Entry 001 - Why this exists

Sibyl is a small mobile tactical roguelike I'm starting to build, and this is the diary of doing it.

The first reason to write it is selfish. In eighteen months I will not remember why I picked one architecture over another, and Future Me will benefit from being able to read a clear answer rather than reconstruct one. Decisions get fuzzy fast; contemporaneous notes don't.

The second reason is more speculative. If Sibyl ends up being any good - and the bar for "any good" in tactical roguelikes is unforgiving - then there might be a small audience interested in how it got built. Not the eventual gameplay, which speaks for itself, but the order. Most indie game devlogs are about art. This one is going to be about an engine that can play itself ten thousand times before it's been seen by a human, and a bot that loses on purpose so we can find the upgrade that's accidentally unbeatable.

The rough shape: hex-grid tactical roguelike, five heroes that play differently rather than hit harder, framed as the classical Sibyls of antiquity, Hades-style meta-progression around the edges, simple graphics, mobile-first. Strong influences from Hoplite, Slay the Spire, Slice & Dice, and Into the Breach, plus a working understanding of why most solo roguelikes ship undercooked.

The next several entries are going to be unromantic and technical: determinism, replay format, the bot's evaluation function, the balance dashboard. The pretty part comes much later. That's deliberate. If the simulated game isn't fun to watch, no amount of art can save it; if it is, the art only has to be sufficient.

Onwards.