What I Learned Failing to Finish a Game in 2024
georgeallen.dev
georgeallen.dev
1. The Art of Game Design, A Book of Lenses by Jesse Schell - https://schellgames.com/art-of-game-design
2. 20 Tips on Making Games by Jordan Mechner - https://www.jordanmechner.com/downloads/library/20tips.pdf
3. Liz England’s blog - https://lizengland.com/blog/
Thanks for the other links.
To leave something in return, here's something I read the other day and kept thinking about it (I'm designing on a PvP motion based game)
"In competitive games, there is little more valuable than knowing the mind of the opponent, which the Japanese call “yomi.”
As a side note, I would even argue that the “strategic depth” of a game should be defined almost entirely on its ability to support and reward yomi."
The Yomi Layer concept is a reminder that moves need to have counters. If you know what the opponent will do, you should generally have some way of dealing with that.
I think the client work pays the bills though, looking at their catalog.
THAT said, there is a lot of intersting things one can learn from John Carmack, so there's an exception to every rule.
this is a great articulation of what i'm trying to say thanks
Quick browse of the portfolio on that website (https://schellgames.com/portfolio) seems to show multiple games that has won awards. All their original games seems to have "Mostly Positive" or above on Steam as well.
What metric are you judging this developer by? I can't say I've played/seen any of those games myself, but clearly they seem to be putting out some games that people play and enjoy.
I don't know their internal metrics by any means, but none of these games look 'hardcore', they all seem casual
Yeah, same here. But my conclusion from that would be "I'm not the target market", not "Those games are clearly shit because neither I or my friends play them".
Again, looking at the games, they've won awards + are highly rated by people playing them. You just think those people are wrong for liking the games, or something?
I've now been coding for like 14 years, and I still haven't done it (besides a number of prototypes). And I'm so burnt out on writing code that I never have the mental energy to push through and get one done.
I was basically the same. Played video games before programming was even on my mind, and first exposure to programming/structured data was trying to mod GTA Vice City vehicles, and then eventually got drawn into programming while trying to game dev by night basically.
On and off I've tried Gamemaker, Unity, Phaser, Godot, Unreal Engine and everything in-between, for the last two decades or something. It always end up the same, game logic so complicated I can't make head or tails of it anymore, and it was really hard to decouple things enough so I could be as confident editing game logic as I am reading/editing other types of codebases.
So I never really got anywhere, until I found Bevy. I'm not particularly fond of Rust, way too verbose and strict for my taste, but ECS turned out to be a god-send for organizing game code (in my case). Suddenly writing decoupled game logic became a breeze, and since discovering Bevy (but really ECS gets most of the credit here), I've even shipped some games during game jams that I'm moderately proud off and placed well in the ranking compared to my expectations.
If you're of similar traits that you need code to be of a certain quality to be able to effectively work with it, ECS might be up your alley too, and worth a try if you haven't already. It made a huge improvement in terms of how flexible the architecture end up being, and made it a lot easier to incrementally work on games.
If if your game never gets published, etc. you could have a game that you and your friends got countless hours of enjoyment from. I personally get a lot of enjoyment from it. Good luck!
It's a different kind of work from my usual though, and it was fun to see my friends in awe that they could write arbitrarily many new cards and almost instantaneously see them in the game.
Most indie projects die before getting there.
You might spend a ton of time and money on art and polish only to suddenly realize your game isn't fun at all. Many such cases.
I can't help but feel that this completely undermines your point - Vampire Survivors is bashed together using rudimentary knockoffs of sprites from games from the 1990s, in an engine which barely supports the idea of particles let alone proper visual effects.* It is the gameplay that carries Vampire Survivors, not the aesthetic.
Game feel is of course essential to producing a good game all-round, but a competent game designer can and will tell the difference between a good game design and a bad one, way before polish and juice are layered on top.
*I don't say this as a criticism - Vampire Survivors is fantastic - but the idea that it's propped up by its look is just daft.
- Make the game legible. A good example would be to reduce the whole game to boxes, you still need to differentiate things, so you might want to color them. Aesthetics in support of gameplay to make the game understandable.
- Add 'game feel'. This is where audio is especially important as you tend to notice the lack of good audio rather than its presence. But also 'juice', animations and what not all layer in.
- Support the fantasy. The name Vampire Survivors carries expectations that boxes do not match.
If you've ever done a lot of playtesting with your target audience one thing you'll find is that missing these elements gives you much worse feedback. Most notably legibility because it's so integral to being able to play a game but the others as well.
Game designers to an extent can get past this but it's still an attempt at extrapolation which is necessarily less concrete. Also if you're new to making games then you're going to make it harder to judge your own work.
The good thing about Vampire Survivors as an example is it shows that you don't need to do much but enough.
The dev team had just come off developing Titan, a cancelled MMO where they had trouble making the core game loop engaging after seven years of development, so they had a lot of motivation to start small and make something good first, then polish later.
Having watched/read a few things about the white boxing stage the general advice is to put in as much polish as you need to do that and no more. If you're trying to prove out jump mechanics literally just some boxes for platforms and a sphere as the character is enough. If you're making a stealth game then you'll need some lighting in your level because it's a core game mechanic.
The majority of games being released today are in a few crowded genres, that frankly the majority care very little for. Go outside that, the space for competition quickly dwindles that just a proper, non exceptional execution would be sufficient for success.
Take Project Wingman for example, considering there is literally only ONE other competitor in Ace Combat 7 for a modern arcade jet fighter game, it was very much axiomatic from the trailers alone that it would be successful, and it was.
The same thing with ELIN, or Forever Winter, Noita by the sheer virtue of ambition of their ideas, even a flawed execution may be sufficient to drive a loyal crowd.
Success is the conjuction between ambition and execution, and if we presume one is dedicated to execute, then there are plenty of ambitious ideas waiting to be realized. If someone tried to make a successor to Mirrors Edge, or mech game in the vein of ACFA, or an Anime-style FPS, and with decent execution, I would be very sure of their success. It has nothing to do with luck here.
I had this problem, didn't figure out the solution till the end of the project- it's bots. Even bad bots are HUGELY important for multiplayer game development because now you can iterate every second instead of every week. I thought bots would be too hard to make for my game, but they really weren't as they don't even have to be good. With LLMs i'm fairly sure almost any type of game can be botted at this point too.
Similarly, when I’m considering a purchase, I tend to focus on the negative reviews (those rated 3 stars or below) rather than the glowing, positive ones. Negative reviews often provide more logical, specific reasoning as they shed light on potential deal-breaking issues. That said, they can sometimes veer into irrelevant complaints that don’t resonate with me. For example, when I’m browsing book reviews on Amazon to decide whether a book is worth reading, I frequently come across one-star reviews criticizing the print or paper quality. If I’m planning to buy the digital version of the book, those complaints become irrelevant to my decision-making process—even though they might be incredibly important to someone else.
In essence, I find value in the nuanced, sometimes brutally honest critiques that failures and negative feedback offer. They paint a more realistic picture—one that helps me make better decisions and understand the world a little more clearly.
> All happy families are alike; each unhappy family is unhappy in its own way.
I’m guessing these days there are placeholder art libraries available?
I have done work in embedded game dev off and on for about 20 years, and I could have done so much more had I had even one ounce of artistic ability.
And (other than hobbyist or OSS), it's very hard to use canned artwork. Everything just needs to be unique for a commercial offering.
But in all fairness, I don't think many of the artists I worked with could code. Just seems to be opposite skillsets (beyond just the creativity).
I think learning to see past this and be able to evaluate "Is this fun?" regardless of it looking like shit is a skill to learn like any other.
A great way to train this is to start playing random games people publish on low-stake platforms like Itch.io. Most of them lack in the art department, but even some of those have really addicting gameplay hooks, or otherwise novel gameplay elements you can notice shines through the awful art.
Hopefully after a while you'll be able to discern more between "Is this not fun because it doesn't look fun, or because it doesn't feel fun?"
I say this because it means you're forced to focus on the gameplay and story. If they aren't compelling, graphics (usually) won't save you anyway.
At least on the surface, thats what I see. Wireframes go a long way.
This is a contrived example but, what if I wanted a Walrus riding a surfboard, wearing a top hat, holding a katana in his right hand, and holding a slice of Hawaiian pizza in his left hand.
Despite the biases people have against generative AI, it will solve a LOT of problems in this space.
Which many will. Just look at the indie games on Steam, a vast amount of them use pre-existing assets.
Jim Sterling poisoned an entire generation of gamedev's minds against it, but there's nothing wrong with doing so.
>You would have to commission an artist to create them, which costs money that you as a bootstrapped independent developer may not have.
If you want a game with quality, unique artwork, likely a style you want to build a brand around and monetize, you should be willing to spend the money on an artist to create it. Using a technology which is trained on the work and style of artists (without their permission, mind you) to extrude an art-like product just to avoid having to pay for it is gross.
>It also takes considerably more time than generative AI does.
Does it? Chances are that "Walrus riding a surfboard, wearing a top hat, holding a katana in his right hand, and holding a slice of Hawaiian pizza in his left hand" is going to be replete with errors, not have a consistent style, have bad geometry if it's a model, and need to be edited anyway. It isn't going to be what you imagined in your head, because generative AI is a mediocrity machine, and it isn't going to compete against a coherent design implemented by real artists who care about their work.
You'd be better off just buying assets or hiring an artist either way. It isn't even that expensive, artists are desperate for work now that AI is eating them alive.
As usual, it seems like we're maybe _almost_ there (if you can generate video, like with OpenAI's Sora, you could probably get a walking animation for a character, and I've seen proof-of-concepts which produce not-rigged 3D models), but it seems like AI can't do a lot of things that you would want for game development.
The one thing it _is_ really "good" (emphasis on the quotes) at is generating static 2D assets like character portraits, HUDs, item/skill icons, etc. Unity's asset store is now full of gen-AI stuff like this (lots of packs of 1,000 spell icons, which are all basically variations of a 'fireball', except maybe this time it's green, etc).
Understanding that the general public should not and indeed cannot be expected to see the potential is also a hard lesson. Games like VVVVVV are unicorns. Never bet on being a unicorn.
Risk of rain comes to mind as a great multiplayer videogame with a small team, it was made with gamemaker studio.
I am curious how is the ecosystem right now and if Godot has become a more attractive option for solo/indie development.
In the same vein, I can recommend this book: https://www.goodreads.com/book/show/34376766-blood-sweat-and...
It shows that even big companies and development efforts can often struggle to create a fun game, even when the people involved have a lot of experience. It is a hard thing to accomplish!
To be sure, there were a lot more game prototypes that got swept into the bin. Often for the same reasons the author mentioned (specifically there were several I knew were going to be too much of a time investment to do properly).
EDIT: I have already posted the first two volumes (disk image) of unfinished games. I can easily create a third volume. Here is volume 1:
https://github.com/EngineersNeedArt/SoftDorothy-UnfinishedTa...
Nice work and best of luck with taking game #3 forward!