Thrall

Greetings Vampire Hunters!

Thrall is undergoing some changes. Turns out, I vastly underestimated the scope of this project. Allow me to illustrate.



The image above is essentially a very simple layout of the story progression for Thrall, and how the development is currently going. Up to this point, I've essentially completed three dungeons and a "boss rush" style of dungeon, which has taken me close to one full year to do. I should mention, these dungeons are the smallest in the game. Each of the bullet points on this image signify another dungeon, each one getting progressively longer.

Problem is, my method of making dungeons up to this point is extremely time consuming, taking months to complete even a relatively small dungeon. When I ran the numbers on how long it would actually take to make every dungeon in the game with my current method, it equated to around 1200 maps, each taking 2-3 hours to make. That's like, three to four years of working every single day, 2-3 hours a day on just maps. No cutscenes, no story, no bosses, no mechanics. Just maps. Four years, no breaks.

I'd like to release this game sometime in my life. Preferably in the next two years. So something had to change. I came up with four potential solutions.

One, I could team up with someone, or a group of people. Couple issues with this, however. First of all, I'm not selling this game, or at least I don't plan to at this point in time. So any help that I get would either need to be paid out of my own pocket with no expected return, or I'd need to find people willing to work for free. I believe in compensating people for their time and their work, so that's a no-go. In addition, teaming up with people will inevitably change some things about the game. It's difficult to teach people a consistent art style, so I'd likely have to make some compromises there. Not my favorite choice.

Two, I could reduce the scope of the game. Instead of around 25 dungeons, I could maybe reduce it to ten, and eliminate some of the branching paths. Maybe eliminate the branching narrative structure altogether. Again, not really the vision I had for the game, and not my favorite choice.

Three, I could change the artistic style of the game. Simplify the dungeon style, maybe go tile-based instead of parallax mapped. But, one of the reasons I embraced this style in the first place was because I couldn't stand how bland the dungeons were in the Pokemon Mystery Dungeon games, which is what the dungeon generation aspect of this game is based on. I just felt that the dungeons needed to have some flavor, some character to them. I didn't want basic bland empty square rooms, with maybe a few swapped tiles and different music. So again, not the ideal choice.

And so, there's the fourth option. Find a way to create procedurally generated dungeons, while retaining the artistic style of parallax mapping. And that's exactly what I did.



This mockup is still pretty square and grid-based, but I successfully managed to create a system that procedurally generates dungeons using a series of parallax mapped rooms. The game constructs a dungeon layout, then searches specific folders for image files to stitch together, then reads the image file name to get collision and spawn point data for that room. The game then creates a parallax and overlay using those images, changes the tile data within the map, and reallocates the player, enemies, chests, and any other events such as exits and item bags.

I also got creative with how spawn points are set up, so that a player will never spawn in the same room as a monster, and you'll never have two monsters spawn in one room. Events will never spawn on top of each other, and the map will always generate a layout that is fully accessible, with no isolated sections. And it's still parallax mapping! I have full control over what each individual room and corridor looks like, how it's decorated, how collision is determined, and I don't have to resort to using tiles in the engine! Every dungeon can still have the flavor and character that I'm going for, while actually being MORE random than it was before!

And the best part? Every dungeon, no matter how many floors it has, will only EVER need 82 small maps to create fully random, procedurally generated maps. 82 sounds like a lot, but it's really not. Especially because they're very tiny maps. Like, this tiny.



Tiny. 82 of those? I can make that in about two weeks working for less than an hour a day, maybe three weeks if I take a day or two off. And that's for every single dungeon in the game. No more spending months on end working on a single dungeon, getting burned out, and wanting to give up.

So, what does this mean for Thrall? Well, first I'm going to remake all the existing dungeons. Yes, I'm basically throwing away nearly a year of work, at least on the mapping side. But the end result will be worth it. I've already nearly finished remaking the Bloodstone Mine, and I only started on the maps a week ago. Next is Eleston, then the Lyre Forest. Those will take some creative thinking, but I'll find a way to make it work. The end result might be a little bit more simple, not as detailed, but still very distinctly unique.

Once the existing dungeons have been recreated, I will release a new demo with the changes. This will also include the changes to the Difficulty system that I mentioned in the last blog. And then, it's onward and upward from there! The game may very well be completed by the end of 2025, like I wanted! And it will have even more randomness, even more replayability than I'd ever thought possible.

So with that, stay tuned! It only gets better from here.

By the way, I made a few videos talking about dungeon generation in Thrall, so if you want to check them out, head over to the Media page and go into "Developing Systems." There's a video that talks about the first method used to generate dungeons, and I've linked the video talking about this system as well!