🎮 TwozyGames
🥌

Button

ORIGINAL

Closest stone to the button takes the end — and the last stone alternates.

Click the game, then use the keys below.
↗ New tab

🎮 1 Player (vs CPU)

Drag back from your stone and release (or ◀ ▶ aim · ▲ ▼ power · SPACE)

👥 2 Players (one keyboard)

Both share one gesture, turn by turn: drag back and release to throw

📋 At a glance

Players
1–2
Same keyboard
Yes
Mode
Head-to-head
On mobile
Yes — 2 players on one phone

In-game: P pause · M sound · R restart · Esc menu · pick difficulty on the title screen.

Free · No download · No account · An original game made by TwozyGames.

🎮 About This Game

Button is curling reduced to the one decision that makes curling a game: where do you put this stone, knowing your opponent throws next. Each end you get three stones, you throw them alternately, and whoever ends up closest to the button — the dot at the centre of the house — takes the end. You do not score one point for that; you score one for every stone of yours that sits closer than your opponent's best. So an end can be worth one, or it can be worth three, and the difference is whether you spent your early stones stacking the house or knocking theirs out of it. The last stone of an end is worth more than the other five, because it answers everything already on the ice. In Button that stone alternates on a fixed schedule — you have it this end, they have it next, all match. That single rule is what keeps the game honest, and it is the one thing here that was decided by measurement rather than by taste. First to five points wins, which lands most matches at around three minutes.

🎯 How to Play

Drag back from your stone and release — like pulling a slingshot. The direction you pull is the direction the stone will not go; it travels the opposite way, and the further you pull the harder it goes. The dotted line and the faint circle at the end of it are not a hint, they are the answer: the stone stops exactly there if nothing gets in its way. Release with almost no pull and the throw is cancelled, so you can drag around and look before you commit. On a keyboard, ◀ ▶ aim, ▲ ▼ set power and SPACE throws. Each end is six throws, three per side, taken in turn. When the last stone stops, every stone inside the house is measured from the button, and the side with the closest one scores for each stone it has closer than the other side's best. If nobody is in the house, the end is blank and nobody scores. Then the hammer flips and the next end begins. First to five points takes the match.

💡 Tips & Strategy

  • Count before you throw. If you are already lying two and they have one stone left, you do not need another stone in the house — you need their line to it blocked. If you are lying nothing with one stone left, a takeout is worth more than a perfect draw that finishes second.
  • The stone that is closest is the only one that matters for the win, but every stone behind it matters for the size of the win. Multi-point ends are how matches actually get decided, and they only happen if you keep stones in the house instead of trading them off one for one.
  • A takeout needs enough power to carry the target off the sheet, not just to touch it. The preview circle shows where YOUR stone stops — if that point is well past the target, you have enough behind the hit to move it.
  • Overpowering is the most common way to lose a stone. Come in too hot and both stones run out the back, which turns your throw into a free clear-out for the opponent.
  • Side walls bounce and the two ends do not. A stone that clips a wall keeps most of its speed and stays in play, so an angled shot into the wall is a real option when a direct line is blocked.
  • A guard is a stone deliberately left short of the house, between the thrower and the button. It is worth playing when you are already lying shot: it does not score, but it makes the answer to your scoring stone much harder to throw.
  • The hammer alternates no matter what happens, so a blank end is not a neutral result — it is an end where you gave up your advantage without cashing it. When you have the hammer and the house is empty, take the single point rather than nothing.

🛠 From the Devs

Button exists because a simulation killed the game that was supposed to be here. The first candidate was a flick sumo — knock the opponent's discs off a table — and it looked fine right up until a positional bot was measured against a bot that simply aimed at the nearest target and shot at full power. Positional play lost, 38% to 45%, and the median game lasted five flicks. There was no decision layer to be good at; it was an aiming test wearing a strategy hat. Curling survived the same test comfortably: a bot that takes out and guards beats a bot that only draws to the centre, 58.4% to 40.7%, at identical aiming accuracy. That gap is the game. But it had one bad number — whoever started with the hammer won 74% of matches, even with the real-curling rule that hands the hammer to whoever lost the end. The obvious fix was to cap ends at a single point, since multi-point ends looked like the snowball. It fixed nothing at all (23.3%), and it quietly made the game shallower, because stacking the house for a big end is exactly what makes placement worth thinking about. What actually fixed it was the boring answer: alternate the hammer on a fixed schedule and let the ends stay worth whatever they are worth. First seat went to 49.1% and the depth stayed. The lesson is one this studio keeps relearning — when a fix is obvious and does not work, the model of the problem was wrong, not the size of the dial.

❓ Frequently Asked Questions

What is the button?

The button is the small white dot at the exact centre of the rings. Every measurement in the game is distance to that point, so "closest to the button" is literally the whole scoring rule.

Why does the last stone alternate instead of going to whoever is behind?

Because it was measured. Real curling gives the hammer to the side that lost the end, and with that rule the side that starts with it won 74% of simulated matches. Alternating it on a fixed schedule brings the opening seat to 49.1% without touching anything else about the game.

Can two people play on one phone?

Yes, and that is the intended way to play it. Button is turn-based and needs exactly one gesture, so the two of you share the same screen and hand it back and forth. There is nothing to react to, so a difference in reflexes changes nothing.

Is the aiming line accurate?

Completely. The dotted line ends where the stone would come to rest if nothing were in the way, calculated from the same friction the physics uses. What it cannot tell you is what happens after a collision — that is the part you are supposed to judge.

More original games

📖 Related Guides