Grapple God
A game about grappling your way out of hell. Built by a team of three in 72 hours for SpeedJam #5.
A game in one weekend
SpeedJam #5 challenged us to make a game built for speedrunning. Over 72 hours, I worked with Hugo Belperron on art and Clyphdrops on music and sound effects to create Grapple God in Unreal Engine 5. Players use a grappling hook to climb a tower out of hell, building enough speed to crash through demons along the way.
I handled the programming and in-engine implementation, including the grappling system, enemy behaviour, and collision and damage handling. With so little time, we prioritised the interactions that made the game enjoyable and used playtesting to guide our changes.
Download and play Grapple God on itch.io ↗
Project challenge
Making the grappling hook feel good in 72 hours
The grappling hook was the centre of the game, so getting its movement right was a priority. Players needed to build momentum and judge when to release the rope, with enough control to make precise moves while travelling at speed.
I built the physics-based grappling system in C++ and tuned the movement beyond the basic rope pendulum. Through playtesting, I adjusted the movement and physics parameters to find a balance that felt responsive and enjoyable.
Combat used the same momentum: players killed demons by crashing into them at high speed. I implemented the collision and damage handling, then added a brief slowdown on a successful kill to give the impact more weight.
Player feedback repeatedly picked out the movement as smooth and satisfying. For a game made over one weekend, seeing people enjoy learning the hook and building up speed was a rewarding result.
Project challenge
Keeping a speedrun consistent
Enemy behaviour needed careful restraint in a game built around repeat runs. Players should be able to learn a route and improve their timing, so I kept the demons' AI simple and limited randomisation.
The small exception was the angle of their fireballs, which varied slightly. That introduced some uncertainty without making the entire encounter unpredictable, so players still had to react while following a familiar route.
Performance also mattered to that consistency. I used Unreal's profiler to find the most expensive parts of the game and disabled costly default engine features where needed. This helped us keep the game running smoothly across different devices within the jam deadline.
What players thought
Players responded warmly to the movement and Hugo's art, with several saying they'd like to see a full release. Some also pointed out rough spots in the level design, which was useful feedback on what we'd need to improve beyond the jam.
