Wow! Thank you everyone who played and rated Cut to Pieces! To be #1 among the competition that Ludum Dare has to offer is a great honor. I'm also really happy that we won innovation as well, which is the category I value the most. And I'm happy for our artist, who's skill has been recognized with a bronze trophy.

So, time to tell you how this game was developed in 72 hours.

Idea

Like last year, I tried my method of meditating half an hour before and after the theme was announced. The idea is to empty the mind, to make it bored in a sense, which then results in more creative ideas. Last year that got me to come up with Vox Regis, this time that was less successful. One idea that came to mind during this session is a game about someone who bought fake paintings, and where you need to go back in time to change history, in such a way so that your painting is accurate. A bizarre idea, of which I wasn't sure how to implement it. I quickly decided to let that idea go, but I liked the idea of doing something with art forgery.

There were a few ideas with asynchronized multiplayer. That, for example, you need to redraw a paiting, and that other players need to guess which painting you drew. There's definitely potential there for some great ideas, but we already made a similar game last Ludum Dare, with Captain Peggles: That Man's Chest. And there's nothing wrong with using a similar idea, but Ludum Dare is also a way to experiment, and to challenge myself. I don't want to be the asynchronous multiplayer guy. I want to see if I can come up with a new innovative and fun idea.

Later on I got this idea of stealing paintings in a museum. But, to prevent the security guard from noticing that the paintings were gone, you would have to replace them with similar looking paintings. You would do so by using stamps or stickers. My older Ludum Dare entry Harvest Heist came to mind, where three bandits need to steal vegetables, while avoiding the farmer. I was actually considering to make it a sequal. But it would be a lot of work, and I wasn't sure if it would be fun. I considered what the game could be like if I took away the walking, where it's just the painting, the stamps, and the canvas. It was then that I came up with the cutout mechanic. What if there aren't stickers, but you need to create your own "stickers" from pictures in a newspaper?

Every Ludum Dare I start with a big question: Will I be able to come up with something fun and unique? In my opinion, for the past five Ludum Dare jams, the answer is yes. That's a very good feeling. Even though I won before in the past, I've always had self doubt about myself as a game developer. Imposter syndrome is very real. But, with three Ludum Dare victories, and the innovative types of games I produce, I feel confident in saying that I am, in fact, a good game developer. That doesn't mean it's easy for me to come up with these types of games, but I should trust myself that I am capable of doing so.

Cutting

To make the game work, the player needed to be able to cut out image. The first part was easy, where I keep track of the mouse coordinates as it makes the cut. I then draw the line using these coordinates. The challenging part was the cutout. How do I figure out which part to cut out of the newspaper. The inside of the circle, of course, but how to figure out what the inside of the circle is? I asked for help in the LÖVE Discord, and got the suggestion to do a floodfill on the outside of the circle. Because, while you might not know the inside of the circle, you can know for certain which part is on the outside. By starting a floodfill from that position, it will fill all but the closed circle. By using the non-filled pixels, you now have the cutout.

At first this floodfill happened over the whole screen, but this gave the game a subtle lag each cutout. So I optimized it. While drawing the cutout, the game keeps tracks of the boundaries of where the mouse has been. By using this rectangle, with an increase of 1 pixel on each side, I have the box in which this floodfill should happen. Because of the 1 pixel increase, I know for certain that the upper left corner will not be inside the cutout.

Cutting out the image is done using LÖVE's ImageData API. Using the mapPixel method you can go through each pixel, and return what the new pixel color should be. Using this method it's a matter of checking if the pixel in the cutout canvas is transparent, and if so to copy that pixel to a new canvas, and to return a transparent color.

Solving

The second problem was to figure out whether the player beat the level. Somehow I had to keep track of the cutouts and their placements, and whether this was good enough to consider the level solved. It was already getting late on the first day of the jam, but I really wanted to get this done early. The way I solved this is by giving both the newspaper and the empty canvas an invisible colored layer. When you cut out part of the newspaper, you also cut out part of this invisible layer. The apple of the first level, for example, has a brownish color. Then, on the canvas, the apple also has the same color. When you submit your painting, the game checks how much of each object's color is being overlapped by something of the same color. All objects have a purple border, which is the safe neutral color. Cutting this out gives you neither good points nor penalty points. Around the border is red, which does give you penalty points.

This system works great. Using Aseprite's spritesheet and JSON, I could export the colored layer in the same image, and then easily figure out what part of that spritesheet is the colored layer. Luckily, exporting and utilizing separate layers with Aseprite was something I looked into recently. The system is not perfect though, and the expanded game will require a system that allows for more options. Some newspaper images were smaller than what was visible in the painting, and therefore would never be able to cover the whole image. You could solve this by creating a smaller area, but then depending on where the player places the image, which should be considered correct, could miss that small area and fail. In this case, a better solution would be that a certain percentage of the available color is part of the cutout, and that the entire cutout is covered by the color on the canvas.

If we didn't have a solution mechanic, the game would still be fun. Players, unaware of this lack of solution system, would still try their best to do well. A lot of players were very careful when making the cutout, and were worried that their slightly wrong cutout might give them a penalty, even though they didn't have to be so perfect. The problem is that if a player beats the game without failing a single time, they can wonder whether it was even possible to lose at all. In an expanded game I would add a percentage, or a 5 star system, to show the player how well they recreated the painting.

Paintings

There wasn't really a plan as to what the paintings would be. I also didn't have time to direct our artist, so I let him do his thing, with sometimes giving him an idea of what to implement. One such idea was the mermaid. A combination of two cutouts to form a single subject was not something I considered when we started making the game, but upon realizing this is something we could do, I saw the true potential in this game mechanic.

I've seen multiple comments praising our artist, and rightfully so. They love the character design, think the paintings look amazing, but mostly think it's very impressive that he was able to draw all of this in 72 hours. All those newspaper articles and cutscenes as well. But to Shores this was business as usual. He's fast, he's experienced, and most of all, this is his comfort zone. He just drew what he wanted to draw, which resulted in fun newspapers, cool paintings, and silly characters.

The game has four different type of solutions.

1. Two items of the same subject. Example: The apple in the first painting, and the apple in the newspaper.
2. Two items of the same category of subject. Example: The old ship in Mermaid, and the modern ship.
3. Two different but similar looking items. Example: The mountains in Hood, and the icecream cones.
4. Two items with similar properties. Example: Medusa's hair, and the spaghetti.

When thinking of the game mechanic, I was mostly considering 1 and 3. It was therefore that I protested when Shores came with the suggestion to use spaghetti as the replacement of Medusa's hair. I figured it was too vague, and people wouldn't figure it out. Shores thought it would be more than obvious. We compromised by adding the handle of a fork, which mimics one of the snakes on Medusa's head, and making it partially a type 3 solution. It turns out that I was completely mistaken. In the far majority of the playthroughs I watched, people quickly figured out that it was the spaghetti they needed, and often ignored the fork.

Steam?

Yes, we are going to expand the game. It's something I didn't want to do at first, but after reading the comments, and watching people play, I realize how much people enjoy this game, and the potential it has. We don't have a Steam page yet, but you can follow me on Bluesky to get updates.

Thanks again everyone who played the game, and thank you Mike for hosting Ludum Dare among all your financial troubles. I hope you will find a job soon. I'll be there for Ludum Dare #59.