I was watching our own play sessions and realized we were paying a compounding cost for v2 complexity: too many systems to explain, too many edge cases to tune, and too little immediate fun when a player picked up the controls. If we don't stop feature growth now, we will have to keep shipping patches that make the build harder to read without making it better to play.
That was the trigger for a hard call: freeze v2 to maintenance-only, start a fresh track, and force every new mechanic to prove it improves the moment-to-moment loop.
Simple combat loops earn trust when stakes stay visible.
Why the old path stopped feeling good
v2 was not broken in the narrow sense. It still ran, and parts of the encounter design were technically interesting. The issue was practical: cognitive load rose faster than joy. Players had to parse too much context before making meaningful decisions. We drifted from a pick-up-and-play target into a study-first target.
That mismatch showed up as readability drag, slower balancing, and blurred ownership of outcomes. None of those are dramatic failures alone. Together, they create the classic trap where a game looks busy but feels flat.
What we made deliberately smaller
- Top-down movement in any direction
- One attack button with immediate feedback
- One monster with clear telegraphing
- Death + restart loop that is instant and understandable
That baseline is intentionally humble. The goal wasn't novelty. The goal was legibility and repeatability. If the core loop is not fun at this scale, extra content only makes the diagnosis harder.
For example, combat feedback no longer depends on layered hidden modifiers. A swing either lands or misses based on distance and facing, and the event feed explains what happened. This improved player learning and debugging in the same move.
For example, restart semantics are explicit: run kills and gold reset, lifetime totals persist, and restart is one button. Players no longer have to reconstruct invisible state rules after a death.
The one major mechanic this cycle: two weapons, one clear tradeoff
After the first playable loop held up, we added exactly one major mechanic: a second weapon archetype. The shortsword is fast and lower damage. The battle axe is slower and heavier.
The tradeoff is obvious in behavior and UI copy, and players can swap in-run with one control input. That adds meaningful choice without adding conceptual overhead.
Why this simplification improved fun
The win is not that the game became easier. The win is that decisions became clearer. When loops are readable, mistakes feel fair. When mistakes feel fair, retries happen faster. Faster retries make difficulty tuning meaningful instead of defensive.
Two practical gains showed up immediately: pacing clarity between weapon rhythms and faster new-player learning from immediate feedback. We did not need a talent tree, stamina economy, or status matrix to get that result.
What we intentionally did not add
We skipped extra monster archetypes, combo trees, procedural loot complexity, and the town economy in this cycle. Those are roadmap items, but each is a complexity multiplier.
The rule now is simple: if a feature does not improve immediate fun or clarity in the current loop, it waits. That rule prevents us from mistaking motion for progress.
Takeaway
Bottom line: simple loops are easier to make honest. They keep stakes visible, reduce design self-deception, and accelerate learning for both players and builders.
What changes tomorrow: add two monster archetypes with readable telegraphs, then verify weapon choice still feels meaningful before touching economy systems.
Review artifacts
Internal contract/evidence stays in notes and reviews rather than the narrative body.