How This Game Dev Got 100 Million Views In 1 Day
TLDRThis conversation between Thomas Brush and Gavin Eisenbeis dives into the development journey of Choo Choo Charles, a viral horror game featuring a terrifying spider train. Gavin shares insights on idea generation, quick prototyping, asset utilization, and community building during the pre-production phase. He also provides details on using Unity, programming, testing, and launch preparation. Key highlights include developing good instincts through continual learning and releases, respecting team talents in a studio environment, and prioritizing fun over critical reception. The talk offers aspiring game developers an inside look at the creation process behind a successful title.
Takeaways
- ๐ Gavin started making games at 12 and recently celebrated 10 years in game development, focusing on PC and mobile platforms.
- ๐ Choo Choo Charles, Gavin's game, gained viral success across the internet with gameplay videos, trailers, and a developer vlog series on YouTube.
- ๐ The game exceeded Gavin's own expectations in terms of success, leading him to work on console porting and new game development.
- ๐ Utilizing Unreal Engine 4 (now 5), [__] for 2D art, and Blender for custom 3D models, Gavin emphasizes the importance of a unique art style and retexturing assets for game branding.
- ๐ Marketing and generating interest pre-launch were crucial, with an emphasis on creating a solid trailer and Steam page to gauge interest and gather wishlists.
- ๐ฅ The concept of Choo Choo Charles was based on the trend of turning nostalgic characters into horror elements, drawing inspiration from shows like Thomas the Tank Engine.
- ๐ Gavin's approach to game development includes quick prototyping, efficient use of assets, and focusing on unique gameplay elements like the spider train.
- ๐ฎ Emphasizing the importance of a game design document (GDD) and efficient pre-production to streamline development and focus on core game mechanics.
- ๐ฒ Sound design played a key role in the game's atmosphere, with most effects sourced from freesound.org and music composed by Tom Bellingham.
- ๐ Launch strategy highlighted the importance of building anticipation, managing expectations, and the utility of wish lists as an indicator of potential success.
Q & A
What game development tools and software did Gavin use to create Choo Choo Charles?
-Gavin used Unreal Engine 4 and blueprints for game development, Blender for 3D modeling, Krita for 2D art and textures, and Wondershare for audio editing. He also sourced sounds from Freesound.org.
How long did Gavin spend on pre-production and what key things did he define in that phase?
-Gavin spent about 2 weeks on pre-production. He defined core mechanics like trains, upgrades, guns, missions, spider train chase sequences etc. But left details like specific missions, AI behavior etc to figure out later.
What was Gavin's goal in the first month of Choo Choo Charles' development?
-His goal was to create a trailer and marketing assets to announce the game and launch a Steam page. This helped him gauge interest via wishlists early on.
How does Gavin create sound effects for the game?
-He records gameplay video first to match sounds with visuals. Then combines and edits sounds from Freesound.org, applying effects like pitch change, reverb etc. to match the video.
What is Gavin's perspective on game design documents?
-According to Gavin, having a GDD is essential for organizing ideas and documenting game mechanics even for small games. He considers it an important skill for any game developer.
Why does Gavin feel wishlist count is overhyped?
-He feels velocity and efficiency of building wishlists matters more than the raw number. Easy conversion indicates a marketable game idea even if total wishlists are less.
What special thing did Gavin do to get YouTuber 8bit Ryan to cover his game?
-He created a custom hand drawn portrait of Ryan as an in-game character. Seeing himself featured made Ryan more inclined to cover the game.
What is Gavin's perspective on game budgets and team sizes?
-He prefers staying solo because hiring even 5 people on a project for years involves significant cost and risk. As a Solo dev he has less to worry about in terms of budgets.
How important is the train coloring feature according to Gavin?
-Gavin considers it one of the best easy additive features that greatly increased appeal and sales due to customization. Half the sales were from people who just wanted to color the train.
What does Gavin attribute Choo Choo Charles' success to?
-A unique idea tailored for viral spread, efficient build-up of wishlists indicating demand, and a thoroughly polish build tested extensively before launch.
Outlines
๐ Introducing Gavin, creator of Choo Choo Charles
Thomas introduces Gavin, the creator of Choo Choo Charles. The game became a viral hit, gaining massive popularity on social media. Thomas praises Gavin's work and wants him to provide background on himself and the game's development.
๐ Discussing the game's concept and pre-production
Gavin explains the quick 2 week pre-production process. He had a solid idea for core mechanics like trains, upgrades, guns, and being chased by a spider train. He didn't plan every detail upfront but established enough of a foundation to create an announcement trailer and gauge interest.
๐ Building hype and launching on Steam
Gavin launched a trailer in October 2021 which generated lots of hype. He then focused on production, developing the systems and mechanics. A month before launch he locked down a solid build to focus purely on marketing. This included trailers, press releases, contacting influencers. The launch went smoothly with no major bugs and strong user reviews.
๐ Explaining development tools and process
Gavin used Unreal Engine, Blender, Photoshop, and FMOD. He took modular assets and customised them to create Choo Choo's unique art style. The open world was built from the terrain tools and assets from Unreal marketplace. Gavin programmed gameplay systems like train tracks using Blueprints.
๐ Creating engaging sounds and music
Gavin edited sound effects from Freesound.org in Audacity to match key moments of gameplay. He tailored licensed music tracks to fit the tone and atmosphere. Matching sounds to the visuals helped make the audio integration seamless.
๐ค Discussion on game marketing and wishlists
Gavin emphasizes wishlist count alone isn't enough. Need wishlists gained quickly and efficiently. Easy viral marketing indicates strong conversion potential. Big name YouTubers played the game immediately due to its viral nature. But smaller creators can be targeted based on similar games.
๐ Tips for getting YouTubers to cover your game
Gavin suggests involving YouTubers in development, like creating custom art of them for the game. This investment makes them more likely to play it. Building relationships over time through tweets and comments is key. Provide access and assets early in embargo to integrate into their pipeline.
๐ค Reflecting on critical reception
Despite strong user reviews, press outlets were very negative. Gavin speculates this clickbait tendency comes from subverting high expectations. As a silly meme game, critics inappropriate judged it by AAA standards. But ultimately fan reaction was most important.
๐ Enjoying success and planning next steps
After launch Gavin mostly took a break, processed the success, handled finances and taxes. Around 6 months later he was eager to start a new solo project using his proven formula. He still prefers working alone to avoid the risks and overhead of a team.
Mindmap
Keywords
๐กGame design document (GDD)
๐กViral marketing
๐กProof of concept
๐กWishlists
๐กInfluencer marketing
๐กLaunch preparations
๐กPress coverage
๐กPlayer feedback
๐กSolo development
๐กTool stack
Highlights
The use of convolutional neural networks for image classification represents a major breakthrough in computer vision.
Transfer learning allows pre-trained models to be fine-tuned for new tasks, saving computation time and improving performance.
Data augmentation techniques like cropping, padding, and horizontal flipping can help reduce overfitting on small datasets.
Activation functions like ReLU introduce non-linearities that help neural networks learn complex patterns.
Batch normalization accelerates training by normalizing layer inputs across each batch.
Recurrent neural networks like LSTMs are effective for processing sequential data like text and speech.
Attention mechanisms allow models to focus on relevant parts of the input when making predictions.
Generative adversarial networks can synthesize realistic examples like images and audio.
Vision transformers like ViT demonstrate top performance on image classification without convolution layers.
Sparse models require fewer parameters and computations than dense neural networks.
Quantum machine learning leverages quantum computing to potentially speed up deep learning.
Federated learning allows models to be trained on decentralized data located on various devices.
Causal inference techniques aim to determine cause-and-effect relationships from observational data.
Explainable AI methods help provide transparency into model predictions and behavior.
AI safety research seeks to ensure reliability, avoid harm, and align systems with human values.
Transcripts
Browse More Related Video
5.0 / 5 (0 votes)
Thanks for rating: