibly helpful.

Anyway, the show must go on. I aim to finish the list of tasks by next week and start some test runs with a friend. One friend, who even helps with free character modeling work, is assisting with testing, recording full gameplay videos to highlight any issues along the way. I’m truly grateful for this friend’s support, which makes a huge difference on this journey. I’m a bit anxious, wondering how demanding this round of testing will be, as the results are always so different from what I imagine. Sometimes, I also find myself hoping that someday, I too might be able to help someone in my field.

I am still preparing the pitch build. For now, I have completed work on events, tutorials, enemy placements, and other added systems up to the intended content. Tasks requiring polishing and fixing have once again piled up into a mountain. After fixing them one by one, I plan to first conduct tests with friends, even without the backgrounds implemented. While testing, I intend to work on the backgrounds in parallel.

It’s tough to act as if I’m not pressed for time, even though I am somewhat under pressure. However, compared to before, things, including me, have improved a lot. I used to always feel like the protagonist in the movie Gravity, trying to focus on fixing something in a space station being destroyed by debris. By changing the mindset to “let’s not think about what can’t be controlled and just do what can be done for now,” I’ve been able to work much more steadily, even in the same situation. Since starting solo development, this experience has really made the phrase “it takes decades for something to go from the head to the heart” resonate. I don’t know how much time I have left, but I hope to leverage this experience to create the game I wanted.

I modified the game events that were handled only with voice in the prologue. By reducing their number and instead showing the actual events, I tried to improve the gameplay experience. Fortunately, the development environment and my experience have changed tremendously in just three years since making the prologue. Buying or creating assets has become much easier, making this more feasible. In particular, the help from AI has been significant. Things that I wouldn’t have even dared to make before, or would have taken weeks or months, can now be done in seconds or minutes, and things that were impossible to modify can now be freely altered. I hope that AI will develop further so I can create a game every one or two months. Sometimes, I still feel envious when I recall old game development videos from the 1980s, where a solo developer would set up two Apple II computers and create a game every few months on his own. What I’m envious of is the part about having two Apple II computers and making one game every few months alone. I often think it would be great if the engine editor and AI were fully integrated, allowing many tasks to be done automatically.

In a previous development log, I wrote about my belief that “user feedback is always right,” but I forgot to mention some exceptions: people who follow me around leaving negative feedback without explaining their reasons. These individuals don’t respond when I ask for details about their comments. They were just like bad weather. Just as you use an umbrella when it rains, stay indoors, or move away when natural disasters are severe, I thought it would be better for my mental health to block and ignore them whenever I encounter them. In the past, I would have cursed at these people and become extremely angry, but I’ve changed so much now that even I am surprised.

Another group consists of people who only insist on their own changes. While not many, these individuals share a common trait: “If you apply my changes, I’ll buy the game.” or “If you don’t apply my changes, I’ll refund the game.” I asked them more about what was uncomfortable, and their conclusion was that they wanted to change the direction to make the game they desired. While I appreciate their passion and interest, I believe that each person should achieve their own dreams through their own efforts, and in such cases, I choose to pass.

While writing, I suddenly remembered something else: when I released the build, I received literally hundreds of emails asking for keys in exchange for reviews. At first, not knowing any better, I was grateful for the interest and gave out keys, but in hindsight, I think it’s better to just consider emails asking for Steam keys as scams. Emails would come claiming to run channels with hundreds of thousands or millions of subscribers. When you look at the email addresses, they trick you by replacing lowercase ‘L’ with ‘I’ or changing ‘m’ to ‘rn.’ Some emails claimed they needed a key because their nephew was sick, but the emails were identical except for the address, with not a single letter changed. Others said they ran a game news page, but when I visited, it was a news page that hadn’t been updated in ten years. There was a wide variety.

I’m working on various tasks, big and small, for scenario events and the tutorial. The video above shows the scene where you enter Active Imagination to proceed with the tutorial. In the prologue, you decided at a specific point whether or not to recall a past memory, and that decision led you to play the tutorial. In Chapter 1, you can practice skills anytime, anywhere, by pressing the Active Imagination buttons for each skill in the conversation window with Pi. The screen transition effect might not seem like a big deal, but it took a lot of time. I created videos like the ones below using Kling, then converted them into text to use. I’m considering remaking them later because the resolution and quality of the text-converted videos are low.

I spent a lot of time thinking about how to transition between reality and Active Imagination. I had no ideas and was struggling, so I started with the falling text ASCII art from the movie The Matrix, but it felt outdated and overused. Then I arranged metal pillars like basalt columns and tried various things with them (not sure why I started that), and eventually, it resembled a huge structure made of metallic basalt columns that I saw in a dream recently. I didn’t realize it until my wife pointed it out. When I put the metal pillar video into the game, it didn’t fit at all. After agonizing over it again, I ended up using the ASCII art transition I initially tried and discarded. I ended up spending over a day just on the screen transition. For now, I’ve included it in the game since it’s passable, but I might change it to something else later. Every day, I oscillate between feeling that the game is playable and feeling that it’s utterly worthless. While it doesn’t affect my pace of creation, it’s quite a tough experience. Considering the current state, the utterly worthless side is probably more accurate, but I hope it gets better little by little. Doing work that almost no one sees is really difficult.

Originally, I wanted to quickly release a series similar to DLCs under the name Axiom of Maria, with titles like Prologue, Chapter 1, Chapter 2, and so on. That was when I didn’t know anything and just dived in recklessly. However, after making the Prologue, I realized that releasing new chapters as DLC-like additions wouldn’t be feasible. In fact, Chapter 1 is almost entirely refactored or newly created, except for some player actions from the Prologue. I decided that I needed to separate them, but Steam wouldn’t let me change the page name. I was so preoccupied with development that I left it as is, which seemed to confuse a lot of people. Interestingly, after receiving feedback that the word “Prologue” wasn’t in the name, I tried changing the page name again, and it worked. Go figure.

I’m always desperately thirsty for feedback, but sometimes certain feedback drags me into my personal hell. Even so, there have been many times when rude and aggressive feedback contained golden nuggets, so I can’t afford to disregard any feedback. To me, this was a harsh and brutal truth, just like how natural disasters harm people without discrimination. On the other hand, positive or vague interest was emotionally very appreciated and even sometimes made me cry, but it didn’t seem to help much in actually making the game. It’s quite ironic.

There have been many modifications and additions to features. These include various background elements for Metroidvania-style progression, status effects like the player catching fire or being affected by electric fields, multiple enemy reactions to attacks, various processes related to the world map, a new suit program UI, a chat window similar to ChatGPT, and so on. Because of that, many bugs appeared in what I had previously made, so I had to fix a lot. As I was resolving the broken parts and working on the story and events, it feels like I’ve remade almost half of the game.
It was tough because the results were too meager compared to the effort. However, seeing myself quietly rejoice as the game’s new form slowly took shape made me realize I could never give up on this. I feel like I’ve spent about two-thirds of my life, and I’m living with the thought that if I just muddle through the remaining one-third, I’ll somehow keep going.
I had focused on the theme, visual presentation, and vibe when making the prologue, but about a month before its release, I realized that I hadn’t been able to present that theme to people as attractively as I thought. I still believe that the theme of Axiom of Maria’s story is interesting, new, and offers a lot to think about, but I think it was my mistake to just lay it all out there right away.
When making the prologue, I think I was so mesmerized by the act of creating the game itself since it had been so long since I last released one. This time, I’m focusing much more on the story itself and the gameplay rather than the theme, but since I’m making it alone until the first test, it’s really hard to check if I’m doing it right. The saying “One is the same as none” comes to mind.