What We Learned

This is the final article in Hage Game's first issue. By the time you read it, we have published 25 original browser games and 30 articles in two languages over about six weeks of work. We did not expect to learn as much as we did. This piece collects the things that surprised us — what we were wrong about going in, what turned out to matter that we did not anticipate, and what we would do differently the second time.

What we were wrong about

We started the project believing the hard part would be coming up with 25 distinct game ideas. It wasn't. By the time we had built game number five, the ideas were arriving faster than we could code them, and we eventually had to close the list because we were generating mechanics faster than the catalogue could absorb. The bottleneck was not creativity. The bottleneck was finishing.

Specifically, the bottleneck was the last 10% of each game — the polish, the edge cases, the tuning of difficulty curves, the writing of the review, the second-language translation, the testing on three browsers. The first 90% of each game took two days. The last 10% took another two days. Twenty-five games times two extra days is fifty days we did not have in our original schedule.

We were also wrong about which games would be most popular. We expected the puzzle games to dominate because they look the most thoughtful; the action games actually scored consistently higher in both our editorial ratings and the informal feedback from early players. Action's appeal is more universal than puzzle's, which we should have anticipated but didn't.

What turned out to matter

The single most important investment we made was the shared stylesheet — the CSS file every page on the site loads. Building it carefully before we built any games meant that adding a new game took half a day instead of two days. Every game's UI is constructed from the same vocabulary of components: stage, HUD, canvas wrap, overlay, controls, help bar. We never had to re-decide how a game's interface should look; we just used the components.

The second most important investment was the article-writing template. We wrote two articles by hand at the start; by article five we had extracted a function that takes a slug, headline, lede and body, and emits a complete double-language HTML page. Every subsequent article used that function. Without the template, the back half of the article catalogue would not have happened — we simply ran out of time.

The third was deciding on the editorial scoring system before we built any games. Knowing in advance what we were scoring on let us evaluate each game's design choices against the scoring criteria rather than against an unstated mental standard. This made the scores more consistent across the catalogue than they otherwise would have been.

What we would do differently

If we were doing this again, the largest change would be to commit to a smaller number — fifteen games instead of twenty-five. Fifteen would have given each game more polish time and produced a catalogue with higher average quality. The bottom five games in the current catalogue (scored 3.5 to 3.6) are present because we set a target of 25 and stuck to it; quality-first selection would have produced a tighter catalogue of fifteen games at 4.0-plus average.

We did not do this because we underestimated the polish cost when we set the target. By the time we knew enough to revise the target, we had committed to 25 publicly and felt obliged to honour it. Future projects: pick a target, then deliberately revise it after the first five units of production, when you actually know how long things take. We will not make the same mistake twice.

What surprised us most

The single most surprising thing about this project was how much the writing mattered. We thought of Hage Game as a games site with some articles attached. By the end, the articles had become as important as the games — they create the editorial voice of the site, they explain why the games are what they are, they give a reader something to do when they have time to read but not time to play. The site is genuinely a games-and-writing project, not a games-with-writing project. We did not see that coming.

The second most surprising thing was that we did not get tired of making games. We expected to be exhausted by game twenty; we were not. The variety across the catalogue — puzzle, action, word, memory, rhythm, skill — kept the work feeling fresh because each game asked a different kind of design question. A catalogue of twenty-five action games would have ended in burnout. A catalogue of twenty-five different categories of game produced its own sustainable rhythm.

What's next

Hage Game's first issue ends with this article. There may or may not be a second issue. We have not committed publicly because we have learned not to commit before we know how long the next thing actually takes. If we do another issue, it will be smaller — fifteen games, deliberately selected — and we will tell you when we have started, not when we plan to start.

For now, twenty-five games and thirty articles. We hope you found something worth your time.

Published · 14 May 2026 · Written and signed by Bill