How I ended up building a Hydraulics game for a Game Jam
I didn't set out to make a hydraulics game.
Learn hydraulics by pumping, routing, and racing fluid through 30 handcrafted levels.
I had a game design document sitting in a folder — something I'd sketched out a while back about a stick figure worker on a hydraulic lift. Three levels. Simple stuff. The idea was to teach the player how pressure moves fluid, one mechanic at a time.
Then things got out of hand. In the best way.
I decided to participate in the GameDev.js program
A game jam from 2026-04-13 to 2026-05-09 hosted by End3r & Enclave Games. Gamedev.js Jam celebrates Web games - build an HTML5 game within 1
It Started With One Idea
The whole game rests on a single physical truth: pressure pushes fluid wherever a valve lets it go.
That's it. That's the game. Everything else — the 30 levels, the relief valves, the timed rushes, the two-pump parallel circuits — is just that one idea wearing different clothes.
The first version was almost embarrassingly simple. You had a pump. You tapped it. Fluid went into a tank. That was Level 1. I almost felt bad calling it a level.
But here's the thing: that moment when you first tap PUMP and watch the dashed line animate down the pipe and the tank starts filling from the bottom — that moment works. It's satisfying. The cause and effect is instant and legible. You understand hydraulics now. You just don't know it yet.
The Design Rule I Kept Coming Back To
Every level had to introduce exactly one new constraint. Not two. Not "and also." One.
Level 2 adds a fork with two valves. One path is wrong. You pick. I kept the levels simple and did not introduce Failure levels early on so that the player would keep playing.
Level 3 adds a reservoir — you can't pump directly to the target, you have to fill a middle tank first and let that tank's stored pressure do the work.
Level 6 is your first series chain. A feeds B feeds C. Open the valves too early and you waste all your pressure across empty pipes.
Each time I added a mechanic, I asked: does the player now understand something real about how hydraulic systems work? If the answer was yes, I kept it. If it was just making the level harder for the sake of it, I cut it.
The game is educational in the way a good puzzle is educational — you don't feel like you're being taught anything. You feel like you're figuring things out.
The Mechanics That Surprised Me
The relief valve was the one that surprised me the most. In real hydraulic systems, a relief valve is a safety device — if pressure exceeds a safe threshold, it automatically vents fluid before something bursts. I added it as a mechanic in Level 11 thinking it would be straightforward. Instead it completely changed the skill required. Suddenly the game wasn't about pumping as hard as possible. It was about pumping gently. Patience became the skill. I love when a mechanic inverts what came before it.
The two-pump system was the other one. Level 12 introduces a second pump button — one for each circuit. You have to alternate between them. Your instinct is to hammer one and ignore the other. The moment you realise that neglecting one side causes it to bleed back down while you're pumping the other — that's a real hydraulics insight. Parallel systems need parallel attention.
The bleed-back mechanic wasn't in the original design at all. It came from playtesting with my son who is 11 yrs old and can figure more things out quickly. He would pump up the pressure, open a valve, and then just... wait. The fluid would keep flowing from stored pressure. Which felt wrong — too passive. So I added a rule: when pressure drops below 0.5 bar, all tanks slowly drain back. Stop pumping and the fluid fights you. Suddenly the game had stakes in every moment, not just when you were making decisions.
Thirty Levels Is a Lot of Levels
At some point someone asked me to expand from 12 levels to 30. My first thought was "that's impossible, I've already used all the interesting ideas." My second thought was "let me try."
The key was combining existing mechanics in new ways rather than inventing new ones from scratch. The relief valve by itself is Level 11. A series chain by itself is Level 6. A series chain with a relief valve, where you have to be gentle but also sequential — that's Level 24, and it's a completely different challenge even though it uses nothing new.
Level 30 is the Grand Finale. It has everything: two pumps, a draining tank, a gated valve, a relief valve, and a 90-second timer. Not because I wanted to pile on complexity for its own sake — but because by the time you've played 29 levels, you genuinely know how all those systems work individually. The finale just asks you to run them simultaneously. It's a natural progession, not a trick.
On Sound
Sound design in a small game like this matters way more than you'd think. For a long time the pump button made a unsatisfying thud. Then I swapped it for a "wet boop" — a low sine wave that decays with a couple of reverb echoes — and the whole thing felt more tactile, more fluid, more right. The background music is a pentatonic loop with an industrial percussion track underneath it. Sweet on top, mechanical underneath. That contrast was intentional — the game is about something industrial, but it should feel approachable.
What I'd Do Differently
I'd do more with the visual language of the pipes. The animated dashes tell you that fluid is flowing, but they don't really convey how fast or how much. Thickness of the dash, speed of animation, colour saturation — there's a whole layer of communication available there that I left mostly unexplored.
Play It
Under Pressure runs in the browser. No download, no install, no account. Just open it and tap PUMP.
If you get to the Grand Finale and clear it — let me know. It took me a while too.
