Repeated Failure

Lately I have been thinking about introductions to games, particularly learning to play through repeated failure. Most games make every effort to make sure the player is completely comfortable with how to play before they start. They go through tutorials, are eased into puzzles or enemy encounters, guided by level design or pop-up advice. Some games stand out as doing things differently. The most obvious examples are the Dark Souls games, where a player is expected to repeatedly throw themselves against a bone wall, as they learn the tricks to defeating encounters.

A screenshot of the game over screen from Dark Souls
You died!

However, I still think there’s room to explore this space. When ADR1FT came out, a game about being the sole survivor of a catastrophe in space with a missing memory, I hoped its opening would just drop you into a leaking spacesuit floating around a wrecked space station with your air leaking out into the void and no instructions, and forced you to experiment to survive. I haven’t played it, but I doubt that’s how they do it.

At Nordic Game Jam 2016, I decided to try experimenting with the idea of learning through hard failure. At GDC, a friend of mine, Denman Rooke, showed me an adaptation he had made of Dark Souls as a card game, where the crawl through the game is simulated with a stack of enemies that you work through one at a time, but when you die, the stack is reconstructed in the same order as it started. This means that on each run, you learn how to be more effective and get further through the deck.

Denman Rooke explaining the card game prototype in a hostel common room
Dark Souls: the card game

I wanted to explore this mechanic further, and I came up with a card game based on Hackers (the movie). The game didn’t work, but I still think some of the design ideas were interesting.

The poster for the Hackers movie in all of its 90s glory
Hackers

Hackers

The theme for the jam was “Leak!” which, like many others, I interpreted as a data leak. From the outset, I wanted to focus on this idea of learning through failure. I also wanted to work on a physical game for a change. I was massively inspired by Denman’s game, but I wanted to create a more social experience, rather than a solitaire game, so I took the core deck crawl, added two players and some bluffing mechanics (the beginning of the game’s downfall).

The premise of the game is that you were a group of rival hackers in the early 90s, trying to break into a corporate data store and steal their shit. However, you can’t do it alone. You want to be the hacker to take the scoop, but you need the others to help you break through the firewalls and other defences. Being the 90s, you’re armed with floppy disks loaded with kilobytes of software, running over 56k modems to attack corporate mainframes and steal megabytes(!) of sensitive data from them. You have a supply of energy drinks to fuel you, and a reputation to keep you honest (or at least not openly treacherous).

Design

The game starts with a draft phase, where the hackers load up their decks from a shared supply. Each hacker is dealt four floppy disks, which they pick one of, and pass the rest on. The intention here is to give each hacker an idea of what the others might have, which gives them a head start against the bluffing that comes later. With three players, they each see their first draft twice, and everyone learns about the definite location of one card outside their own deck.

Once the hackers are kitted out, they make their first run at the corp’s mainframe. To do this, they load up one of their disks in secret (by playing the card face down), and then all reveal together. Each disk has a load time, which determines the execution order. The first hacker looks at the top card from the mainframe stack in secret. The revealed mainframe has two of three possible layers of defences, with consequences for failure, and each disk has two hacking capabilities, of which the hacker must choose one to execute.

The first hacker must run the mainframe, and the results of that run are public. Each other hacker then takes it in turn to choose whether to execute their code or pull out of the current round and lose some reputation. If the mainframe’s defences are defeated, the card is placed face up, the disks return to the hackers’ decks and the process repeats. If not, the hackers become locked out and come closer to being detected, the mainframe stack is restored (although some of the mainframe failure actions result in rearranging the stack), and the hackers must start again from the top.

In amongst the mainframes are databases, which only have a single defence, and the first hacker to break the defence claims the data. Data is how the game is scored, and so this gives hackers an incentive to risk going first into the unknown, which is a potentially harmful action for them.

The hackers, therefore, need to make sure they have all the defences covered in the correct order so the group can succeed and progress, but also need to avoid taking penalties as individuals. This results in a drive to be later in the execution order, so as to avoid responsibility for taking the deck, but that drive should be balanced out by the fact that late execution pretty much guarantees that the hacker will never get any data.

The game continues until the hackers are detected, any of them run out of energy drinks, or any one of their reputations runs dry, giving them incentives to work together. At the end of the game, assuming they didn’t run out of reputation (which disqualifies them), each hacker tallies up how much data they managed to steal, and the one with the most to brag about wins.

Issues

At some point I added checkpoints in the form of back doors, to try and make the game less punishing after a long run. Playtests showed that people were put off by the idea of having to repeat the process of going through the entire stack again from the beginning.

The fact that hackers’ decks are fixed from the beginning of the game means that I found that players wanted to always play the same disks every round, since they knew what the outcome would likely be. Changing up the disks every round would have been too slow, so instead, I added a penalty, so if a hacker plays the same card twice in a row, they take a hit to their reputation for unoriginality. This worked to an extent, but added an extra layer of mental bookkeeping in a game crammed full of hidden information to remember, and it relied on players being honest, which in a game that encourages secrecy and lying, isn’t a good thing to rely on.

Overall, the game is just trying to do too many things. The bluffing mechanic is at odds with what is essentially a memory game. It adds a layer of competitiveness that isn’t really supported by the rest of the design, and turns out to be fairly dull due the lack of depth of player actions. Combining that with the need for players to remember the order of the mainframe stack, meant that players basically ignored the bluffing aspects and just focused on breaking through the defences.

The game would have been better as a more cooperative experience, with all the players working together to plan their actions to make the best progress through the corp’s mainframe stack. If I revisit the game, I would definitely try to make it a cooperative game, rather than a secretive competitive one, though perhaps with a mole, or some sort of faction mechanic to add a competitive edge to keep things interesting.