Notan

At the end of last year¹, I made a long post about the games I played during that year and what I can learn from them as a game designer. I was hoping to make it a yearly tradition starting from that one. But then, you know what happened to cohost.² Luckily I was slowly working on the second edition throughout this year, so I can share the smaller version before the lights go out. There are some rough edges in this post and I wanted to fix them by the end of this year, but I don't have the time anymore. No one does here. So, enjoy the slight messiness I guess.

Alephant

Le Slo's puzzle games are notorious for their difficulty. So much says how his Alephant, a wordless game about language, got nominated for Most Challenging Game in this year's Thinky Awards. But difficulty alone can't make for a good³ game. The game has to show reasons to pursue the challenges. In Alephant, the reason is the stories told in its level design.

I'm not talking about "narrative" stories here. They are not "ludonarrative" either. A story here means an idea conveyed to players about how a level is constructed or how it should be solved. When a level is designed with care, I can feel the designer talking to me:

Don't you think this trick is cool?

How would you untangle this constraint?

It almost feels like having a conversation with the designer. The puzzles in Alephant are focused and packed. Le Slo knows how to effectively tell stories in a small space.

Another unique point of this game is level ordering. The levels are not numbered and there's no immediately clear order⁴, but by the time you finish most of the levels in Chapter 2, you'll realize it has some patterns: In each chapter, the levels are organized into 5 rows. The center row has the most "basic" puzzles. They are relatively simpler than other levels. The top row and the bottom row each has only 1 level. The level at the top is a conversation-themed level, which adds an extra flavor to the regular rules. You can even tell it's a special level by the ox symbol in the level select screen.

The level at the bottom is, however, a massive liar. The bottom level in chapter 1 looks pretty normal. There's no new type of objects, and no special symbols in the level map. But actually it is a wolf playing an innocent sheep. It uses a game mechanic not tutorialized anywhere in the game, which means you can't solve this puzzle while you are thinking under your existing knowledge. You would even stumble on a different new mechanic, only to find it's not related to the solution at all. After all, this "wrong" mechanic is properly introduced in the next chapter. It's just that hard to find the "right" mechanic in this level.

The bottom level in Chapter 2 is similarly difficult. But remember, this game has patterns. It becomes much more manageable once you realize you can reuse the untutorialized trick from Chapter 1. To be honest, as a puzzle game developer myself, I would struggle to tutorialize this mechanic too. Even more so if I had to employ this game's approach and make players naturally learn it through experimentation, as opposed to shoving it into an obvious tutorial level which spoonfeeds the players with the core idea. Le Slo made an interesting design choice to put it away in the corner of optional levels and signal it with patterns.

Gordianaut (from CosmOS 9)

In late 2020, an assortment of small games was released on itch.io and Steam as 10mg Collection. The idea is that 10 different teams worked on 10 small games, each taking 10 minutes to play, and sold as a bundle of $10. I don't have the backstories on how this particular project came to fruition, but I've witnessed some of my fellow gamedevs inspired by it to make small commercial games. Indie devs, especially those with no prior experiences, tend to spend an infinite amount of time making their "dream games", only to get burnt out and not releasing them in the end. In a way, 10mg was counteracting against this gamedev's nature. Empowered by the idea of making small-scoped games, the folks from a puzzle game community⁵ eventually made a space-themed puzzle game bundle, CosmOS 9.

Gordianaut from CosmOS 9 is a short puzzle-platformer with one simple core mechanic: You can control platforms while you are on them. It is exactly how it sounds. There are no mind-blowing surprises like in Baba Is You or The Witness⁶, but it does well what it does. It thoroughly explores the design space, presenting small yet unique ideas in each puzzle. While some levels are fiddly by nature, I liked how each puzzle feels different in its own way.

As a player, I appreciate that people are making shorter games. A massive game with hours of gameplay has its own appeal, but it's infeasible to chase all these big fish in the endless ocean of indie games. Just from 2024, I missed Islands of Insight (Feb), Entwined Time (Apr), Lorelei and the Laser Eyes (May), Isles of Sea and Sky (May), Star Stuff (Jun), Maxwell's puzzling demon (Aug) all because I've been playing 3 big games (Alephant, Can of Wormholes and corru.observer) this whole year. I haven't even finished any of these 3 yet. Sometimes I just want to pick up a small game in a weekend evening and reach the ending by the night. Gordianaut was a perfect game in this regard.

More games I wish I had the time to write in detail

• Can of Wormholes: Interesting mechanical twists and solid level design. This game never ceases to surprise me. No wonder it won Game of the Year in Thinky Awards this year. It has an innovative hint system too.
• Hempuli's Covemountlikes (the first 20 and the rest): Hempuli made 20+ block-pushing games in the span of 5 months earlier this year. Truly astonishing work. In particular I want to mention That's a Warp. It has similar mechanics to one of my old prototypes, and yet it feels so different. I also recommend it as an introduction to this series; it's relatively approachable (but not easy) and you can have a taste for the chaos of these games.
• UDLR-Modify: Unique mechanics and strong level design. Nice bump sound too. I heard they are working on a post-jam version and I'm excited about it.
• Sokobrawn: This game's style of level design is quite unusual. It has minimal tutorial for the core mechanics, and then leaves the rest for players to figure out. Some mechanics are used only once in the entire game and you need to find them in already-challenging levels. It's something I wouldn't do myself, but certainly interesting.
• corru.observer: It feels kind of cheating to mention this game when I've already written about it last year, but I can't skip a game I've played the most throughout 2024. I'm at the end of EP3 (the 4th episode) now and it feels a completely different genre from the beginning of the game. It's really cool. Sometimes I'm lost in the breadth of the gameplay and I need to remind myself I'm playing this game for the story.

---

1. Technically speaking, it was the beginning of this year that it was published. I was aiming to get it done within 2023 and missed the target only by a week, so I think we can all agree to call it approximately the end of last year.

2. 

3. The quality of good is subjective. There can be many ways to define a good game. Here I loosely use the word to describe games that evoke net-positive feelings in/after playing them.

4. This design pattern to have no signposted level order itself is not particularly new in puzzle games of this kind. There are Stephen's Sausage Roll, A Good Snowman Is Hard To Build, and Recursed, just to name a few.

5. It is thinky-puzzle-game, a Discord server for puzzle game developers and players alike.

6. Actually I was spoiled the "surprise" of The Witness before playing it, so I can't know if it's really "surprising" myself. Also I haven't properly played it yet. I'm not sure if it's still worth playing after everything.