About Bysoft Games
We are a very small studio. That is deliberate. Not a stage we are trying to grow out of.
Why small
A game with one clear mechanic can be finished, tested and polished by a handful of people. A game with six systems usually cannot. What ships instead is six things that each work at about seventy percent, and nobody remembers it. We would rather make something narrow that feels finished.
Being small changes what we can be honest about, too. When a feature turns out to be bad, we delete it. There is no roadmap review to sit through first. The cascade bonus in Pebble Merge is the clearest case: the version we designed first measured worse than having no bonus at all, so we cut it down to almost nothing and shipped that.
How we decide things
Opinions about game feel are cheap and usually wrong, so we try to measure instead. Concretely, that has meant:
- Checking colour separation with a perceptual colour-difference formula rather than by eye, because the first palette we liked turned out to be nearly indistinguishable for the four smallest stones.
- Running thousands of simulated games to test whether a mechanic rewards skill or luck, before building the mechanic.
- Treating any change that makes numbers look much better as a bug report about our own method, not as a win.
We publish those measurements when we can. Not because anyone is waiting for them, but because writing them down is the only reliable way we have found to catch ourselves reasoning backwards.
Where our games live
Every game we ship is a folder of static files. HTML, JavaScript, a small physics library, and sprites generated at build time, served from an edge network. No application server behind them, no database in front of them. That is partly stubbornness and partly arithmetic: a game that is just files cannot have its user table breached, cannot go down because a background job died, and costs roughly nothing to keep online for years.
It also forces a useful constraint. Without a server we cannot run an authoritative leaderboard, so we do not pretend to: leaderboards in our games are local, and we say so on the screen rather than implying a global ranking we cannot actually verify. Designing around what we genuinely cannot do turns out to be more honest than building a server to fake it.
How we ship
A game is not finished when the mechanic works. In our experience the last third of the effort goes into things nobody notices directly: how long a stone keeps moving after it lands, whether a sound sits in key with the one before it, whether two tiers of pebble can be told apart at a glance on a cheap phone in daylight. We treat those as the actual product.
We also test the parts that are easy to skip. The physics runs headlessly in a test harness, so we can play thousands of games overnight without rendering a single frame. Colour separation is checked with a formula rather than by eye. Neither of these is glamorous, and both have caught changes that felt right in the hand and measured wrong.
What we do not do
- No loot boxes, no randomised purchases, no gambling-adjacent mechanics.
- No engagement tricks: no loss-framed notifications, no artificial scarcity, no “your friend just beat your score” pressure.
- No tracking beyond what a page needs to function. Our games keep your best score in your own browser. We never see it.
- No advertising inside gameplay. Ads, if any, belong on pages around a game, never between a player and the thing they came to do.
Working with us
If you run a games portal and want to license one of our titles, or you are a developer who wants to know how something here was built, work with us. We read everything.