A Power-Up That Leaves the Cave Alone
TUNL's laser burns through crystals, mines and whole boulder islands. Adding it without changing a single rock in anyone's daily cave meant it could not be a new coin at all. It had to be an old coin with a new label.
Until the laser, boulders were the one hazard you could not shoot. Bullets spark off them, only a bomb breaks one, and that is by design: a boulder is a rock island with two guaranteed passes, and flying a pass is the skill. The request was for something stronger than a bullet that could break them anyway. The weapon was the easy part.
The weapon
Pick up a laser and for two seconds a beam runs from the ship's nose to the first wall. Crystals, mines and cannon shells in its line are gone at once and pay points like a bullet hit. A boulder stops the beam, starts to glow, and bursts into rubble. I tried two other forms first on paper, a stream of piercing bolts and one charged shot, and both read too much like the ammo coin you already have. A beam is unmistakably a different weapon.
The constraint: one cave for everyone
Every player flies the same cave on a given day. The cave is not stored anywhere; it is calculated from the date by seeded random number generators, one stream per kind of object, as I described in the daily cave post. The catch is that the streams are sequences. If the coin spawner draws one extra random number, every coin after it rolls differently, and two players on different versions of the app would fly two different caves on the same day.
That ruled out the obvious design. Giving the laser its own weight in the coin roll would shift every other coin's share and re-roll every coin from the fourth sector on. I also rejected an upgrade to the ammo coin, which would be unreadable since ammo is spent in under two seconds, and a hangar ship ability, which would turn a power-up into something you buy.
A relabel, not a roll
What shipped is a relabel. The game keeps a cursor, a position further along the cave. Once the cave has passed it, the next coin that would have come out as an orange ammo coin comes out as a laser instead, after every placement rule has already had its say. Then the cursor moves on by about twenty seconds of flight.
The jitter in that step does not come from a random number generator either. It comes from a hash of the position, so it is just as repeatable but consumes nothing from any stream. The result: the cave's geometry is identical to a build without the laser. The cave test replays every day both ways and requires the two to match byte for byte, except that some orange coins now say laser.
Measure the cadence, do not read it off the constant. Because the laser waits for an orange coin, its real interval is the cursor step plus that wait. With the step set to 35 seconds, a replay of twenty days showed lasers arriving only every 40 to 95 seconds, the first one around score 400. At 20 seconds: about every 17 seconds in the fourth sector, every 28 to 40 seconds deeper, the first one at a median score of about 340.
The laser starts in sector 4, together with the boulders it exists to break. Every run below score 233 is exactly as it was. There is a cost: a laser takes an orange coin's place, so deep in a run you see a little less ammo. That is a fair trade for a stronger weapon.
"The laser doesn't break the boulders"
That was the first report from actually playing it, and it was right. The first version needed the beam to stay on a rock for about a third of a second to burn it through. But the beam is straight and sits at the ship's height, and players do not fly into a boulder, they fly its pass. So the beam only grazed the edge for a tenth or a quarter of a second. Of twelve real boulders flown past with the laser on, one broke.
The fix is that the first touch commits: a boulder the beam touches at all ignites, glows and sizzles for a moment, then bursts. Eight of the same twelve broke; the other four were never touched. Aiming the beam automatically at the nearest rock was the other option, and I left it out, because the laser is strong enough already. The lesson for me was to test a hazard along the lines players actually fly, not along the ship's line in a test.
Small things that mattered
- The coin. The first laser coin was a ruby rod in a collar, and in the first screenshot it read as a syringe. It is now a cut ruby firing a short beam.
- The ending. The beam used to fade with a smooth pulse in its last moments, which was easy to miss. It now flickers like a flashlight with a weak battery: short irregular dropouts that grow toward the end, with the hum stuttering along. The flicker is only a look. The beam keeps hitting through every dropout, and the pattern is hashed from the clock, not drawn from a random number generator.
- The length. It began at 2.5 seconds and felt too long in play, so it is now 2.
- The effect. In an autopilot replay over eight days the laser was on for 0 to 7% of the frames and shield hits fell by about 3%. It is a moment of power, not a difficulty lever, and that is what I wanted.
The pattern I keep reusing
The laser is not the first feature shaped by this rule. Repair kits, the wrench a mine or cannon shell drops when you shoot it, live in their own list rather than among the coins for the same reason: put them among the coins and the daily cave forks. In a game whose promise is one shared cave a day, the question for every new feature is the same: does it draw a random number, and if so, from whose stream?
The first laser of the day usually turns up somewhere past score 300 in today's cave.