A Deathless Roguelike

3 September 2026

In Love Me Not you are a flower that climbs out of a cave. The cave has been cursed by a witch, so the level is regenerated above you whenever you fall.

This is similar to the procedurally generated levels in infinite runners like Doodle Jump, except that instead of erasing the entire level on a failure, Love Me Not only erases the most recent sections. This style of level generation is novel to me. For the purposes of this post I’ll call it spool generation.

I quickly became attached to spool generation. Procedural generation has a special place in my heart, but traditionally it’s a one sided conversation. The level is generated and statically presented to the player. Spool generation opens a dialogue. The player can accept or reject sections of a level. Rise to a challenge or retreat. Surpass a difficult area and then see it erased.

At the same time, the generated spaces can be alienating. They are amorphous and struggle to distinguish themselves. They never feel homely or familiar. They can’t just be received.

I wanted to write a little bit about how spool generation was worked into Love Me Not.

I often simplify a player’s time with a game into a “novice period” and an “expert period”. These are usually characterised by different goals. In Love Me Not a novice’s objective is to escape the cave by any means necessary, while an expert’s objective is to escape the cave as fast as possible.

Spool generation had some natural effects on the novice period:

1. You only get one major attempt at a level section. Improvisation is favoured over memorisation.
2. Repeated falls don’t result in repeated climbs.
3. The magnitude of failure is important. Small falls result in small regenerations.
4. Repetition and luck in hard sections will eventually result in easier sections being generated. Dynamic difficulty is implicit.
5. The variety of multiple levels is expressed in a single continuous session. Novice players can experience longer stretches of focus.

The expert period was more fiddly. Time attack is a natural fit for this style of game, but it makes spool generation irrelevant. Any kind of fall is going to jeopardise your personal best, so you’re better off resetting the run. You never see the level regenerate.

This made the mechanic feel like an onboarding novelty. I wanted to find a scoring system that could bring spool generation back into the decision space of an expert player.

Time freezes seemed to work. After you fall far enough, Lomen the Lizard spawns above you, marking your “best height” with her tongue. If you ascend back above the tongue then the clock is briefly frozen. The freeze gets shorter each time it is activated.

The time freezes created a new gesture. From time to time, optimal play requires you to intentionally fall, spawn the lizard, and quickly climb back up. Because the climb regenerates part of the level, deciding where to fall becomes quite a complex decision.

Here are some things I consider when deciding when to cash in time freezes:

1. I’m about to reach an overhang which can propel me downwards faster than falling.
2. The next section looks difficult. If I fall now I might get an easier one instead.
3. The next section looks easy. If I fall now I might waste a time freeze I need later.
4. I’m approaching the finish line. If I don’t fall now most of my time freeze will be wasted.
5. I know a fly is directly below me. If I fall here I can use it to quickly spring back up.
6. I’ve just missed an edge I was hoping to land on. Escalating the fumble into a tumble could convert it into a helpful time freeze.

The time attack now feels strange. Previously it had been a familiar meditative state, a pursuit of perfect execution. Now it’s more critical and reactive. I’m reading the layout of the level, plotting how I could move through it, and deciding if I even should.

Thank you for reading. I’d love to hear your thoughts.