Hello! I'm Tom. I designed a game called Gunpoint, about rewiring things and punching people, and now I'm working on a new one called Heat Signature, about sneaking aboard randomly generated spaceships. Here's some more info on all the games I've worked on, here's the podcast I do, here are the videos I make on YouTube, here are some of the articles I wrote for PC Gamer, and here are two short stories I wrote for the Machine of Death collections.
The winners of the 48-hour game-making competition Ludum Dare were just announced. It’s a regular thing that awesome indie devs pile onto partly as a test of their skills, partly as a thought experiment, and partly just to jam and share ideas. The theme this time was Enemies as Weapons.
I didn’t enter, though lots of absolute beginners do, and it’s the sort of thing I love. But while I now feel capable of starting a small project like that, I don’t feel capable of finishing one in any kind of time. I was also pretty goddamn determined to get elevators working in Gunpoint by Monday, because four days on that is just kind of sad.
But I did have an idea for one. I think I like the competition’s theme because it’s suggests changing the state of enemies in a more interesting way than from alive to dead. Either to make them accidentally destroy things you want destroyed, as in games where you shove them into something or dodge their attacks, or to turn them to your side.
My game would be called Defect and Serve, and it’d be about bent cops. You’re a criminal who’s just been caught, unarmed, during an elaborate bank job. The arresting officer hints that he’d let you go for a hundred grand, and while he’d like to just take it from you, if you’re dead or caught there’ll be no explanation for the missing money.
Your only option is to pay up – tossing the money to him in bundles of $50,000 by clicking. Above his head is a meter labelled “Principles:” followed by two blue pips, and each bundle of cash makes one of them explode in a little shower of bills. When they’re both gone, he leaves and heads for the opposite end of the building.
It’s from a top-down view, and once that cop leaves you’re free to figure out a route to the exit avoiding the other roaming police. Each has a number of Principles pips, usually between 1 and 10. It’s possible to get out without running into anyone else, but if you do, you’re rooted to the spot and must chuck them a $50k bundle for each pip of Principles they have before you can go.
Once you’re out, the next levels are a series of heists where you start with limited funds, and make your way to vaults and safes to get more. Your objective is to steal more money than you spend bribing, obviously, so you avoid the very Principled cops like the plague.
A few levels in, you’re told you can also toss some extra cash to a cop after you’ve already turned him. This creates a red Loyalty pip, and if you give him as much Loyalty as he had Principles, you can order him to radio an all clear to the rest of the cops, making your current location safe for a while.
If you can double a cop’s Principles in Loyalty pips, you can actually click on another cop to order him killed. The turned cop will head for him, but won’t strike until the two of them are on their own. It’s up to you whether to distract or bribe the other guards to help him achieve that, or get on with sneaking through the building and hope he has the chance at some point.
Eventually you will encounter the odd White Knight cop, who’s impervious to bribes. If he’s guarding your objective, your only option is to have him killed. At this point cops with low principles become essential tools rather than minor threats, and you’ll actively try and run into one on his own so you can turn him to your side. Once you have, it’s worth intentionally alerting the White Knight to lure him away from colleagues, giving your turncoats the chance to take him out without getting caught.
I’m actually enjoying sticking with Gunpoint, it’s just impossible to resist coming up with a concept for a theme like that. I’ll probably have another Gunpoint progress report soon – I’m close to another milestone involving the AI, and I’ll be in a good position to figure out what kind of challenges are going to be fun in this frame work.
Even as it rapidly approaches an actual game, it seems to be getting further and further from the story-heavy finished article I had in mind when I started. I pictured it story-heavy because writing is trivial, but I’m starting to realise scripting is not. And it’s a low value type of work, only really good for one play through, and little to do with the medium’s strengths. I’m starting to wonder if there’s another type of content I can use to link a bunch of puzzles, or a more efficient way to convey story without visually depicting a lot of non-interactive events.
More Game Jams
Lack 26: Would a randomly generated story be possible? Wait, don't punch me into the ground for being a big silly just yet.
The story is split into separate events, each event can then lead to several other types of events, sort of like a flow-chart. Eventually you'd be able to cycle through to a new case beginning if you solve the case, or, if a character got killed then that would disable certain outcomes and options in a case or you may open up new options by making a mole/snout/acquaintance/ally/girl-friend/enemy (all given a humourous random name, of course).
It would require a fair bit of writing, but if you keep the story light and not particularly word heavy then you could probably get away with writing just a couple of case archetypes and then working on the options within a case.
That's the basis of the idea anyway.
LeSwordfish: I wouldnt particularly like that, if the writings good, as, you know, taken as read, i'd really dislike the idea that i'd be missing out on some of it.
Tom Francis: The backend is feasible, but the tricky part is the specifics of how you depict the generated story. If you write and script a number of predefined events, players will notice when they come up again and switch off. If you make them very abstract, like The Sims' pictoral thought bubbles, it's hard to care and sometimes hard to even understand.
I do want some elements to be systematic like that, though. If there are major characters, I want to decide how they'll treat you if they hate you, and how they'll treat you if they like you, then just let your behaviour determine which state they end up in.
I think there's mileage, and game potential, in the idea that secret agent missions tend to revolve around an important object or event with at least two interested parties. A VIP to be killed or protected, data to be stolen or destroyed, etc. If I can account for both eventualities, I can both let you choose who to work for, and let you fail your mission without requiring a restart.
Bret: Sounds interesting.
Man. Game's changed a lot in the making. But it has more facepunching now, so all to the good.
Jaz: "I’m starting to wonder if there’s another type of content I can use to link a bunch of puzzles, or a more efficient way to convey story without visually depicting a lot of non-interactive events."
Full. Motion. Videooooooooo.
Brodie: I know the game (Fib) was written in 48 hours, but would it have really taken that long to make it do something besides sit there forever upon player death?
Tom Francis: Don't you respawn? I respawned. The most easily fixable problem I had with it was that you hit spikes a fraction of a second after the next screen loads, if you're still holding the movement key. It's the one time in gaming I longed for a low wall to block my progress.
DanPryce: Valve probably asked themselves the same question with Portal - weave the story within those puzzles. Granted, a lot of that was GLaDOS subtley expositing, but there are visual ways you can get story across. Like the bits behind the walls in Portal, with the 'there is no cake' scribblings; audiences pick up on environmental clues like that and they build their understanding of the story. Plus, a newspaper clipping never goes amiss.
Tom Francis: They talk a bit about this in the interview that's going up later today. The trouble is that art is the most expensive thing for me to produce by myself, and I seem to be particularly bad at environmental art.
A persistent voice in your ear like GLaDOS would definitely give me some scope for explaining things I can't effectively depict. But I'd need a twist on it. GLaDOS works because her motives aren't clear, but the usual Female Voice In Your Earpiece cliche is getting super old now.
Narration is one idea I've been toying with. I'm going to need tutorial messages, and it'd be a fun challenge to work those into a brooding noir narration. The main problem is that currently, the plan is not to force too much of the hero's character on you. I want you to decide whether the guy's laconic, surly or friendly. Harder to do if I write a bunch of your innermost thoughts for you - but not impossible.
Brodie: I didn't respawn!