A chain reaction in this puzzle is a deterministic wave that starts from one selected source unit and only earns the full 1,000-point score after all five 5-by-5 workbenches are cleared, one tick group at a time. The field is already populated when you load a level, and you choose exactly one unit to launch the experiment. That source emits on tick zero. Any uncleared unit within its Manhattan range activates on the next discrete tick, and every newly activated unit emits its own wave one tick later. The reaction continues until every unit is cleared or no new unit can be reached. Clearing every unit on a single field awards exactly 200 points; five full clears therefore total exactly 1,000. This is not a domino placement game. Nothing is built or connected before the trigger fires. The puzzle is to read each unit's visible power and position together, predict which existing unit can start a directed cascade that eventually covers the whole board, and trigger that source once. Because candidate units are processed in stable board-list order, the same trigger always produces the same tick groups, and a player who projects the wave correctly can predict the result before pressing the button.
How the Five-Field Run Reaches 1,000
The chain reaction game ships with five pre-built experiments on fixed 5-by-5 grids. Each experiment places several powered units on the board before you trigger anything. Every unit displays a power value from one to three, and that power is the maximum Manhattan distance of the unit's outgoing wave. The experiment ends either when every unit has activated or when no more uncleared units can be reached by any active unit.
A full clear of one experiment awards 200 points. Clearing all five experiments produces the exact maximum score of 1,000 points. There is no partial credit for an incomplete cascade. If a chosen source leaves active units behind, the run earns nothing on that experiment, and the failure is recorded. You cannot reduce or change this scoring. The field is populated, the rules are fixed, and the same trigger on the same field always produces the same tick groups. Independent tests lock the complete success set for each level, so the puzzle's correct sources form a finite, verifiable list rather than a hidden preference.
What Each Unit's Power Number Means
Each unit shows a small integer between one and three on the board. The number tells you how far that unit's wave can travel in a single frontier, measured as Manhattan distance. Manhattan distance counts only horizontal and vertical steps between two cells. A unit at row 2, column 2 and a unit at row 2, column 4 are exactly two steps apart because they share the same row. A unit at row 1, column 1 and a unit at row 3, column 3 are four steps apart: two down plus two across. Diagonal jumps do not count as a single step.
A unit with power one therefore reaches only adjacent uncleared units. A unit with power two reaches any uncleared unit within two combined horizontal and vertical steps. A unit with power three reaches anything within three combined steps. Distance can span empty grid cells, so a strong unit may skip over blank space to land on a distant target. Higher power does not always win, however, because a high-power unit stranded in a corner may reach only weak neighbors that also reach only weak neighbors.
| Power | Maximum Manhattan distance | What it reaches in one frontier |
|---|---|---|
| 1 | 1 step | Adjacent uncleared cells only |
| 2 | 2 combined steps | Any uncleared cell within a 2-step radius |
| 3 | 3 combined steps | Any uncleared cell within a 3-step radius |
Tick Groups and the Manhattan Frontier
The simulation uses discrete ticks rather than a continuous animation. The unit you select becomes tick zero and emits immediately. Any uncleared unit inside that source's Manhattan range activates on tick one. Tick one units then emit during tick two, and only the units activated on tick one form the new frontier. The process repeats one tick at a time.
Two rules govern which units activate next. A unit activates only the first time it is in range of an active frontier, and candidates are examined in stable board-list order, so the same trigger always produces the same tick groups. This is the part that makes the game a deterministic graph puzzle. There is no random spawn, no hidden probability, and no physics timing. The tick recorder lists every observed wave as T0, T1, T2, and later steps, while cleared units shrink and display the tick on which they reacted. These values are computed result data, not a decorative animation.
Multiple units activated on the same tick form one simultaneous frontier. Only that frontier emits during the next tick, which is why branches are valuable when they cover separate regions: one wide frontier can light up many units at once. A unit at the back of a long chain cannot skip ahead, because each wave only travels from the immediately previous frontier. The chain is therefore as slow as its narrowest relay.
Plan and Run a 1,000-Point Experiment
Here is the exact procedure to clear every unit on a single experiment and continue toward the full 1,000-point run:
- Read the board. Note each unit's grid position and its visible power number. Power is Manhattan range, so count horizontal plus vertical steps without diagonal shortcuts.
- Project tick zero. Decide which existing unit, if any, can reach every other unit eventually through a chain of frontiers.
- Select the source. Tap, click, or keyboard-focus that unit, then confirm the selection. Only one unit can be the source.
- Trigger the field. Press the trigger control. The source emits on tick zero and newly reached units activate together on the next discrete tick.
- Watch the tick recorder. Read the recorded T0, T1, T2, and later groups as cleared units shrink and label themselves with their tick number.
- Score the experiment. A full clear awards exactly 200 points and unlocks the next experiment. An incomplete cascade earns no points, but it can be repaired.
- Repeat across all five experiments. Five full clears reach the exact 1,000-point maximum.
This is the same workflow you can verify inside the Chain Reaction game. The board exposes only the product rules, unit positions, visible power values, and your current selection before a trigger runs. It does not show the correct source, a future tick list, or a hidden answer in visible text, accessibility labels, or data attributes. Tick groups appear after the trigger because they are part of the result, not a hint.
Reading the Tick Recorder When a Wave Stops
If the wave stops before every unit is cleared, the tick recorder tells you why. Look at the last recorded tick group. That group is the final frontier: the set of units that emitted during the last successful tick. Every uncleared unit on the board sits outside the Manhattan range of every member of that frontier. No further activation is possible from that frontier alone.
The fix is to choose a different source at the start. A successful source after one failure clears the failure record before the next level. Repeating the identical failed trigger is deduplicated, so a double press does not consume another mistake. A second genuinely different failed source closes the run for that level. The rules are deliberately lenient on careless double taps but strict on genuinely bad choices, which is why the game keeps the failed-trigger classifications out of the production level objects and verifies them only through an independent test oracle.
| Situation | Behavior | Notes |
|---|---|---|
| First failed source | Counts as one mistake | A different source can still be tried |
| Same failed source re-pressed | Deduplicated | Does not consume a second mistake |
| Second distinct failed source | Closes the run | The level ends with no points awarded |
| Successful source after a failure | Failure record cleared | Before moving to the next experiment |
| Full clear | +200 points | Five full clears total 1,000 exactly |
To read a failure quickly, ask three questions. Is there a remote low-power unit that no current frontier can reach? Could a different starting unit have lit up that endpoint earlier through a different relay chain? Is the chosen source itself a dead end, reaching only weak neighbors that also reach only weak neighbors? A practical repair strategy starts by reading power and position together. High power alone is not enough if the unit can reach only weak dead ends. Look for a source that reaches one or more useful relays.
Pick a Source That Actually Clears the Field
Before triggering, project the wave in frontiers. Mark the source as tick zero, identify every target inside its range for tick one, and repeat from only those newly reached targets. Branches are valuable when they cover separate regions. A remote low-power unit is often the hardest endpoint, so check which relay can reach it before committing. A source is good if its tick-one frontier contains at least one high-power unit, or if it spans two separate regions of the board. A source is bad if its tick-one frontier is small, its high-power neighbors are already out of range, or it leaves a remote low-power unit stranded at the back of the chain.
For a focused look at the first selection step, the trigger chain reaction walkthrough covers the projection in more detail. Keep the failed-trigger repair guide nearby when an experiment does not clear, since a single mistake can be repaired only once before a different choice closes the run.
Five Fields, Five Different Traps
The five experiments progress from a short three-unit relay to longer fuses, a corner-to-corner handoff, a four-way bloom, and a final multi-branch surge. The grid size never changes, and the rules never change. Difficulty grows through placement, power distribution, simultaneous waves, and directed reachability. The first experiment is a short three-unit relay, the shortest of the five experiments. The middle experiments test directed reachability, where a source must activate a specific relay rather than the nearest neighbor. The final experiment is a multi-branch surge, where one source must light up several distinct regions through several distinct relays on the same tick.
Independent tests recalculate each wave using separate code for every possible source, lock the complete success set for each level, and prove every remaining source stops early. So when an experiment resists you, the answer is a different source, not a different rule. Everything runs locally in the browser. Source choices, tick records, mistakes, progress, and scores are not uploaded. No account, subscription, extra package, or network calculation is required. The game is well suited as a compact reasoning break, a visual introduction to directed reachability, or a way to practice separating simultaneous state changes into clear discrete steps.