alone I art Post-Mortem


As we were packing our stuff after making Eggscape, someone said something in the lines of “Let’s do this again in December!”, and since that day in August we’ve been talking about participating once more in Ludum Dare.

As the final week till LD #22 began, we followed the theme voting closely, coordinated our votes and shared the possibilities of each theme that interested us. Having learnt a lot with LD #21, we were confident that this time everything would work out better, even with two fewer members.

The whiteboard calls the shots

The Idea

Something that we might never get the hang of it is figuring when in our timezone will the theme be announced. Finally, at 2 of the morning, “alone” was crowned the winner. After fumbling around with multiple ideas, it was about 5am and only Manuel, the other designer, and I were still awake while the others slept.

We started talking about a game where you played a kleptomaniac at his girlfriend’s house, having dinner with her parents, and stealing stuff when they weren’t looking. This sounded fun, but there wasn’t much variation, and we suspected it would require a lot of scripting to be interesting.

The best part about that idea was picturing the main character trying to conceal every item beneath his clothes, so that evolved into a shady guy wearing a trenchcoat on a museum, that would stray away from his tour group to steal paintings. You couldn’t remain alone for too long or the guide would start searching for you, but this was dismissed since it could limit the player’s movement too much.

Always great when the artist nails it better than we do

We then settled with a lone shady character walking around a museum, stealing art when he was alone and no one was looking. His trenchcoat would get as large as the concealed art, preventing him from navigating through narrow doorways. If the painting you were carrying was blocking your progress, you could swap it for a smaller one to proceed. There would be security guards on a constant patrol, and visitors walking around and stopping to admire the art, but running away and warning the guards if they spotted you snatching something.

The tour group made a brief comeback, as a moving hiding spot for the player, much like the monk groups in Assassin’s Creed, but that was cut earlier due to time constraints and issues that where delaying the development of more important non-playable characters.

Not even the size of his desk could constrain our artist’s talent

What Went Right

– An environment that nurtures creativity™

This time everyone was in the same place. There was no design or art by proxy, which made things much easier, specially since the extra hand in art determined the quality of our entry.

There was constant communication between everyone, and when there wasn’t, we quickly rectified that, which help maintain a good mood throughout the development process. If someone started sighing or getting upset, we asked what the problem was and tried to figure out solutions together.

The whole project gained a perfect harmony pretty quickly: the art style match our vision from concept to execution, and our sound choices were on the spot. All this thanks to communication™.

Pair programming in action

– Good “marketing”

We wanted our game to be different. For that, we decided to keep an eye on the other participants. We didn’t want to steal other’s ideas, we wanted to avoid them, so we set up our Twitter clients to follow #LD hashtags. Eventually we started talking with other ludum darers, giving support and sharing links.

This lead to mutual interest and hype between projects. We can’t be certain of this, but hopefully they remembered to rate our entry as we did theirs.

Our updates at the Ludum Dare blog were also well spaced between to avoid spamming people and getting pushed by other posts. The Tumblr blog we made last time wasn’t forgotten, and we tried to use Facebook and Twitter in a controlled manner. Tried.

And then everything was about notch’s game

– A team of fast learners

For the second time competing in Ludum Dare, there were some amazing first times.

Manuel was in charge of level design, which he never once worked on before. We needed content, a lot of it, and we didn’t have time to prepare the game for multiple levels, so we all though that a huge museum was best. Though we now realize we were wrong, the map he made is still cleverly challenging. He even added color coding for areas that were the most far away from the exit and contain the most valuable art.

To avoid pulling the programmers from their tasks, we decided to use the DAME map editor to build our level. Learning how to use it was surprisingly easy for me but there was still the issue of exporting it to Flixel. We were exporting our map in LUA, but Flixel wouldn’t identify entities like the paintings, so when everything was rendered, it appeared as a solo background. So basically, Pedro, our engine programmer, spent one hour learning LUA to write an interpreter! Just wow!

From paper to editor

– Adding value

To encourage exploration and add value to the game, we made the Gallery one of our top priorities. This Gallery is accessed through the main menu and contains all of the paintings the player steals. Having classic, well known paintings wasn’t funny enough, so we made our “own”.

We have in total 72 unique paintings in the game, all of them are scattered in the museum, and each have a title on them. Most are nods to other games, while others are Internet jokes.


What Went Wrong

– Again with the framework

Last time we had problems with our tools, so this time the programmers decided to research in advance for something that suited their needs the most, and found they Flixel. Unfortunately, this engine turned out to have it’s own brand of headaches.

We had problems swapping animations, changing and rotating the guards’ cone of vision, setting up spatial sound and developing better pathfinding than just by points and that avoided getting stuck in objects. Most of these features had to be dropped to meet the deadline.

But worst of all, it was picky with our sound files. Some wouldn’t play completely, others wouldn’t embed when compiling the SWF file, giving multiple errors when trying to stream external audio files. We nearly missed the submission window because of that, forcing us to release the game without sound. Only on the following morning, after the Jam had ended, that we figured out that the problem was related to file size, so we got back to work and solved some minor bugs.

This time we were able to test
a prototype early in development

What’s Missing

We made our best effort to make this game rich on content. From all the features we had to cut due to unforeseen problems with the engine, the quantity of paintings and the gallery menu were elements that we had to deliver. Though we should have spent some time looking up how to save the player’s progress.

The guards took so long to implement that we had to drop the visitors and change the setting to a museum at night. The placeholder circle around the guards, used to test their cone of vision that we didn’t finish, stayed in the game with a different color and opacity, and we gave a flashlight to the sprite to pretend it was a circle of light.

There’s a lot of art missing from the level, mostly museum props used either as decoration or obstacles, because not everything was exported from DAME.

And after these 17 days of voting, we read your comments about the game and completely agree that it should have had a smaller map or multiple exits.

It’s not that big…


Although we had our share of problems, everything worked out better than last time. We took our time planning who would do what and how.

Manuel worked on the game design with me, designed the level and made additional art (paintings and props). Fred took care of the AI and sound. Tiago made all the characters and animations. Pedro and Daniela programmed the gameplay mechanics and menus. And I translated the level into the editor, while also making mockups and acting bossy.

We prepared in advance, setting up a local SVN server instead of using Assembla like last time. We forced ourselves to be original by identifying the theme’s common patterns and avoiding them. We tried to sleep at least 6 hours per day and were well stocked up on food.

There was a lot of stress in the final minutes of the Jam, but it was a great experience nevertheless!

Regarding future Ludum Dare events, we’re talking about of at least finishing Eggscape before competing again, but we also have other side projects up our sleeve. So we can’t decide on Make A Game’s fate just yet. I mean, third time’s the charm, right?

You can play “alone I art” at Ludum Dare or at


Here’s a video of Banksy that Manuel showed us when we thought of the trenchcoat guy and swaping art. Totally had forgotten about it!

Tools Used

FlashDevelop, Flixel engine, DAME map editor, Audacity, Tortoise SVN, Adobe Photoshop CS4, Google Docs.

This article was originally posted on January 6th, 2012 at Ludum Dare.