Match / Clear
Match 3+ identical tiles in a row to clear them; refills can cascade into chains. Minimal rules, maximal juice.
What is it
Line up three or more identical tiles and they vanish. You've definitely seen this — Candy Crush, Anipop, and Sheep a Sheep are all built on it.
How it plays
Each move you swap one tile to match three in a row; they clear and score, tiles above fall in to refill, and sometimes one clear cascades into another — the longer the chain, the better it feels.
Why learn it
The rule fits in one sentence, but "juice" and "replayability" are pure design craft — this is the best first lesson in game feedback design.
At its core, match-and-clear is a very short, trustworthy feedback loop: the player makes one swap and the rules immediately return a deterministic result. The hard part is never the clearing itself — it is the credibility of three things: invalid swaps must roll back, cascades must resolve in a fixed order, and refills must stay explainable. Player trust in the rules is built from exactly these invisible engineering details.
Why This Mechanic Matters
For Players
Match-and-clear answers a primal appetite for closure — turning a messy board into order, one small, certain, instant win at a time. Its rules are readable at a glance and its feedback is immediate, which makes it one of the rare mechanics that works across all ages and cultures.
For the Ecosystem
Match-3 is one of the most proven monetization models in casual games: near-zero onboarding widens the top of the funnel, level-based structure supports endless content expansion, and move/stamina economies interlock cleanly with rewarded ads. Candy Crush's decade-long reign is no accident.
Transferable Craft
Learn match-clear and you've learned the portable craft of feedback-curve design — reward multipliers, chain pacing, difficulty ramps — knobs you'll keep turning in numbers, combat, and idle games alike.
Learning Goals
- Beginner: understand the minimal match → clear → refill loop and reproduce one swap's full feedback in the demo
- Intermediate: write the complete resolution contract for a single swap (rollback, batched match collection, cascade order, gravity refill)
- Advanced: design deadlock detection and constrained reshuffles, and verify level health with quality metrics (legal-move count, chain depth, deadlock rate)
Core Patterns
② Instant Playable Demo
Match / Clear Demo
Loading game…
Rule Hints
Rule: swap/click → form ≥3 identical → eliminate → fall/replenish → chain (optional)
Lab Guide: treat the demo as an experiment
Chain Multiplier Experiment
Do: Drag "Chain Bonus Multiplier" from 1x up to 3x, then deliberately set up a second chain
Observe: The higher the multiplier, the more dramatic the score from chained matches. Watch whether players start leaving an obvious match untouched just to wait for a bigger chain — that's a sign the reward is quietly changing how people play.
Move Budget Experiment
Do: Squeeze "Moves per run" from 30 down to 8
Observe: Fewer moves means every move needs more thought. Compare the two settings: does the 8-move version make players think more carefully, or does it just make them lose faster? If it's the latter, it's simply too few moves, not a more interesting challenge.
Deadlock Watch
Do: Keep swapping until no valid move remains on the board
Observe: This demo lets the run get stuck in a deadlock on purpose; a real game must detect this automatically and reshuffle the board — players should never be left staring at a board with nothing they can do.
Replay Determinism Experiment
Do: Pick the same starting board and the same swap sequence, play it through twice, and log what each move cleared, how pieces fell, and the score (toggle "Fixed Board" in the demo to try this hands-on)
Observe: Both runs should match exactly. If they don't, it's usually because the code deletes while still scanning, or the refill's randomness isn't fixed — this is the foundation for a match game that keeps its promises, and for safely tuning it later.
Legal-Move Quality Experiment
Do: Once the board settles, find every adjacent swap that would match, and label each as "score only," "advances the objective," or "spawns a special tile" (toggle "Highlight Legal Moves" in the demo to try this hands-on)
Observe: Having a move available doesn't mean the board is interesting. If every legal move pays off about the same, players are executing rather than choosing; if only one move is clearly right, it turns into an answer hunt. The best boards mix high- and low-value moves so players can judge for themselves.
③ Behavior System Breakdown
3.1 Problems Solved
- Every move gives an instant result: matched pieces disappear, unmatched pieces don't — easy to understand even on your first try.
- Matching just 3 gets boring fast, so matching 4 or 5 pays off more, encouraging players to plan ahead instead of matching whatever is closest.
- Blockers and small goals on the board give players a reason to think about which move is worth making, instead of tapping at random.
- If the board ever has no possible match left, the game automatically reshuffles so the player is never stuck.
3.2 Minimal Ruleset
- Board and pieces: start with something simple, like a 6×6 board filled with a few colors or shapes. Three or more of the same piece in a row or column counts as a match.
- How to play: players can only swap two pieces that are next to each other. If the swap doesn't create a match, it swaps back automatically, as if nothing happened — the game never tricks the player.
- How matches clear: after a swap, find every match this move created first, then clear them all together, instead of clearing them one at a time — otherwise the result looks inconsistent.
- Falling and refilling: cleared pieces leave gaps that the pieces above fall into, and new pieces drop in from the top; then check whether the new layout happens to match again (a chain).
- What happens when stuck: if the board has no possible match left, the game must reshuffle the pieces automatically so the player always has a move to make.
3.3 System Loop (optional)
Player picks a move (swap two adjacent pieces) → check whether it creates a match of 3+ → if yes, clear it, let pieces fall, refill, and check for a chain → if no, treat it as if nothing happened → once the board settles, check whether any move is still possible, and reshuffle if not → hand control back to the player.
Design Notes
Juice comes from the curve, not the animation: chain score = base × multiplier^(chain−1), so the second chain is worth far more than the first — animation is just the packaging for that curve.
Difficulty must come from readable boards: move budgets, blockers, and color count are all knobs, but the premise of any difficulty is that the player can see what went wrong. Silently removing legal moves in the backend is the cheapest difficulty — and the one that kills retention.
A deadlock is not a failure; it is a teaching moment: detect it, hint a legal move, and reshuffle under constraints. These three decide whether a stuck player learns something or churns.
④ Common Combos & Variants
Match / Clear + Progression
Turn points earned from matches into upgrades or unlocks, giving players a long-term goal to work toward instead of just a quick thrill.
Match / Clear + Combat
Turn matched colors or match size into attack or healing effects, so clearing matches also means fighting — blending the two into one activity.
Match / Clear + Random
Make each playthrough's board a little different for variety, while still making sure players always have a way to win instead of relying purely on luck.
⑤ Advanced Design & Algorithms (collapsed)Click to expand
Design Warnings
- If chain rewards are too generous, players start hoping for lucky big combos instead of carefully thinking through each move: check whether players are still advancing the level objective once big chains show up — if the two keep drifting apart, the reward is winning over the strategy.
- If a level gives too few moves or too little time, players won't have time to think and will end up tapping randomly: don't judge only by clear rate — look at failed boards for a move that was actually better but missed. If the real problem is 'no time to read the board,' the issue is pacing, not difficulty.
- If new pieces are secretly stacked against the player, they'll feel like "the game is out to get me": it's fine to quietly help after a run of bad luck, but the board must never stay unsolvable, and it must never clear itself while the player does nothing.
- When several special effects happen at once, decide ahead of time which one goes first: spawn position, then whether it chains into another special tile, then area effects last — otherwise the same move can produce different results at different times, making the rules feel unreliable.
- Lowering chain rewards a little can sometimes make play more varied: once players must balance objectives, blockers, and board shape together, chains stop being the one correct answer and become just one option among several — whether to lower it should depend on whether players still advance objectives and use special tiles.
Algorithm Examples
- How to find every possible match on the whole board in one pass, without missing any or checking the same spot twice: a common approach is to only re-check the two rows and columns touched by a swap, merge the hit cells into one set, then classify that set's shape as a straight line, an L/T shape, or a crossing match.
- How to tell when the board is "stuck" (no possible match left): try a temporary swap between each cell and its right and bottom neighbor and see if it would match; if any pair would, there's still a move — the same check also powers hint generation.
- How to quietly reshuffle the board when it's stuck: shuffle only the regular pieces, keep fixed level objects and locked cells in place, then verify the new layout has no immediate match and still has at least one legal move before using it.
- Keep the board's data separate from what's drawn on screen: store each cell's color, whether it's special, and whether it's locked in one place, rather than reading the rendered picture — this is what makes reshuffling, replay, and testing reliable.
- One match can trigger several rounds in a row (a chain): work out what each round clears, how many points it adds, and whether it spawns a special tile before updating the board, instead of changing the board bit by bit — that keeps the outcome from depending on processing order.
- If you want to seriously tune this ruleset, track over time: legal-move count, deadlock rate, invalid-swap rate, average chain depth, and how often 4/5-matches occur — compare settings on the same level to see whether a change actually helped.
⑥ Related Case Articles
Match-3 Mechanics: Crafting Satisfaction via Score Curves and Constraint Strategies
This guide dissects the non-linear mapping between chain length and score in match-3 games, detailing technical implementations for blocking infinite chains through dynamic generation and movement constraints. It further explains how to calibrate feedback timing to sustain player engagement, equipping developers with tools for numerical balancing and level pacing.
Match-3: Leveraging Elimination Mechanics to Drive Retention and Pacing
A deep dive into Match-3 core mechanics, focusing on the intensity stratification logic for 3/4/5-match eliminations. Through quantitative analysis of satisfying elimination triggers and dynamic checkpoint control strategies, we reveal how to drive user retention via rule design. Content covers constrained random generation algorithms and level pacing, aiming to help developers build high-engagement elimination experiences and improve game balance.
Number Merge (2048 Variant): Failure Probability Control and Perceived Randomness
Deep dive into the core mechanics of number merge variants: dynamically adjusting failure probability thresholds to softly control difficulty, avoiding unwinnable states while maintaining tension. The article details the pseudo-random generation algorithm logic, ensuring number distribution appears random yet remains controlled, precisely balancing player experience with numerical stability to help developers build high-retention elimination levels.