Chain Reaction is a five-level trigger puzzle where you select exactly one source unit on a populated five-by-five board and watch a deterministic wave spread through nearby powered units on discrete ticks until every cell clears or the propagation stops. The rules are explicit and short: each unit shows a power value from one to three, and that number equals its maximum Manhattan range — the count of horizontal plus vertical steps it can reach with no diagonal shortcuts. Your chosen source emits on tick zero, every uncleared unit inside its range activates together on the following tick, and every newly activated unit emits its own wave one tick later. Only the frontier of units activated on the immediately previous tick emits during each new tick, so the wave advances in clearly labeled groups the game records as T0, T1, T2, and onward. Clearing every unit on a level awards exactly 200 points, and clearing all five levels produces the exact maximum score of 1,000. There is no placement phase, no domino stacking, and no random element — one selection launches the entire experiment, and the same trigger always produces the same tick groups.
How to Play Chain Reaction
Every field follows the same three-step loop. The grid size, the range rule, and the scoring never change across the five experiments; difficulty rises through placement and power distribution, not through new mechanics.
- Inspect every unit's grid position and visible power. Power is Manhattan range, so count horizontal plus vertical steps without diagonal shortcuts.
- Select one source unit by touch, click, or keyboard, then trigger the field. Newly reached units emit together on the next discrete tick.
- Study the recorded tick groups if the wave stops, choose a different source, and try again. Clear five fields at 200 points each for exactly 1,000.
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. Each board begins populated, so your only decision is which existing unit becomes the source.
Power, Manhattan Range, and Distance Counting
Each unit's power value is the single most important number on the board, because it tells you exactly how far that unit can send its wave. Power is integer-valued from 1 to 3, and it represents Manhattan distance — the number of horizontal steps plus vertical steps between two cells. A source with power 2 reaches any uncleared cell at distance 2 or less; a power-3 unit reaches anything at distance 3 or less. Diagonals are not shortcuts you can buy cheaply: from (1,1) to (3,3), the Manhattan distance is |3-1| + |3-1| = 2 + 2 = 4 steps, which a power-3 source cannot cover directly. The same Manhattan count lets a wave cross empty cells, so a power-3 unit sitting in one corner can still activate an uncleared unit three orthogonal steps away regardless of which cells sit on that path.
| Power value | Maximum Manhattan range | Crosses empty cells | Reaches the four orthogonal neighbors |
|---|---|---|---|
| 1 | 1 step | Yes | Yes |
| 2 | 2 steps | Yes | Yes |
| 3 | 3 steps | Yes | Yes |
Candidates inside a frontier are examined in stable board-list order, so the same trigger always produces the same tick groups. A unit activates exactly once, even if several neighbors reach it on overlapping ticks, and that activation belongs to the earlier of those ticks. Reading power and position together is the only preparation the puzzle ever asks for; once you commit to a source, the board returns the result as computed tick data, not as an animation.
Tick Groups and the Simultaneous Frontier
The wave does not ripple continuously. It advances one discrete step at a time, and the units that fire on each step form a single group called a tick. Tick zero contains only the source you selected. Tick one contains every uncleared unit within that source's Manhattan range. Tick two contains every uncleared unit within the range of any unit that activated on tick one, and so on. The crucial rule is that only the immediately previous frontier emits during the next tick — units that activated two or more ticks ago do not re-fire. This is why the same trigger on the same board always produces the same tick groups, and why the Tick recorder can list the wave as T0, T1, T2, and later steps with no ambiguity.
Multiple units can activate on the same tick and form one simultaneous frontier. A power-3 source can therefore reach several targets at once, and those targets become a single tick group rather than a chain of individual steps. Branches matter: when a tick-one frontier covers two separate regions of the board, the wave keeps advancing in both regions at once on tick two, which often clears the field in far fewer ticks than a single straight line would suggest. The Tick recorder exposes these groups as computed result data, so a stalled run tells you exactly which frontier reached its last uncleared cell and which frontier failed to advance further.
The propagation ends in one of two ways. If every unit activates, the field clears and you earn the full 200 points. If the previous frontier has no uncleared neighbors inside its combined range, the wave stops; cleared units shrink and display the tick on which they reacted, and you can read those numbers to see exactly where the chain ended.
Scoring, Failure Repair, and the 1,000-Point Goal
Scoring is fixed and visible. Every cleared field awards exactly 200 points, and five cleared fields sum to 1,000 points with no bonuses, multipliers, or hidden extras. A trigger that leaves active units earns no points for that attempt, but the run is not over. One failed source can be replaced by a different source, and the field restarts cleanly. Pressing the same failed trigger twice is deduplicated — the system counts it as one mistake, not two — so a double press or accidental repeat does not consume your repair. A second genuinely different failed source, however, ends the run on that field: the game treats that as a final deadlock, and you must restart from the beginning.
When a corrected source succeeds after one failure, the failure record is cleared before the next level opens. This means you can take one calculated risk per field, learn from the tick groups it produced, and still finish the level cleanly. The repair mechanic exists to make the puzzle about reasoning rather than punishment: every restart, whether through the Restart button or by triggering a new source, leaves the board in the same state it began in, so a fresh inspection always starts from the original placement. For a closer look at how the repair mechanic plays out across the harder fields, the guide Can You Fix a Failed Chain Reaction Trigger? walks through specific repair scenarios on the multi-branch levels.
Controls, Board Layout, and Accessibility
The board itself never resizes and the controls never change, which makes keyboard and touch play straightforward to learn. The field uses a fixed five-by-five grid that stays within a 390-pixel viewport, and every unit control is at least 44 pixels high so touch and pointer taps remain comfortable. Status updates are announced with live text, so the current selection and tick groups stay readable without depending on animation. Pointer and touch play use native unit buttons: tap a unit to select it, then press the Trigger selected unit button. Keyboard play is scoped to the reaction field — focus a unit, use the arrow keys to move through the available units, press Space to select, and press Enter to trigger. Restart, Trigger, and Next experiment retain their normal button behavior when they hold focus, so the full run is playable without a mouse.
Before a trigger runs, the page exposes only the product rules: unit positions, visible power values, and your current selection. It does not place a solution index, correct-source flag, future tick list, hidden answer, or success marker in visible text, accessibility labels, hidden fields, or data attributes. After the simulation, observed tick groups appear because they are now part of the result; before the simulation, the page stays strictly neutral. Independent tests verify the complete successful and failed trigger classifications out of band, so the only way to read the answer is to run the wave.
Reading the Board Before You Trigger
Because the board is deterministic, the most useful skill is reading power and position together before you commit to a source. High power alone is not enough: a power-3 unit sitting alone in the corner may only reach weak dead ends. Look for a source that reaches one or more useful relays — units whose own range covers a separate region of the board. Then imagine the reaction in frontiers: mark the source as tick zero, identify every uncleared target inside its range for tick one, and repeat the process from only those newly reached targets on tick two. Branches are valuable when they cover separate regions, because a single straight chain usually stalls before it reaches the far side of a populated board.
A remote low-power unit is often the hardest endpoint on the level, so check which relay can reach it before committing to the trigger. If no relay on the field can activate that endpoint, consider selecting it as the source itself — the source emits on tick zero, so a hard-to-reach unit clears immediately when triggered directly. 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, and the grid size and rules never change across them. Difficulty grows through placement, power distribution, simultaneous waves, and directed reachability, not through new mechanics.
Source choices, tick records, mistakes, progress, and scores are stored only in your browser; nothing is uploaded, and no account, subscription, or network calculation is required to play the full five-level route to Chain Reaction. The simulation uses explicit deterministic rules with no random spawn, hidden probability, physics timing, or server response, so every score you record reflects your read of the board rather than luck.
For a deeper look, see Cocktail Mixing Game Rules: Controls, Scoring, Miss Logic.