The Making Of…
How two developers learned the importance of picking your battles
A two-man team learned the importance of picking your battles in the sublime Into The Breach
Format PC, Switch Developer/publisher Subset Games Origin US Release 2018
Hold up Into The Breach alongside its peers – the likes of XCOM, Phoenix Point or Mutant Year Zero – and what stands out is the game’s simplicity. Its tactical battles play out on an eight-by-eight square grid, adding up to a campaign that can be won or lost in the time it takes to complete a single mission of XCOM. At first glance, it might not seem like the kind of game that would take over four years to make. But Into The Breach’s simplicity was hard-won – not least because, for most of the development process, its scope was much wider.
In 2014, the two-man team of Justin Ma and Matthew Davis, collectively known as Subset Games, were still recovering from the unexpected popularity of their debut game, FTL,
which had gone from experimental project to early Kickstarter success story to IGF winner in the space of 18 months. “In retrospect, it came together at a blistering speed,” Ma says. “The core game idea was kind of in place in like two or three months.”
That success and speed, however, came at a cost. “Coming off FTL, we were both pretty burnt out,” Davis says. “We weren’t really in a hurry to do anything new, which is why we did Advanced Edition at a more casual pace after finishing FTL. And even after that, it was still a good four or five months before we started anything new.” The pair weren’t sure they even would make another game. “We had to feel like we could cancel it at any time,” Ma says. “If Matt decided he hated working on games, then we could just step away.” This was part of an effort to recreate the happy-accident design space of FTL, while also trying to create a more comfortable development cycle than the 18 months they’d just endured.
Eventually, Davis landed on an idea that seemed worth pursuing: a tactics game, inspired by the many hours both developers had poured into Firaxis’ XCOM reboot, but also by mini-Roguelikes Hoplite and Desktop Dungeons.
“No randomness; interesting meaty design that’s really elegant and clean – I thought it’d be really interesting to try and make something like that,” Davis says. “I didn’t realise just how hard it was going to be.”
As his idea started to take shape, Davis drew further inspiration from an unexpected
source: 2013 Superman movie Man Of Steel, infamous for a final battle where its hero demonstrated a troubling lack of regard for human life. “Entire cities would be destroyed, and it’s just background noise, and no one really seems to care that a few hundred thousand people probably just died,” Ma says. “Matt wanted to make a game where protecting the city was the highest priority, where you’re willing to sacrifice yourself to save the city. That
was the core design pillar. Although it doesn’t say much about the gameplay, it’s a feeling that we wanted to evoke for players, and we used that to guide the design.”
This feeling helped Subset land on the game’s theme – mechs versus kaiju seemed like the most obvious fictional example – and its broader design, such as the idea of telegraphed attacks. Each kaiju shows what it’s going to attack next turn, giving the player chance to counteract it. “That was a kind of fumbling discovery that we had relatively early on,” Ma says. “It would be a long time before we decided to focus on that as an actual design pillar, but we noticed that it was interesting – a design space I hadn’t really seen
explored.” It created a combat system where everything was perfectly predictable, without the dice rolls of XCOM – or indeed FTL, which had been criticised for its randomness.
As the turn-based combat started to come together, there was still the question of the other half: the strategy layer that would link each battle and, hopefully, give players a reason to care about the buildings they were protecting. “We fell into the trap that I think a lot of games fall into,” Ma says. “It’s really hard to make, basically, two games at once. Two games that can stand on their own and yet also complement each other.”
Subset worked through multiple prototypes, often borrowing from boardgames the pair were playing at the time. One version of the game was influenced by Pandemic, with Vek infection spreading between nodes on a map. Another was more akin to worker placement games such as Agricola. The focus was tightened down to a single city for the player to manage. “Over time that gradually kept shrinking, to be smaller and smaller. For a while there was just a single city where you have four battle maps,” Ma says. “We spent a year on this prototype, where you repaired buildings over time and you’d choose when to engage with threats as they appeared.”
Asked what the problem was with each prototype, Davis’ answer is always some variant on the same phrase: “Ultimately, it just wasn’t fun.” As the self-described pessimist of the duo, Davis would always be the first to acknowledge something wasn’t working, but every time Ma agreed with his assessment. And so the pair scrapped months of work, over and over again. “We would try something and then it would fail, and we would try something else and we’d fail,” Ma says. Where the core of FTL had come together in a couple of months, it took Subset two-and-a-half years to even figure out what this new game would be.
The promise the pair had made to one another in the early days of development – that they could always just step away from the project – looked increasingly tempting. “We often discussed dropping it entirely,” Ma says. “I don’t know if we were close to pulling that trigger. But it was definitely on the table.”
“WE DISCUSSED DROPPING IT ENTIRELY. I DON’T KNOW IF WE WERE CLOSE TO PULLING THAT TRIGGER.”
So what stopped them from canning the project entirely? “Maybe a bit of a sunk-cost fallacy,” Ma admits. But there was also a sense that this strategy game idea, nebulous as it still was, had something special about it. “I think if one of us had an idea that sounded better and more doable, we probably would have dropped it. But neither of us did. We had lots of dumb small ideas, but nothing that was so obviously clear.”
Around this point, the pair did briefly step away from Into The Breach to explore one of these side ideas. They tried prototyping an endless runner game instead, “just to explore some new mechanics and have some fun”, Ma says. “That was harder than we expected, and maybe it pushed me back into thinking, ‘all right, let’s just finish this thing that I feel like we can do’. That was a bit of an impetus to try harder and actually finish it.”
They returned to Into The Breach with a new philosophy, one that Ma neatly summarises as: “Screw it”. Or, as Davis puts it: “We cut out all the crap that didn’t work.” Everything which wasn’t working – which wasn’t fun – was immediately tossed out. The strategy layer was jettisoned entirely. Instead, players would simply pick their next mission from a list of three. This put the focus entirely on the one aspect which was working: combat.
Between bouts of wrestling with the strategy layer, Subset had also been working on the game’s other half, and combat had developed nicely from that early concept of telegraphed attacks. There were still issues to be ironed out, especially when it came to UI. (“So many times, we put the game in front of a playtester and asked them, ‘What’s going to happen?’ before they click shoot,” Davis says. “And they would get it wrong. Every single time.”) But as every ounce of fat was trimmed away – no more management of cities or multiple squads, no more cannon-fodder support units, no more passive abilities – they finally had a core loop that felt polished, unique and, ultimately, fun.
“Frankly, if we made that decision earlier, we would have saved years of effort,” Ma says. And the pair had no qualms about going back to the drawing board and erasing past work, he says: “As long as the game that you have left is good enough, it feels wonderful to just cut away at everything.”
From this smaller base, Ma and Davis were able to build outwards. This, they both agree, is how Subset does its best work. “We made something as small as we could possibly make it, and then we just kind of tacked on the features that we thought were absolutely necessary for the rest of the game,” Davis says. They added small pieces to the design, ready to drop anything that wasn’t perfect, or at