Dev Log

Twice in twenty-four hours the assistant told me that the right decision here fundamentally changes the game. Not as a cost to weigh against a benefit. As the recommendation. This is a fan adaptation. The premise is the video game.

I went to Protospiel with something that looked like a finished game and came home to pen and grid paper. In my professional life, going back to wireframes is never a good sign — it means there is a fundamental problem in the usability of the system. A publisher advised me to prove the core loop before making the rest using hand written tiles for speed-to-test. I took his advice. And it took me all day.

Protospiel Chicago, September 2026. I packed the game as neatly as I could, and it looked like an actual game prototype, not a bunch of cardboard. However, not one playtest made it to the end. The reason was a design flaw that had been evident since May.

I've spent months trying to define fun so I could design toward it. Then a tower showed up with a button on it, and every argument got shorter.

Somewhere between paper prototype and a finished, cohesive object. Better cards on flimsy stock. Tiles that are gorgeous and clash with their own border. Props pulled from wherever. A month of work that's visible, and not yet done.

Taking a month to focus on a visual upgrade and art production. Ten decks, one shared frame — and a footer line that turned out to be the whole rulebook in miniature.

I didn't wake up planning to reorganize the repo. I woke up wanting to go back work on the manor. These are not as different as they sound — and somewhere in the middle is something I'm still figuring out about working with AI.

Overcoming a career's worth of print anxiety in one session. The first custom tile order has shipped. Here's everything it took to get there.

In Post 03, we cut room rarity. The problem it was solving didn't disappear — it just went quiet. A German farmer's game just gave it a new voice.

Three wrong answers, one right one. How to teach the player room orientation.

A collaboration-with-A.I. journey. I set out to design a board game. Somewhere along the way I became a person who arbitrates git workflows.

Two months in. The prototype finally knows what it wants to be.

Three hours. Two players. Day 3, one room from the Antechamber. A session that kicked me off my high horse and handed me the most valuable design feedback yet.

Three dice. Two kinds of cards. And the moment a rulebook stops being a list of instructions and starts being a grammar.

It's Memorial Day weekend. I should be on a boat. Instead I'm sitting with a list of things I might cut from this game.

Two playtests in, I replayed the video game from the beginning and took notes on everything I'd stopped seeing. The item system needed a rethink. And four boxes labeled Confidential changed how Special Keys work.

A solo run that made it deep into the Blue rank before the house won. What the Fixed Stars booklet fixed, what the Commissary didn't, and two overlapping locks that ended everything.

A surprise Saturday drop-in. A friend who has never touched the video game. Forty minutes later, he wins. And somewhere in between, he says things that only a completely fresh set of eyes could say.

Nine constellations, a 50-box tally grid, a permanent campaign resource that accumulates across every run. All of it gone. What's left is a six-row reference card and a single d6 roll. The question was whether to be faithful or whether to be playable, and I put off answering it for weeks before realizing it was the same question.

A three-phase CSS cleanup across twenty-five files. The codebase got bigger. The site got easier to change. Those two things are not a contradiction — and it took doing the work to stop confusing one for the other.

An audit against the video game's authoritative source found six room tiles with the wrong gem cost on the back. One more was over-priced in the opposite direction. The prototype handled it the way paper prototypes are supposed to — a marker, a scribble, done.

Campaign 4, Day 1 — complete. The first Day 1 win across four campaigns, reached the long way around: thirty-five rooms placed, a near dead-end pulled apart by a single draft, and a streak of favorable draws at the top of the mansion.

Campaign 4, Day 1. Still no gems at 45 minutes. A Music Room roll that felt brutal in red ink. And the moment the Apple Orchard unlocked, and nothing at the table was ready for it.

Campaign 1, Run 3 mid-session. The prototype works. The table doesn't. Plus: starting coins confirmed, Secret Passage contrast failure, and treasure maps drawn upside down.

Campaign 1, Run 2. Eight rooms placed, 42 steps remaining, 0 gems from start to finish. Steps aren't the problem. The gem economy is.

Campaign 1, Run 1. Fifty steps, forty-three coins, and nowhere to go. The Vault at Rank 2 broke the economy in four turns and told us something important about flat probability pools.

Twelve errors in a rules doc that felt finished. How they got there, why they survived, and what it took to find them — including apostrophes that broke the room counter and a mechanic attributed to the wrong room.

A session of looking things up, getting things wrong, and making cleaner decisions because of it. The fruit system, the Showroom question, the Telescope, and why the card is the item.

A week off, a trip to New Orleans, and an accidental visit to the Sazerac House — a building organized exactly like Mt. Holly Estate, down to the draft mechanic.

On being eight years old, writing to Nintendo, and why this is technically game number two. The origin story of the adaptation and why it needed this many documents.

Rarity is gone. The bag is the only deck. Movement is declared, not tracked step by step. A full design session rebuilt core assumptions — and produced the Room Glossary, a reference card system, and a much cleaner rules document.

First playtest session. The active deck was a precarious stack. The tiles were too small. The pawn moved step by step and the overhead was real. Everything that broke, and what we built to replace it.

The mathematical bones of the adaptation — the step economy, the Slot 1 reprieve, conditional drafting, the lock dice system. The first articulation of what this project is trying to be.