Friday, April 12, 2013

Accessible Puzzle Design in Pushmo


Pushmo is a puzzle game released in late 2011 on the Nintendo 3DS. After the 3DS suffered without a great exclusive game for the better part of the year, Nintendo's Pushmo came on the scene and seemed to succeed on every front: it was charming and accessible, had a lower barrier to entry than most 3DS games (it could be downloaded straight from the eShop for only $7), and its gameplay  nicely demonstrated the system's 3D screen. On the recommendation of a friend, I picked up Pushmo last year and played it on my commute for several months. While it's not perfect, I found it contains some important lessons about design puzzle games for a broad audience

Pushmo does a few things extremely well to achieve broad appeal. It selectively borrows from platformers while keeping the player's focus entirely on puzzle-solving. Its more challenging puzzles start off simple and accessible to avoid scaring off the player. Lastly, Pushmo enables players to develop advanced strategies for completing puzzles, by providing visual checkpoints to help break a puzzle into its component pieces. These elements come together to ease casual players into the game and encourage some to play though and complete its most difficult puzzles.

A puzzle game, not a platformer.
Pushmo appears to benefit from being designed as a puzzle game first, and everything else second. On its surface, the game controls like a platformer - players run, jump, and grab blocks in 3D space. However the game smartly avoids becoming a platformer by placing all of the challenge in its puzzles, and removing any pportunity for the "platforming" to cause unnecessary friction.

Whenever the player accidentally falls or makes a bad move in Pushmo, they can rewind time to undo, and a reset button allows the player a fresh start whenever they'd like. The game prevents players from accidentally walking off of ledges by pausing Mallow at the edge of a block. It even prevents the player from accidentally landing on the "reset" button when jumping from the puzzle. These details often go overlooked in puzzle-platformers, but Pushmo's developers knew they wanted to test a player's mastery of puzzles, not their manual dexterity. Pushmo successfully encourages experimentation by making sure players are never punished by the controls.

The first step is always the hardest, except in Pushmo.
Puzzle games can often seem impenetrable to players, especially when they are just starting out. Consider, for example, a Sudoku puzzle in its starting state:


There are 459 potential starting moves (51 squares x 9 digits) in this puzzle, and an inexperienced player has no frame of reference for which ones to prioritize. Advanced Sudoku players know how to quickly sort this information and focus on just a few potential moves. They know that the first or last columns are good places to start, because each has only four empty spaces. That narrows down the number starting of moves to a manageable 16 (4 squares x 4 digits), and they can iterate through them, eliminating options as they go. All of a sudden, they're on their way to solving the puzzle.

Similarly, Pushmo starts by showing the player a large grid of blocks to manipulate. Like in Sudoku, players have to keep making moves until they reach the goal (solving the puzzle). In Pushmo, each block has 4 possible states - pushed all the way in, or pulled out into the foreground up to 3 times. The Pushmo puzzle below contains 25 pieces. An inexperienced player might feel overwhelmed by the 100 possible starting moves and nearly infinite combinations of moves that can follow them. Without some encouragement, the player might even turn the game off or start playing a less confusing game.


However, Pushmo is kind to its players and lets them dip their toes in the water before taking the  plunge. As it turns out, the first step in most Pushmo puzzles - even the advanced puzzles later in the game - is usually very easy, gravity limits the options. Mallow starts at the bottom of the puzzle, and can only manipulate blocks that exist at his current elevation. In the example above, that means there are only 3 blocks that can be involved in the player's first move. Already, we've narrowed things down from 100 moves to 12. Since Mallow can only jump high enough to climb the lower-left block, the first decision is an easy one. In this puzzle and many others, the player's decision tree for their first few moves is fairly linear, as in the first figure below.


Later in the puzzle, things will not be so easy. The graph of possible moves becomes much more complex and nonlinear, as in the second figure above, often requiring players to backtrack or manipulate the same block multiple times. However, the player's "momentum" has already picked up by this point, and the game has rewarded them with enough forward progress that they are less likely to get frustrated and quit. By starting each puzzle with a few simple moves and ramping up the difficulty later, Pushmo avoids overwhelming the player before they're having fun.

Careful observation and planning.
In Pushmo's advanced puzzles, things start getting hairy. Not only are the later levels bigger and more complex, but they introduce new elements like switches that pull every block of a certain color to the foreground, or ladders that act as portals to their twin elsewhere in the puzzle. These elements do more than add variety: they serve as visual checkpoints that the player can use to plan their strategy for completing the game's largest puzzles.


The puzzle above - one of the last in the game - is visually complex and contains many small, hard-to-reach blocks. There are also two pairs of ladders - stepping onto a red ladder teleports the player to the other red ladder as long as both blocks are pulled into the foreground. Finally, there are red an dark green switches, each of which will pull all blocks of that color to the foreground - helpful for blocks like the lizard's red tongue, which have no nearby footholds. Using all of these visual landmarks, the player can break the puzzle up into stages, as in the image below.


There are five stages to this puzzle, starting at the bottom and ending with the goal at the top. The player can mentally break the puzzle down into these stages by working their way backwards from the goal. To reach the goal, the player can use the blue ladder. For the blue ladder to work, the player needs the red blocks to be pulled out, so they need to hit the red switch, and so on. Players can recognize the "critical path" by following the placement of switches and ladders. Having these short-term goals helps make Pushmo's puzzles more feasible as they scale up, allowing the designers the freedom to create puzzles from larger, more complex images.

Every puzzle has multiple solutions.
These are the three elements of Pushmo's design that I feel gave it the biggest opportunity for success, by making the game more accessible while still containing many challenging puzzles. The game focused on puzzling, not platforming, and removed any unneeded friction. The puzzles start off easy and ramp up as the player moves through them, which helps get the player acclimated to each space without feeling overwhelmed. As the puzzles get larger, the game introduces new elements that help guide the player and turn large challenges into sets of smaller ones.

However, I suspect there may be those who disagree with my assessment. On the flip side of the coin, there are at least two good arguments to these points that I can think of (and there are probably many more):
  1. Pushmo gives the player the information necessary to form a plan and strategically work their way through the puzzle. Many "casual" puzzle games, however, employ the opposite strategy. Some of the most popular puzzle video games, such as Tetris and Bejeweled, make it relatively difficult to plan too far ahead. The player has a limited understanding of what the board will look like two turns from now, and even advanced players often make moves based on current information instead of looking forward. Often, but not always, these games have no "win" state - the challenge is to see how long you can "survive" or how many matches you can make. By using the "keep trying until you succeed" model, Pushmo avoids failing the player but also delays positive feedback to the player until they've completed the puzzle. There is no "partial credit." It's possible this is more alienating to casual players.
  2. Because the first few moves are usually fairly straightforward, every puzzle starts off the same way. It is very rare for there to be an alternate strategy at the beginning of a puzzle - the best move is usually to pull all of the bottom row all the way out. It's possible that Pushmo becomes less interesting for players as time goes on because this strategy doesn't evolve much even in the later stages.
Have you played Pushmo? Have you played its sequel, Crashmo, which I am yet to check out? Do you have examples of any of the above from other puzzle games? I would love to keep the conversation going in the comments.


Thursday, April 11, 2013

Catching Up

It has been over nine months since the last post was published on Thumbstruck! As my excuse, here's a quick update on what I've accomplished in nine months:



  • Completed 14 more games, including Rayman Origins, Pushmo, and The Walking Dead (all highly recommended). This brings my grand total to 42 games completed since I started tracking them (September 2010).
  • I proposed to the woman of my dreams, and against all odds, she said yes. We our planning our wedding and getting married later this year.
  • I helped launch a lot of new stuff at work. If you own a Windows 8 computer/tablet, there's a decent chance there's a game on your computer with my name in the credits. Hooray!
  • I helped run a game with Arkadium at the Come Out and Play festival on Governor's Island. The game is called Super Secret Spies, and despite the sneakiness, there are tons of photos online.
  • I attended Practice, a game design conference hosted by NYU Game Center, and heard talks from some truly remarkable people. No single event has reinvigorated my resolve to create great games more than this conference. If you're interested in hearing about any of the talks, let me know!
  • Got my first smartphone! I am now the proud owner of a Nexus 4 running the latest version of Android, which also means I have new free-to-play games to try out. My commute is Triple Town, and Triple Town is my commute.
  • I returned to TCNJ for the Global Game Jam 2013. With my good friends (and dream team) Bryan Mayer, Ryan Epp, Chris Hallberg, and Nicole Pieri, I built _scope, a psychadelic game of Set-meets-Spirographs, with just a hint of childbirth. For real.
  • My writing has appeared around the net a bit recently, including this opinion piece on iMedia Connection about the then-new Facebook App Center.
  • I wrote a  feature story on Gamasutra covering five tips for better playtesting. This was recently re-published in the March issue of Game Developer Magazine, which was distributed to subscribers and at GDC. Let me know what you think!
And that's it! If you're interested in chatting about some months-old news, I'm happy to take a trip down memory lane. Leave a comment!

(Thanks to Vinny for pushing me to start writing again.)