Pawnborn
Fair by construction, not by promise
Every game says it is fair. The word costs nothing to print, which is exactly why it means nothing on its own. The interesting question is not whether a studio promises fairness — it is whether the promise is the kind of thing that could survive a bad quarter.
Two kinds of promise
The weak kind is a policy. We don't currently sell power. Policies are true until the revenue meeting where they are not, and by then several hundred thousand people have arranged their lives around them.
The strong kind is a constraint the system is built around, where removing it would mean rebuilding the game rather than changing a setting. That is more work up front and it is the only version worth anything.
What that looks like here
Pawnborn's version is a single sentence: every involuntary battle between two prepared players is fair chess — identical pieces, equal clocks, colours by a coin. Kingdom power buys stakes, information, preparation and opportunity, and never material, tempo or clock.
There are exactly three sanctioned exceptions, all of them consensual or non-human: fights against the world's monsters, which fight strange by design; a border left so undefended that pretending an empty field is an army would be the dishonest thing; and a wager two players freely agreed to. That list is closed. It is not a starting point that grows.
What makes it a constraint rather than a policy is that it is enforced downward through everything. There is one table in our design documents whose right-hand column is "effect on the chessboard", and almost every row in it reads none. A proposed feature that would put something in that column does not get a debate about balance; it gets rejected on the grounds that it is a different game.
The rating is a closed human loop
A smaller example, which we like because it is the sort of thing that is very easy to get wrong by accident. Your chess rating in Pawnborn moves only when two humans play each other. Any game with a machine in a seat is unrated — for both sides.
This costs us something real: sleeping players' defences fight a great many games, and none of them will ever move a rating. What it buys is a rating pool with no inflation in it at all, and one sentence we can say without an asterisk: the number next to your name came from beating people.
And because a machine can fill a seat, we made the machine announce itself — the board says, in the world's own words, that this seat is a sleeping rival's defence rather than the rival, before the first move, to both players and to anyone watching. Knowing your opponent is human is only worth anything if you know it while you are deciding whether to play.
Money
The same reasoning gives the business model. We sell beauty: heraldry, court themes, the architecture of your kingdom. Nothing sold affects a game, and the rarest things in the game are earned and cannot be bought at any price.
There is a commercial argument for this and we may as well say it out loud, because a promise that also makes money is more durable than one that only costs money. In a market full of pay-to-win strategy games and chess cheating scandals, being the game where fairness is structural is not a sacrifice. It is the most interesting thing we have to sell.
The two Covenants are written out in full here.