The Conversation is a retelling of my meeting with Dan Stubbs, who is developing the probably-too-ambitious-for-its-own-good game, The Hit. In part one, we discussed the squandered promise of the GTA open world model, in part two, the restrained ambitions of AAA. and part three touched on financial success and multiplayer. This is the final part.
While we headed towards his flat, Dan Stubbs made a stray observation that perhaps Mike Bithell is “betraying” his audience making a Metal Gear Solid game when they probably bought into him for more Thomas Was Alone. But we didn’t discuss the problem of developers trying to buck their stereotype (like Chris Park from AI War to Tidalis) and focused instead on Metal Gear Solid, because Stubbs loved that game, even though it is well-known for, er, cutscenes.
“Metal Gear Solid is pure opera,” he mused. “It has events that are so much larger than life and is fantastic because of it. People moan about the insanely long cutscenes but they’re as much a part of it as anything else.”
“It’s why people buy it,” I added.
“I think Metal Gear Solid is one of the best narrative games because it is so self-aware about what it is doing, about the fact that it is so completely ludicrous and over the top. But it’s playful – and I think this is one of the most important things – there’s far too many cutscenes that aren’t, just portentous doom-laden things, where Very Serious Things happen.”
We crossed the road.
“I’m generally more forgiving of so-called bad design these days,” I said. “I mean, we can find a way around it, you know, but if we said look I’m not going to use cutscenes, then we’re going to have to do without Metal Gear Solid. I think a lot of people would be poorer for it.”
“It would be very, very sad if there was no Metal Gear Solid.”
“Then again, on the other hand, Bioshock Infinite.”
After Stubbs did a bit of shopping, we reached his flat and I was faced with a terrifying dilemma of journalistic ethical proportions: Stubbs cooked an omelette for us both to eat. Corruption is everywhere. I interpreted this as a nod and a wink that I had to give The Hit favourable coverage.
That’s probably why I began to focus on the The Hit. I asked, “Do you worry about losing the ability to stay determined?”
He shrugged it off. “No. If I was going to give up I would’ve given up at the start of the year. Or when the Kickstarter failed.”
Dan Stubbs launched a Kickstarter during the summer so he could keep working on The Hit daily, but it did not succeed. I suspect the Potato Salad Kickstarter stole all the funds that would have flowed into his coffers. He looks back on this as a positive outcome: “No, no, £20,000 would have been really nice and I’d have carried on making the game – but I do like being able to refocus. The fact that the Kickstarter failed means I don’t have to make exactly that game. The update I did with the story bullets and stuff was partly me playing around, seeing how people would respond to that idea while I had a bit of publicity, because I do want to do that or some variation on it.”
The conversation branched out into the clarity and feedback of modern systems. AAA design has improved to make sure players are never in the dark about what a game expects of them, but has the dial has swung too far that way? “Free games, art games, various indie games are the only work that has been addressing that,” I said.
“I don’t know,” Stubbs responded. “I think there’s two sides of it. It’s like the cops in The Hit. This is something I’m having difficulty getting across to players because players expect the game to be GTA-lite or something. The cops are not in there as things to be shot, they’re not in there as enemies or targets, they’re there so that the world can shout NO at you. They will kill you unless you run away and do something else.”
I remembered the Ultima series as a precedent for this. “If you stole from somebody in Ultima IV and they saw you do it then you’d get the wrath of the guards and they were like TANKS. The best thing you could do was leave the city. Unfortunately, you could go straight back in afterwards and the game would forget.”
“Actually, Ultima IV was different because if you weren’t a good person you couldn’t complete the game. As soon as you stole, you’d knocked yourself back a few points. Even attacking first in any encounter was frowned upon as not particularly valorous. But, yeah, the guards were not there as a challenge but as a message: you’re not supposed to be doing that. You could beat them, but it was always traumatic.”
Stubbs mused about players in Assassin’s Creed who should be weighing up how to get past guards to reach a treasure chest but: “Players generally end up twatting them in two seconds because the game’s so easy. And that’s kind of what I mean about the lack of indefinite language in videogames, that there’s nothing really there. It’s like a wall that you can step over. There’s no decision.”
“That was kind of the issue I had with Dishonored because it was all about customizing your experience and you can get through any way you want and it doesn’t really make much difference. If you want to slaughter everyone there are some narrative consequences to that but – you know what, if you’re the kind of person who really enjoys killing people in games, it’s going to let you do that and you can have a whale of a time. Look at them disintegrate when you stab them in the neck! Thief and Thief II would baulk at the idea of killing people left, right and centre if you were playing the harder levels; you know one death is going to stop the game. Admittedly that’s a bit authoritarian but now we have whatever way you do it is okay!”
Stubbs describes an experiment he wrote called Moiety. “It was an exercise in a minimalist environment with minimal but strictly defined language. There were two types of character in Moiety. There were the white characters who were these ghosty wraiths who would run towards a goal which, if you headed towards the same goal would win you the game. So they were YES. There were the other characters, ghostly black wraiths with white faces – they would run towards you and kill you. So they were NO. It’s just a world composed of YES and NO and trying to make that as clear as possible without forcing anything on the player. But yeah that was me playing around with turning YES and NO into definite gameplay things rather than AAA that tries to make everything YES so that you’re always winning all the time. Wandering around a world of maybes where all these NPCs tell you what an amazing girl you are. Yeah, Dark Souls got it right.”
After omelettes are consumed, Stubbs starts showing me bits and pieces of The Hit and other prototypes. At the time, he was still using prefab buildings – procedurally-generated cities were yet to come – and was working on an interaction system. Posters were generated procedurally using in-game characters, and Stubbs showed me the “studio” underneath the city where these posters were generated. He demonstrated a little of the pseudo-text he was developing for these posters to give the city more character, something that looks like writing at a glance but is actually gibberish.
He showed me some of the game’s tricks as well, how 2D models of the NPCs pop into 3D as we get closer to them. “Ignore the fact that they’re walking through buildings.” Then an NPC limped past. “I don’t know what’s up with that animation glitch.”
“Maybe it’s actually deliberate,” I suggested.
“Yeah, let’s go with that. Everything is on purpose.”
Then two characters walked by with the exact same limp-glitch.
“Basically I want those guys in there as background detail. They’re not actually characters yet. Although you will be able to point your gun and kill them if you wanted to. But–”
I finished his sentence: “But that’s not the point of the game.”
“No, the point of the game is to have a world of people who you shouldn’t kill because you don’t have that many bullets and the police are very good shots. I’m going to make it so if the police catch you and you’ve got a gun, they’ll just open fire on you and kill you in seconds. Whereas if you throw the gun, run away, they’ll become much worse shots.”
I didn’t want to get carried away with the graphics but it reminded me of Eskil Steenburg’s Love. Stubbs didn’t admit it was the inspiration for the current style of The Hit but said, “Yeah, I played that for a bit. It’s got a gorgeous look.”
“It’s another one of those games that make me so nervous about getting hooked on the graphics. Everybody fell in love with it and it sounded cool and then people played…”
“It’s an amazing game but so hard to get into. I don’t know if it needs a tutorial but it’s one of the reasons The Hit is set in a familiar environment. I don’t want to go for some kind of abstract fantasy world, Minecraft or Love or whatever, so the player knows what it is they can and can’t do in the world. I don’t want to be holding the player’s hand for the first three hours of play because that’s going to lose an awful lot of players. I want the players to think the cops are targets for the first ten minutes…” He finished the sentence with a laugh.
Stubbs showed me the only publicly-released part of The Hit, The Room. The player is confined to a single room, the layout of which is partially random. “None of this is generated on the fly but these portraits can be. I’ve got a system that will do that.” There was a window from which we could look out at the street and see NPCs marching around. I’m a sucker for open worlds. Looking out the window gave me this overwhelming desire to go outside and explore.
“As you’re moving around,” Stubbs continued, “things can change in the room without you looking at them. I’m testing out this kind of system because I want to be able to introduce story elements as the player is moving around without it being obvious that I’m doing so. And if you’re in a really crowded street and somebody walks up behind you, that’s a believable thing. The fact that you weren’t aware of them ten seconds ago doesn’t mean they weren’t there, it just means you hadn’t noticed them.” He pointed out something in The Room that changed colour when we weren’t looking. “I’m kind of wondering how many people actually played and noticed that anything was changing.”
The Hit isn’t unique in building a system that tries to change details while the player isn’t looking, but for Stubbs it seemed to be a critical component of his project. I confessed that I hadn’t seen any changes but observed there was some text on the corner of the screen indicating what the player was looking at.
The game is aware of what you’re looking at all times but Stubbs explained the game would have a system where objects were small packages of code, so that objects could be taken from one player’s world into another where such an object had never existed. I was reminded of Alternate Reality from the 80s which had a similar system of storing objects as code rather than a bundle of simple pre-defined attributes.
We got back to the NPC and contract-killing concept Stubbs wrote about on Gamasutra. “One of things I wasn’t sure about,” I said, “is that if you follow a person, they become a proper NPC and slowly gain more character traits the more you focus on them. At the same time, though, this is a game about contracts so I was just wondering… I wasn’t exactly sure how those two went together. You get contracts – but how does that relate to individuals being fleshed into NPCs?”
“The contract system is the simplest version of the game which is what I want to release before the end of the year. I want it to be something that’s enjoyable, that’s an entertaining experience, because like I said I want to build the city first and figure out how to make the dynamic narrative on top of that.”
I realised The Hit was just bait. Stubbs was starting out with something that felt like a GTA-type game to get players in the door. His real goal is a much grander project, a living, breathing multiplayer multiverse in which unscripted things happen.
There were still so many gaps in my understanding, though. I still couldn’t put together a picture of how this was going to work so I headed back to my problem with “emergent narrative” where developers hoped a game engine could completely replace an author. When games procedurally generate story, the result is usually coarse in structure.
I said, “If I need to collect an object then it’s skinned as “my mother needs this medicine” or “we need the special sword we’ve lost ten years ago”. It’s just a classic fetch quest that’s skinned. And then there can be other archetypes like “assassinate guy”. There’s a certain narrow range of mission types you’ve got to play with and I’m wondering how you will avoid an AI Director getting locked into “well I’ve got these mission types, I can reskin them then send them out into the world” because we’ve only talked about hit contracts in the game. But clearly there’s something more going on inside your head…”
Stubbs then talked about an event library that the AI Director can sift through. It would then throw relevant events in front of the player, trying to guide the player through certain story hoops. But eventually he admitted: “I don’t know how much I want to talk about this.”
The problem is these advanced aspects are slated for future years, so anything could change between now and then. Nonetheless, I kept pushing my concern that humans are very good at spotting patterns. How sophisticated does the AI Director have to be to fool players into thinking the world is real and anything could happen? Procedural generation is great for something like Quarries of Scred but to claim it can produce infinitely interesting variations in gameplay is a pretty tall order.
“I don’t want to make systems that are infinitely deep,” he said. “I want to make systems where it’s ridiculously fast to add new content to the game. I want to get to the point where one person can be playing it like a single-player game and five other people can be DMing the game and creating the assets and stories and characters on the fly.”
He then told me to stop recording for ten minutes so he could expand on how he planned to do that – off the record. I turned off the recorder and listened.
Well, the idea was interesting but without a prototype – the proof can only be in the pudding. We will have to wait to see if Stubbs’ ideas are the sound foundations for something interesting.
The recorder was turned back on and I addressed the issue of The Hit being bait. “You’re telling people what you’re planning but it’s incomplete information. I wonder if that’s actually a problem for you because you’re not even saying ‘I’m going to make The Hit’ but something more like ‘here’s The Hit and here’s some other things I am interested in’. Maybe you’re putting out lots of different messages.”
“This is partly a problem of being so close to it. Like for the Kickstarter, you can’t put out something nebulous. People will be put off by that, you have to go ‘THIS is the thing’ so I tried to do that.”
“Because The Hit is essentially phase one and that’s not actually what you’re interested in. You’re interested in phase 4 and 5 and The Hit is a way of getting there.”
So the Kickstarter failed. How to get to phases 4 and 5? “You’ve got the Patreon now as well, right?”
“I’ve got the Patreon but I haven’t been pushing that. I have a problem with taking money without giving something in return and I’ve… I’ve been depressed and incapable of putting words down. So I haven’t been writing the articles lately or blogposts. And that’s not just the Kickstarter it’s a whole bunch of stuff. A lot of it is the fact that I’m writing something so insanely ambitious that I’ve put myself through the grinder quite a bit. But it is actually coming together.”
“It’s difficult, really, isn’t it, when you talk about it a lot? From my own experience, once I start talking about bigger projects I worry that I’m just talking a lot. When I was younger I used to do that all the time, talk more than actually doing the work. And I think you’re in a difficult place where you need to talk for PR to get interest but at the same time you feel like a little bit like a charlatan because the stuff isn’t there yet.”
“Yeah, I’m avoiding talking about the stuff that’s a long way from being there.”
He walked me around a current version of The Hit and pointed out an office building with procedurally-generated interiors. In fact, the map was extremely small but kept wrapping around. He’d flipped his love for open-world games on its head. “Instead of making a game that has thousands and thousands of square kilometres, I’m just making something that loops around. So if I want to stick an event in the game, I can always put something in close to the player. Basically the whole NPC level-of-detail stuff is just bringing the game to the player. Making so that whatever you do, whatever direction you push it, the game will deepen in that direction like a fractal. The more you zoom in, the more interesting it gets. So instead of having the player climb all over the map, the content will always exist right around the player.”
Another red flag. I asked: “How is this actually going to work with a multiplayer server? Because if this is player-centric, how does it work with multiple viewpoints?”
Stubbs launched into another long explanation about how data is shared between players but I’ll admit I couldn’t quite figure out how an AI Director tightly focused on one player’s experience would work as multiplayer. Considering Stubbs’ love for Dark Souls and The Castle Doctrine, I guess it’s something like this: each player gets their little story bubble, but other players in the same city can easily puncture that and mess things up.
But Stubbs clarified that it wasn’t something he was building at the moment and everything was still up in the air. “It might just end up being a single-player game but it probably won’t because I find multiplayer really interesting.”
Stubbs went on to show me plenty of prototypes, such as his test which put 10,000 NPCs in motion, even though he doesn’t want to have that many characters on screen at one time. Lots of NPCs, though, became a sort of calling card for Stubbs’ work as can demonstrated in his zombie endless runner and, Go! Commando! Go!, a racing game which revels in being ridiculous.
But we had been talking for hours and, slowly, our conversation had begun to peter out. Less revelations and game design notes were making it to the surface and we called it a day. It was time for me to return to London.
Stubbs escorted me back to Plymouth station because I had absolutely no idea where I was. We talked a little about Stubbs’ past, his career history, but the interview, the conversation was over. The voice recorder had been turned off.
In the months since I visited Plymouth to see Stubbs, he has continued to work on The Hit and his Twitter account occasionally reveals a bit of leg. Yet I still do not know what to make of The Hit and whatever it is supposed to become. It’s a monstrous project that Dan Stubbs is clearly passionate about, but will it actually become something concrete?
Stubbs described his failed Kickstarter as a return to freedom, but that means he doesn’t have to ship any more. The design of The Hit is in permanent flux, with Stubbs’ ambition unfeterred by the pragmatism of just having to get something out the door to survive. Sometimes it feels there’s more focus on procedurally generating a city than the qualities that would distinguish it as a game. Sometimes I think I’m watching the development of Subversion all over again.
But perhaps it does not matter. Even if The Hit fails as an enjoyable multiplayer romp, perhaps it will light the way for other developers to get it right. Even if The Hit fails as a project, it will not devalue the conversation we had one day in August.