groupthink Windows game

I've been working on groupthink since mid 2014. It started off as a challenge to get a game completed before my daughter was born in September, and has turned into kind of an endless mission to build as big a top down block pushing multiplayer puzzler as I can make before life pulls me in other directions.

There's a lot of inspiration behind groupthink. An old game for game boy called Kwirk, Portal 2's two player mode, lots of Sokoban-like games and other block pushers. But the catalyst for why I made this game was because I wanted to play a simple, top-down oriented puzzle game with my wife. In this vast internet I assumed such a game existed; sokoban had so many variants and clones out there that surely somebody made a version of it for two people to play.

Unfortunately no such game seems to exist. After a couple of months I sat down and banged out a simple sokoban clone, while keeping it easy to expand and of course making it easy to make multiplayer in the long run. In a few hours I had a couple hundred lines of code and a little P pushing around little B's to little dots on the map.

I wasn't really interested in making a simple sokoban clone so I added some doors, and keys, and conveyor belts to move objects around. I wanted the puzzles to be less like a spacial oriented puzzle game like sokoban and Kwirk and more like Portal's puzzle style-- logic, timing, object placement, and putting together the pieces. So this meant I needed more and more pieces with more and more abilities.

Soon the box destinations were multi-colored and the game was no longer about putting boxes in place, but about getting to the exit after solving the puzzle's components. When you solve one color of the sokoban puzzle, it opened up other parts of the map, opened a door, or turned a conveyor belt around.

All along I was getting a little nervous about actually making maps. Yeah, I've got all of these pieces coded up, but I had yet to make any real maps that could be challenging to players. Is it possible to make a challenging puzzle when players could see everything they needed to solve the puzzle?

You have to remember that Portal's puzzle style is the direction I want to go here. In portal you can't see everything at once-- you have to look around the map and figure out the steps in a 3D environment. I don't have that advantage here.

So when I finally started making maps here, I was super anxious that two or three months of work were kind of pointless, and I'd never be able to make any puzzles for this thing that were at all challenging. My first set of puzzles were kind of underwhelming. It had 15 maps in all and ended with a tangle of conveyor belts and you had to get a box to the center of it all. They were mostly a tutorial but who was I kidding; I just wasn't ready to apply my mapmaking skills yet.

I'm a coder after all. So I went back to the code to make more objects, thinking they'd make the puzzles for me.

Lasers soon followed along with reflectors and splitters to redirect them around the map and light up laser "receivers" in the correct order, and buttons were added that players could either stand on or push a box onto. Now players had to wonder if they needed to put a box on a button, use it to solve a sokoban puzzle, or use that box to block a laser that was keeping a door closed. Things were getting more complex, but not enough.

Added to my anxieties about making puzzles in my puzzle game came the whole "interface" thing. At this point all I have is keys for an interface-- press F6 to load a map file, F5 to save, F3 to delete a map, and while you're editing, you better remember every key that I assigned to every object. It all had to change. I can't have players referring the readme.txt to figure out how to start a game.

So I begrudgingly added a menu system. Which of course takes days and days of extremely boring work of moving buttons around little by little. But finally I had a real system for interacting with my game! Along the way I gave the game some real graphics, since nobody wants to look at the shit I was generating via graphic library calls.

Back to making puzzles. Now it was time to really step out of my comfort zone of coding and daydreaming of how awesome these puzzles were going to be. My first few puzzles were of course simple "how do I play this thing" maps that I hoped conveyed how to play the game. I remember the first time I realized I could use one of the object blocking "red tiles" in order to prevent a conveyor belt from moving a box until the player turned the red tile green. That was kind of special moment for me, and to this day red walls on conveyor belts are some of my favorite puzzle pieces. It's something about controlling boxes from across the map that I think gave a map an extra challenge-- now you are not only figuring out where a box has to go, but you're also figuring out how to manipulate the conveyor belts to get the box there.

I ended up with 24 maps and I called that the single player campaign. I don't think the maps were particularly challenging; I was still experimenting with the puzzle pieces and figuring out how to ge them to do interesting stuff and make challenging puzzles. For the most part I avoided sokoban-style space management. The lasers were the least used part of the game, which is a shame because they took the most time to make. I'm still just getting the hang of lasers, and if you play through any of the maps you'll see that they generally lend a supporting role in the maps I have so far, with only a few puzzles devoted to lasers.

And so on I went to two player puzzles. If single player puzzles were daunting, now I had to utilize these same pieces to create maps that couldn't be completed by one player, giving me more to think about. Now that I think about it I had an easier time making multiplayer puzzles than I did single player. I can box in, separate, and give separate tasks to players and assume that part of the puzzle is just going to be figuring out who needs to go where at what time. I think the first batch of two player puzzles was 11 maps long.

The three player maps were easier-- I was getting used to my tools and my game. Maps were no longer something I sat and thought about for a while before building, I just started putting shapes on the window and began letting the map kind of build itself. Before I knew it I had 6 or 7 maps for three players.

I was really excited to share what I had so far with the reddit "gamdev" community and so I put it up and got a bit of feedback from one guy. He told me he hated that I didn't have gamepad support. Another guy played through almost every map; only one couldn't be played because I had broken it somehow along the way while fixing some bugs.

A week or two went by; I added gamepad support. That thought the two player maps were really difficult and confusing-- which is good I guess, but it reminded me that I never made a two player "tutorial" map set. I made the two player maps right after putting out the most difficult stuff I could make for one player, and falsely considered two player mode a continuation. I had assumed anyone playing the two player mode must have gone through the single player. Back to the drawing board for multiplayer modes.

Another suggestion I got was for some more feedback in the game-- the game gave no indicator for what you should do at any time, doors opened randomly when you pushed buttons, sokoban puzzles tore down walls. Color coding the puzzles to show wtf was going on was something I was reluctant to do because I wanted players to have a sense of mystery--- like how portal had 3 dimensions to look around in, I had a no feedback until you completed one puzzle; it made sense at the time. The few people who reviewed it convinced me that I needed to give players some hints, otherwise I'm just hiding the puzzle from the players.

Another issue I was having was that I had to send out a new version of the game in order to add more "core" maps. I had a system set up where users could upload map data to pastebin.com and then share the URL; it was the easiest way to let people share maps... but that wasn't helping me because I didn't really want players to have to use a mildly wonky copy/paste system to play the core game campaigns.

And so I came up with the idea that I'd have buttons for some campaigns that you couldn't hit until I released the map. Currently four player advanced maps are downloaded when you click the play button--- this allows me to add or change maps, but most importantly it lets me update everyone's game with content automatically and without forcing them to go download the latest version.

This led to my current favorite feature-- Monday/Thursday Challenge. This is a single player feature that hopefully will one day keep players coming back for one more puzzle twice a week. Every few days I'll be adding a map to this online map file and the game will tell the player as they load up if there is new map to play since they last loaded. I have some challenging puzzles already designed for this feature, which as I write this currently has three maps.

So that's pretty much where I'm at. I'll write a separate document about the puzzle design that gets into that part of the game development a little deeper. If you somehow got through this whole thing, thanks for reading!

-daveyeah