Transcription

Your support is greatly appreciated! https://www.patreon.com/8blit

S06E02 I Built & Published My First ATARI 2600 Homebrew

INTRODUCTION

If you’ve been following the channel for a while, you might know I’ve spent the last year and a half working on a little passion project. Well, after a ton of coding, testing, and bug fixes - so many bug fixes - it’s finally finished. I’d like to introduce my brand-new game: Tic-Tac-Toe for the Atari 2600. It’s a two-player game with a single-player mode, featuring 16 different ways to play, making it, technically, the first fun Tic-Tac-Toe game on the platform. In this episode, I’ll show you how I accidentally brought this from concept to reality, with a little help from some friends. Pitter patter.

CONCEPT

The game actually started life as example code for a previous video, where I covered how to program the CX-50 and other Atari-compatible keyboard controllers. You can check that out in the episode titled “Atari 2600 Keyboard Controllers Explained”, and I’ll drop a link to it in the description below.

In that early build, pressing one of the top nine buttons on the keyboard would place an X in the corresponding spot on the board. However, once I got the basic board working, I realized it wasn’t actually a great example, mainly because we never touched the bottom three buttons. Plus, building a Tic-Tac-Toe board you couldn’t even play felt a little lame. So instead, I decided to flesh it out into a fully playable game. I ended up replacing the example code for that video with something else… which, in my opinion, turned out to be a much better showcase for CX-50 programming anyway.

BUILDING THE GAME

Early development was strictly two-player, using keyboard controllers. You played a single game, took turns placing your marks until it ended, win, lose, or draw, and then it just restarted. Honestly? It was pretty boring. So I started brainstorming ways to make it more exciting, which led to a whole cascade of new features.

It was a bit of a domino effect. As I added one feature, I’d get an idea for another. Ultimately, that led to four unique gameplay modifiers. You can choose a Single Game, a Best-of-Ten Set, a Timed Mode, or an Obscured Mode. And the best part? You can mix and match them. Just using the two-player mode gives you eight different ways to play right off the bat. You can play a standard, relaxed game with no time limit… or a best-of-ten set where each move has a timer that gets progressively faster, and neither player can see their marks on the board. It gets crazy fast.

As I was wrapping up the gameplay mechanics, it hit me: not everyone has a second player sitting next to them. So I built a robot friend for you to play against. Our new friend has a basic AI that’s sometimes clever, and sometimes… a little dumb. It knows some solid counter-moves, providing a decent challenge for non-Tic-Tac-Toe grandmasters, especially if you’re playing a timed, obscured set! With the addition of the Robot AI, the game instantly jumps from 8 to 16 different play modes.

But wait, there’s more! To make the timed modes even more challenging - or a little easier - I hooked up the console’s difficulty switches to control the turn speed for each player. Now, if you’re playing against someone who’s struggling, you can flip their switch to give them a little extra time to think about their move.

At this point, the game was feature-complete. Except for one major detail: not a lot of people actually own the CX-50 or compatible keyboard controllers. So, I went back to the drawing board and revised the input logic to support standard joystick controllers. Instead of pressing a key that corresponds to a grid position, you move the joystick toward that spot on the board and press the fire button. Now, if you’ve watched this far and you’re disappointed that it uses joysticks instead of the original keyboards, I’ve got great news for you. When you buy a physical copy, you’ll actually receive the ROMs for both versions: the Joystick Edition and the Keyboard Edition!

TESTING

After adding the sound effects and music, I sent the ROM out to some friends for QA testing. The feedback was really positive, even though we found several bugs, mostly video glitches and unstable screens during gameplay transitions. We also discovered that a player could beat the Robot AI every single time if they placed their first two marks in specific spots. So, I added an extra counter-move and reworked how randomization factors into the Robot AI’s decision-making process. It took several rounds of testing to get it as polished as it is now, but those minor fixes made a world of difference. Don’t skimp on your QA!

FINDING A PUBLISHER

In my episode titled “There and Back Again”, I chronicle my journey of programming my own Atari 2600 game, from getting my first Atari VCS for Christmas in the early 1980s, to hand-entering ones and zeros from a book into an Atari 800XL, all the way to creating 8Blit, the channel you’re watching right now. Throughout all of that, I never actually finished a game. Until now, of course. If you’d like to check that out, I’ll leave a link in the description below.

At first, I really only thought about AtariAge for publishing, since I’d used their forums in the past. It seemed like the logical path. However, it didn’t pan out. After getting ghosted, I wasn’t feeling very hopeful about getting a physical release. I looked into self-publishing, but even with the support of my patrons, I couldn’t afford the manufacturing costs for the box, manual, label, cartridge, chips, and circuit board. Nor did I have the time for all the labor - cutting and assembling boxes, burning chips, soldering boards, packaging, and shipping.

That’s when I remembered a message I’d received a year earlier. Matt Shea, the owner of IdiotBox Games, had reached out to ask about publishing a game. At the time, I didn’t know much about IdiotBox Games, but over the years, the name kept popping up, and I was especially intrigued by his game, Spice Harvester. On the 8Blit Discord, we actually have two members who have also published with IdiotBox: Roy Riggs, who made Tangleweed Joe and Kaiju, and Kameleon70, creator of Destroyer 2600. I’ll drop links to both of their games in the description below. So, with all that in mind, I reached back out to Matt to see if he’d be interested in publishing my Tic-Tac-Toe game for the Atari 2600.

WORKING WITH THE PUBLISHER

Let me start by saying that working with Matt has been an absolute pleasure. He’s really passionate about the Atari 2600 and publishing physical games. First, I sent Matt the latest ROM builds for him to try out, and it wasn’t long before he responded, super psyched to work together.

Over the next couple of months, Matt did his own playtesting and found a couple more bugs, including another transition that briefly rolled the screen. Fixing the code would’ve been easy… if I actually had space left in my 4K ROM. Out of 4,096 bytes, I had exactly seven left, which is basically nothing. This required me to get a little creative and refactor the code to make some room. Eventually, I managed to free up around 300 bytes for my patches, with enough left over to add even more music.

While I was working on that, Matt was searching for an artist to create the box art and reached out to Stephen Schisler. After looking at his website, I knew Stephen would be perfect. I feel like his art style really captures how I remember early ’80s pop culture. I’ll link to his website in the description below.

A few months later, Matt sent over the finished box artwork, and I was right - I loved it. It fit in perfectly with the game and the Robot AI. After a couple more months, I received a package from Matt containing a complete-in-box copy of the game: manual, stickers, and all… along with 20 uncut printed boxes for me to sign for early pre-orders. After practicing my signature a couple hundred times before committing it to the boxes, I carefully made my way through the stack. All that was left for me to do was repackage everything and ship it over to Stephen so he could sign them as well.

Thankfully, Matt reimbursed me for the customs duties when I got the package, as well as the cost to ship it along to Stephen. Overall, I’m really thankful Matt took on publishing the game. It’s a little nerve-wracking to relinquish control and trust someone else with something you’ve been working on for over a year… but it was definitely the right decision.

ZERO PAGE HOMEBREW

One of the greatest, and also most nerve-wracking, experiences is watching a game you’ve labored over for so long get played live online for the very first time. On July 31st, ZeroPage Homebrew streamed a live unboxing of several new IdiotBox releases, including the world premiere of Tic-Tac-Toe.

For those of you unfamiliar with the channel, ZeroPage Homebrew showcases new independently created games, primarily for Atari systems. Hosted by James, they feature live gameplay broadcasts, developer interviews, and the annual Atari Homebrew Awards. I was super psyched when I learned my game would be featured on the show, so I made sure to be in the chat just in case they ran into any issues.

Overall, I think they really enjoyed playing my version of Tic-Tac-Toe. With all the different play modes, there’s a little something for everyone at a party to keep the action going. I’ll leave a link to their stream in the description below.

GAME SOURCE RELEASE

Several months ago, I promised I would release the source code for the game… and that’s a promise I’m keeping. The 8Blit channel is all about helping others learn how to program Atari 2600 games, and I think a fantastic way to learn is by having full access to the source code of a complete commercial game. In fact, this really is the next logical step. Every single technique you’ve learned in our previous episodes was used to make this game. There’s almost nothing in it that hasn’t been featured in a video at some point.

Not only will the source code be freely available, but it’ll also come with companion 8Blit episodes where I’ll break down the code structure, data tables, game theory, and the Robot AI. But… you’re going to have to wait just a little bit longer. I still need to produce those companion episodes, and I also don’t want to take the spotlight away from all the hard work everyone put into making the physical launch happen. With that said, it’ll likely be sometime next year before I set it free.

NEXT STEPS

What’s next? Well, more episodes, more example code, and possibly even a brand-new game sometime next year!

If you’d like to grab a physical copy of Tic-Tac-Toe, you can find the link in the description below. If you enjoy 8Blit, please consider liking the video and subscribing to the channel - every bit of support really helps. And if you’d like to help power the future of the channel, you can become a patron or check out our merch store. All the links are right down in the episode description.

That’s all for now. Thank you so much for watching, and I’ll catch you next time!

Back to top of page