Tyrbo Make your game

How to Make a Game Like Candy Crush Saga, Without Copying It

Tyrbo · · 7 min read

Candy Crush Saga, from King, did not invent matching three tiles. What it got right was everything around the swap: a move budget on every level, a goal that changes from one level to the next, special pieces that reward bigger matches, and a long winding map that always shows the next stop. Those ideas now shape most mobile puzzle games.

Below is how that structure works from a designer's side, the parts a level-based match-3 needs, and how to plan, build and tune your own in Tyrbo. You will finish with a puzzle game that has its own theme and its own rules, which is the only kind worth publishing.

  • Match-3 puzzle game concept art in the bright and bold lookMatch-3 puzzle · Bright and bold
  • Match-3 puzzle game concept art in the soft pastel lookMatch-3 puzzle · Soft pastel
  • Match-3 puzzle game concept art in the candy cute lookMatch-3 puzzle · Candy cute
  • Match-3 puzzle game concept art in the cozy cartoon lookMatch-3 puzzle · Cozy cartoon
  • Match-3 puzzle game concept art in the 16-bit pixel lookMatch-3 puzzle · 16-bit pixel
  • Match-3 puzzle game concept art in the neon night lookMatch-3 puzzle · Neon night
  • Match-3 puzzle game concept art in the paper craft lookMatch-3 puzzle · Paper craft
  • Match-3 puzzle game concept art in the anime bright lookMatch-3 puzzle · Anime bright
Concept art Tyrbo made for match-3 puzzle games. None of it is from Candy Crush Saga.

What makes Candy Crush Saga work

The loop is short: look at the board, find a swap, watch it resolve, check the goal, repeat until the moves run out. The reason people keep playing is that each level is a small, different puzzle. Five things make that happen.

The move limit turns a toy into a puzzle

Swapping tiles with no limit is relaxing and aimless. Give the player twenty moves and every swap becomes a decision: the obvious match now, or the setup that makes a special piece next turn. The budget is the difficulty dial. One move more or less can turn a level from frustrating to fair.

Goals that change every level

Some levels ask for a score. Others ask you to clear a layer hidden under the tiles, bring items down to the bottom, or collect a number of one color. Each goal makes the player read the same board differently. A score level rewards big combos anywhere; a drop-item level makes the columns under the item the only ones that matter.

Special pieces from bigger matches

Match four and you get a piece that clears a line. Match in an L or T and you get a blast. Match five in a row and you get a piece that removes a whole color. Swap two specials together and something bigger happens again. This is where skill lives: strong players spend moves building specials rather than taking the first match in sight.

Blockers and board shapes

Not every board is a rectangle. Holes, odd outlines, tiles locked in place, layers that need two hits, things that spread if you leave them alone: each is a new problem that arrives on its own level and then mixes with the others. Blockers are how a level designer gets hundreds of distinct puzzles out of one rule set.

The map and the lives

Levels sit along a path, and the path is a reward in itself, because the next stop is always visible. Lives refill over time, which paces sessions and gives the game a reason to be opened again later. That pacing system is a design choice of its own; you can keep it, soften it or drop it.

The parts a game like it needs

Systems

  • A grid board with swap, match detection, gravity and refill, plus the rule that a new board never opens with a match and always offers a legal move.
  • Cascade resolution in clean steps, with input locked until the board settles.
  • Special pieces and their combinations, written as a table: which match shape makes which piece, and what each pair does when swapped together.
  • Goal types: score, collect, clear a layer, drop items, each with its own counter on screen.
  • A move counter, stars and a fail screen that offers a retry.
  • A level map that remembers progress.
  • Boosters used before or during a level.

Content

The content of a match-3 is its levels, and they are the expensive part. Each one is a board shape, a starting layout of blockers, a goal, a move budget and a color count. Plan the first stretch carefully and treat every later level as a small design job, not filler.

Tuning

  • Color count. The single strongest difficulty lever. Fewer colors mean more cascades and easier levels.
  • Move budgets. Set them generous, play, then tighten one at a time.
  • The ramp. One new idea per level, a breather after every hard one, and wins on most of the early stages.
  • Luck. Some randomness is the joy of cascades. Too much, and players feel the game decided for them. Watch for levels won or lost on the opening board.

The match-3 build guide has starting numbers for a first ten levels if you want a draft to play against.

Making your version in Tyrbo

Start with a sentence that has a twist

A theme alone is a reskin. A theme plus one rule change is a game. Three prompts that already go somewhere else:

  1. "A match-3 in a flooded museum where every match drains water, and the level is won when a sunken statue is fully uncovered."
  2. "A match-3 under a night sky where you match stars into constellations, and a finished constellation becomes a special piece you can aim."
  3. "A match-3 on a vegetable allotment where tiles you leave alone slowly grow into bigger, more valuable versions, so waiting is sometimes the best move."

What Tyrbo asks

Expect a handful of questions, each with a pre-selected answer: board size, how many tile colors, which goal types to start with, how many levels in the first plan, whether to use lives. If you have no strong view, keep the suggestion. The one worth real thought is the goal mix, because it decides what your levels will feel like.

Reading the plan

The plan lists every part of the game with a title, a pitch and concept art, before any of it exists, and seeing it takes only a free sign-in. Check that it names your special pieces, your blockers and the goal types, and that the first levels bring them in one at a time. Changing a line here costs nothing.

Build, play, tune

Building is on a paid plan. Your board starts from Tyrbo's match-3 starter on its tested grid board, so swaps, gravity and refills behave from the first build, and your twist is written as code owned by your game. Scan the code on screen and the build opens on your phone.

Then shape it in plain sentences:

  • "Level 6 is won on the first board too often. Add a sixth color."
  • "The drain is too slow to notice. Lower the water visibly with each match."
  • "Players never build constellations. Show a hint after eight seconds."

Ask Tyrbo for a test panel inside the game: jump to any level, refill moves, change the color count live. Tuning levels means replaying levels, so anything that shortens the replay is worth asking for. The match-3 game maker page shows the looks you can give your tiles, and the puzzle game maker covers the wider category.

Making it yours, not a copy

Plenty of match-3 games share the swap. What King owns is Candy Crush Saga itself: its name (and the word "Crush" used to suggest it), the logo, the candy tile designs, the characters, the music and sounds, the text, and the exact layouts of its levels. Use none of them. Rebuilding a level grid tile for tile is still copying, even with your own art laid over it.

The easiest trap in this genre is the candy theme itself. Bright sweets on a board will read as an imitation to players and to store reviewers, so pick a world that gives your tiles a reason to exist. Then change at least one rule so your game plays differently:

  • Tiles that grow, rot or freeze over time.
  • A board that scrolls upward as you clear it.
  • Specials the player builds on purpose from a recipe, not by luck.
  • No lives at all, and a softer move limit that ends a level with a partial reward.

Google Play's policies turn away apps that impersonate other apps or infringe intellectual property, and a puzzle listing that echoes a famous one in name, icon or screenshots is exactly what that rule is aimed at. Matching, move limits and special pieces are mechanics, and mechanics are generally free to build on. This is not legal advice; for a specific question, ask a lawyer.

Publishing it

When the first levels feel right, Tyrbo gives you a signed Android build you own, commercial rights included, wired for ads and in-app purchases. Extra moves after a failed level and booster packs are the usual places for them in a puzzle game; how generous to be is your call. Follow the step-by-step checklist to put it in the store yourself, and read how to publish a game on Google Play for every stage.

Questions people ask

Can I legally make a game like Candy Crush?

Matching tiles, move limits, level goals and special pieces are used by many puzzle games and are generally free to build on. King's name, art, characters, music, text and level layouts are not, so make each of those your own. This is general information, not legal advice.

How long does it take to make a game like Candy Crush?

There is no single answer; it depends on how many levels you want and how carefully you tune them. The board and its rules start from tested parts, so most of your time goes into level design and playing each level until it feels fair.

Do I need to know how to code to make a match-3 game?

No. You describe the tiles, the goals and your twist, confirm a few pre-filled choices and read the plan before building. Moves, colors, specials and levels all change when you ask in plain words.

Can I add boosters and extra moves to my match-3?

Yes. Boosters can be earned, bought or both, and extra moves after a failed level are a common rewarded ad. Every build comes wired for ads and in-app purchases, so what to offer and how often is the design choice left to you.

Make the game this guide is about.

Make your game

Sign in to get the plan and its concept art. Both are free.

Candy Crush Saga is a trademark of its owner. Tyrbo is not affiliated with or endorsed by them; the name is used only to describe and compare.