Outlast three AI bombers. A sunbeam sweeps the arena, and any bomb still in the light when its fuse runs out fizzles.
How to play: arrow keys or WASD to move, Space to drop a bomb (tap the arena on touch screens). Fuses last 2.6 s. Watch the beam’s direction and speed, then drop your bomb so it goes off in shade. The AIs do the same, so expect them to wait for the right moment.
Power-ups drop from crates: R longer blasts, + extra bomb, S speed, ☾ eclipse charge (used automatically on a bomb that would otherwise sit in the beam). Blasts destroy power-ups they reach.
Autoplay: tick “AI plays for me” to watch the AIs play all four seats. The match restarts by itself after each round.
Learning AI: when on, the bots learn from how they die and tune three habits: how long an escape route they accept, how wide they treat blasts, and how much timing margin they want around the beam. What they learn is saved in this browser.
Lessons from Solar Flare
The big idea is timing: even the best-laid plan and the most powerful tool only work inside the right window of opportunity. Here is what playing, or just watching the AIs, can teach.
1. A window beats power
In the game: a bomb that goes off inside the beam does nothing, however well placed. In real life: a launch, a pitch, a deploy or a hard conversation can fail purely because of when it happens. Ask “is the window open?” before “is this good?”
2. Read the forecast
In the game: the HUD counts down to the next beam and shows its direction, and a chime marks when it starts. In real life: leading indicators do the same job: calendars, budget cycles, change freezes, market conditions, or the hours when the person you need is actually free.
3. Waiting is an action, but not forever
In the game: Normal and Hard bots hold fire until their bomb will detonate in shade. Easy bots rush and waste bombs. In real life: patience is only a strategy when you know a window is coming. Acting too early wastes effort, and waiting without a deadline wastes opportunity.
4. Plan your exit before you commit
In the game: bots only drop a bomb if they can reach safety in time, and many early deaths come from a bot’s own bomb. In real life: rollback plans, exit strategies and financial runway should exist before you launch, not after.
5. Systems chain together
In the game: one blast can trigger nearby bombs, which can catch you from an unexpected direction. In real life: dependencies create cascading failures. Leave slack between tightly coupled parts, and ask what else your action sets off.
6. Prepare for the off-window
In the game: an eclipse charge lets one bomb work in the light, but it is scarce, so spending it well matters. Range, speed and extra bombs make each window more productive. In real life: reserves, relationships and skills let you act when conditions are poor, but they are limited. Save them for moments that matter, and invest in being able to use a window quickly when it opens.
7. Learn from failure in small steps
In the game: the Learning AI sorts each death by cause and nudges three habits a little at a time. Big corrections would make it either timid or reckless. In real life: run blameless post-mortems, name the cause, and change one thing at a time.
8. Observe before you act
In the game: turn on “AI plays for me” and just watch. In real life: watching others is cheap data. Compare Easy with Hard: Hard is faster, but its real edge is that it almost never wastes a bomb in the light.
Try this
Run autoplay on Easy for three rounds and count the fizzles, then repeat on Hard.
Play with sound on and use the chime as your cue that a new window has started.
Turn on Learning AI plus autoplay for ten rounds and watch which cause of death shrinks in the panel.
In one round, drop bombs the moment the beam leaves a tile. In the next, drop them just before it arrives. Compare how many fizzle.