Guide and planner · The Dream Game Sprint
The roadmap we use: five stages (the 5 P’s) day by day, what “done” looks like each week, and the rules that keep a small game small enough to finish. Set your start date and it becomes your calendar.
Pick your Day 1. The track shows the five stages, the five playtests (sessions where you watch someone play) and the dates that matter. Everything saves in this browser.
By hand vs with the coach
This calendar works on paper. The Dream Game Coach turns it into your one next step every day: it runs your playtests and checks, keeps your dates, and warns you when your scope starts to slip. See Premium
The plan assumes about 1.5 hours a day, around 45 hours in total. That fits people with a full-time job. Hobbyists who finish games usually manage 10 to 20 hours a week, often in short sessions: one developer made 12 games in a year working 30 to 45 minutes before bed plus weekend mornings.12
Expect a big share of that time to go to things that aren’t building. One first-time solo developer logged their hours: about 43% went to fixing bugs, testing, admin and marketing.3 The plan leaves room for that.
Behind schedule? Cut scope (how much you’re making). Never cut playtests, and never cut sleep.
Studios use these words loosely. Use them strictly: each one has a test you can answer with yes or no.45
| Milestone | Day | What it means | The test |
|---|---|---|---|
| Toy (main action, no goal) | 3 to 5 | A rough version of your one main action, built from plain grey blocks. No goal yet. | Can a stranger play it for a minute and want another go? |
| Last pivot (last big change) | 6 | Playtest 1. Keep the idea, twist it, or change it. After today, new ideas go on a list for the next game. | Did someone who isn’t you enjoy the main action? |
| Feature freeze (no new features) | 14 | Every feature exists and the game plays start to finish, even if rough. Nothing new gets added after this. Your art style is locked too. | Is everything on your Must list in the game? Then stop adding. |
| Content lock (everything is in) | 22 | All levels, art, text and sound are in. Stand-in art and sound are gone. From here you only tune, fix and polish. | Is there any grey box, “TODO” or stand-in left? |
| Release candidate (ready to ship) | 28 | A build you’d be happy to ship today, with no known bugs that break the game. | Would you upload this exact file if launch were in an hour? |
| Ship (it’s live) | 30 | The public release, then a short post-mortem (a look back). | Page live, game playable from a link, first post sent? |
Tick things off as you go. The dates follow your Day 1.
A playtest means watching someone else play your game. You don’t need many people per round. In Nielsen’s model, one tester finds about 31% of the problems and five find about 85%. After that, each extra tester finds less and less, so three small rounds beat one big one.6
With people who had never played it, watching from start to finish. Testing began with one half-finished room.7
After each round, group notes by pattern. Fix what two or more testers hit, and make one moment they loved even better. Park the rest.
Adapted from Schell Games’ “ffwwdd” playtest questions.9
A game that keeps growing (“scope creep”) is one of the problems that comes up again and again when researchers read hundreds of game post-mortems.10 These rules keep a 30-day game a 30-day game.
Stores take longer to let you in than your game takes to build. Start the slow ones on Day 1, or plan them for your second release.
| Where | What to know |
|---|---|
| itch.io | Free, and you can go live today. Keep the page private until it’s ready: a new page only gets one turn on the “Most Recent” list.13 |
| Steam | $100 per game, and about 30 days from paying to your first release. Best as your second release.14 |
| Google Play | New personal accounts must run a private test with real people for two weeks before release.15 |
| Web portals | CrazyGames and Poki put your game in front of their players for a short test first. Poki asks for web exclusivity: your game can’t be on other websites.16 |
In a study of 149 adults working on a four-week goal, the group that sent weekly progress reports to a friend did best.17 But announcing who you’ll become (“I’m making an epic RPG”) can make people work less, because it already feels like progress.18
Share progress, not promises.Each week, post one GIF (a short looping clip) and two lines: what you built and what’s next.
A devlog (a short progress post) on your itch.io page or a post in the AI Game Bootcamp community both work. Don’t expect devlogs alone to bring players; treat them as a way to keep yourself on track.
Within a week of launch, while you still remember, write down what happened. This look back is how one finished game becomes a better second one. Use this template:
AI Game Bootcamp Premium
This guide gives you the plan. Premium gives you the coach and the lessons to follow it all the way to launch.
The 30-day plan is the Dream Game Coach’s schedule. Outside sources were checked in October 2026.