Horizon Zero Dawn started with a small team.
And prototyping games is really hard.
I’m going to explain how they managed to build their game mechanics.
This is my synthesis of publicly available GDC Vault talks on Horizon Zero Dawn, not original reporting.
At the beginning Guerrilla started Horizon with a team of 8 to 16 people.
The main composition was:
• Designers
• Artists
• Coders
• Animators
They had a lot of Shooter experience from past games, but no Open World experience.
The main idea was to quickly create prototypes.
If they worked, they iterated on them.
If they didn’t, they ditched them completely.
The main question was → How do we do combat against robots?
The solution:
Come up with the most complex robot possible with a lot of parts.
• Vulnerabilities
• Shooting armor plates to expose weakness
• Disconnecting weapons to use against them
They wanted to test lots of things.
They prepared the prototype to be able to know how the combat would feel.
And most important of all, if that combat would be fun.
At the same time, they continued prototyping other aspects of the game.
The Open World would need to be created.
With the same idea about trying out difficult things they decided to create a complex outpost.
That would help figuring out what they would need to make an Open World game feel alive.
With this prototype they could figure out things like:
• How would the architecture look?
• How many NPCs are needed
• Would the settlement be big?
• Which are the ways the player can traverse the settlement?
And they continued prototyping this and many more things.
For instance, player abilities and weaponry would be another major topic. As well as prototyping mounted traversal, especially from a game design perspective.
This was not an easy task as the mount should be able to ride through a dense forest while the player is shooting.
In the end Guerrilla ended up prototyping lots of things quick. And it was not cheap, a lot of prototypes were thrown away or needed to be redone.
And the main problem was that they had cool protos, but there was not a game there yet.
They shifted their focus to the Core Loop of the game.
The loop became:
1. Explore
2. Interact with things
3. Gather
4. Progress
When combined, the first glimpse of the game started appearing after two years of work.
From this experience the lesson is that a final game vision rarely shows up at the start.
Having a direction helps, but prototyping separate concepts is what lets the good ideas surface.
Recap. Guerrilla started with 8 to 16 people and shooter experience but no open-world experience, so they prototyped fast and threw most of it away. Cool prototypes still weren’t a game. What turned them into one was committing to a core loop, explore, interact, gather, progress, after about two years of work.
What this doesn’t cover. This is the process, not the tech. How the combat, the mounts, and the outpost AI were actually built is the hard part, and it’s a separate post.
