A game should end
Ideally while you still want another game. We avoid draws where we can.
About
Where the studio is heading, why we started small, and the rules we design by.
Why old games
Tafl was played across northern Europe for the better part of a thousand years. Backgammon's ancestors were rolling in Persia. Nine Men's Morris is cut into the cloister benches of English cathedrals, where bored monks scratched a board and played through the sermon. They lasted because people kept choosing to play them.
That is a longer playtest than any studio can run, and we take it seriously. Most digital versions of these games copy the rules, add a wooden texture and supply an opponent that is either hopeless or unbeatable. We want to give them proper art, opponents whose strength has been measured, and a reason to come back to them.
Some old games we rebuild whole. From others we take mechanics that have lasted (the shove, the throw of lots, the race, the ring you can be pushed out of) and build new games on them.
The road
This is the plan in order, with no dates attached. We are on the first step.
Dohyō, on Android: a small, complete game with a tutorial, eight computer opponents, four worlds and twelve languages. It is in closed testing now. We wanted to finish one game properly before starting on the others.
Names and art live in the theme and not in the rules, so a new world is just a data file. That lets us add settings without rebuilding the game each time, so a player who likes the rules but not the setting can pick another.
A physical edition of the same game, with the app beside it as opponent, referee and teacher. The box and the app would share one tested ruleset, so you could play either on a given evening.
Pieces you reach for, on the table you are actually sitting at. The harder question is whether that adds anything a good wooden board does not. We think it could, for playing alone and for playing someone far away as if they were across the table.
What we hold to
These are about how a game feels to play. When we disagree about a design, they settle it.
Ideally while you still want another game. We avoid draws where we can.
A turn with only one sensible move is not much of a turn. Every turn should involve a real choice.
If players have to read the rules twice, the design needs work. You should be making real decisions within five minutes.
If a piece or option is always the wrong choice, we fix it or remove it.
Behind the scenes
A game design is a guess about how a game will behave. Usually the guess is settled by argument, by instinct, or by releasing the game and seeing. Simulating the ruleset settles it before anything is built, and the most useful result is finding out where the guess was wrong.
In Dohyō it found a bias towards the second player that nobody had noticed, and the fix removed a rule as well as the bias. It also ended a Monte Carlo tree search opponent we had already finished, because a much simpler evaluator beat it. That one is written up in the devlog. And it turned "the Warden feels weak" into a specific answer (the piece lacked movement, not strength) and a change we could check.
Taste still makes the decisions. The simulation handles the arithmetic.
Old mechanics, carefully tested, for people who like the craft of play.
Get in touch
Dohyō is in closed testing on Google Play, and we need people to play it on their phones and tell us what the simulations could not. If that sounds like you, join the testers list.
Press, collaborators and anyone who wants to talk about strategy games are welcome to get in touch too.
Two-player abstract strategy · Android · hotseat and single player · offline, no account, no telemetry · built in Godot · twelve languages · four interchangeable themes.
Status: closed testing on Google Play. No release date announced.