Recreate Flappy Bird’s flight mechanic | Wireframe #29

From last year’s issue 29 of Wireframe magazine: learn how to create your own version of the simple yet addictive side-scroller Flappy Bird. Raspberry Pi’s Rik Cross shows you how.

Flappy Bird: ridiculously big in 2014, at least for a while.

Flappy Bird was released by programmer Dong Nguyen in 2013, and made use of a straightforward game mechanic to create an addictive hit. Tapping the screen provided ‘lift’ to the main character, which is used strategically to navigate through a series of moving pipes. A point is scored for each pipe successfully passed. The idea proved so addictive that Nguyen eventually regretted his creation and removed it from the Google and Apple app stores. In this article, I’ll show you how to recreate this simple yet time-consuming game, using Python and Pygame Zero.

The player’s motion is very similar to that employed in a standard platformer: falling down towards the bottom of the screen under gravity. See the article, Super Mario-style jumping physics in Wireframe #7 for more on creating this type of movement. Pressing a button (in our case, the SPACE bar) gives the player some upward thrust by setting its velocity to a negative value (i.e. upwards) larger than the value of gravity acting downwards. I’ve adapted and used two different images for the sprite (made by Imaginary Perception and available on opengameart.org), so that it looks like it’s flapping its wings to generate lift and move upwards.

Pressing the SPACE bar gives the bird ‘lift’ against gravity, allowing it to navigate through moving pipes.

Sets of pipes are set equally spaced apart horizontally, and move towards the player slowly each frame of the game. These pipes are stored as two lists of rectangles, top_pipes and bottom_pipes, so that the player can attempt to fly through gaps between the top and bottom pipes. Once a pipe in the top_pipes list reaches the left side of the screen past the player’s position, a score is incremented and the top and corresponding bottom pipes are removed from their respective lists. A new set of pipes is created at the right edge of the screen, creating a continuous challenge for the player. The y-position of the gap between each newly created pair of pipes is decided randomly (between minimum and maximum limits), which is used to calculate the position and height of the new pipes.

The game stops and a ‘Game over’ message appears if the player collides with either a pipe or the ground. The collision detection in the game uses the player.colliderect() method, which checks whether two rectangles overlap. As the player sprite isn’t exactly rectangular, it means that the collision detection isn’t pixel-perfect, and improvements could be made by using a different approach. Changing the values for GRAVITY, PIPE_GAP, PIPE_SPEED, and player.flap_velocity through a process of trial and error will result in a game that has just the right amount of frustration! You could even change these values as the player’s score increases, to add another layer of challenge.

Here’s Rik’s code, which gets an homage to Flappy Bird running in Python. To get it working on your system, you’ll first need to install Pygame Zero. And to download the full code, go here.

If you’d like to read older issues of Wireframe magazine, you can find the complete back catalogue as free PDF downloads.

The latest issue of Wireframe is available in print to buy online from the Raspberry Pi Press store, with older physical issues heavily discounted too. You can also find Wireframe at local newsagents, but we should all be staying home as much as possible right now, so why not get your copy online and save yourself the trip?

Make sure to follow Wireframe on Twitter and Facebook for updates and exclusive offers and giveaways. And subscribe on the Wireframe website to save up to 49% compared to newsstand pricing!

6 comments
Jump to the comment form

Avatar

Also check out the Raspberry Pi learning to play FlAppy Bird using AI – https://www.youtube.com/watch?v=QfOzZpMFFl8&safe=active

Reply to Dan Aldred

Avatar

Never heard of NEAT until I came across this version of flappy bird.
https://github.com/techwithtim/NEAT-Flappy-Bird
I was surprised at how fast it learned.
Gives some idea of how simple collision avoidance is in Tesla Autopilots? It should learn to play Bunny/Frogger better than me?

Reply to Gavin McIntosh

Avatar

Really awesome…Reminds me of my video game days…

Reply to Ishan Thapar

Avatar

Really amazing … It reminds me of my video game days …

Reply to empleosgt

Avatar

Nice one. I always love flappy birds

Reply to Robert

Avatar

oh my goddness, great

Reply to mobilbahis

Leave a Comment