Starfield’s mixed reception underlines the increasing challenge of triple-A game development

It’s getting harder and exponentially more expensive to make big games. We know this, because game developers keep telling us this. And because we can see the evidence all around us, in hugely anticipated games that either hobble out of their studios beset by bugs and missing features, like Cyberpunk 2077 at launch (it’s much better now) or in games like Starfield that simply don’t inflame imaginations and attract the same unanimous praise like Bethesda Game Studios’ previous titles did.

It isn’t because the developers working on these titles aren’t working hard enough, or lack talent or passion. Quite the opposite. In order to achieve the scale, the mechanical depth, and the visual fidelity that passes for roughly par in a modern triple-A gaming, multiple studios with headcounts in the hundreds need to work together across the globe for years, managing extremely complex pipelines.

This isn’t a new phenomenon. There were rivets popping out of the old triple-A structure in the pre-COVID days when Mass Effect Andromeda and Anthem released to rapturous indifference, both games carrying the feel of something designed by too large a committee and with nothing fresh or incisive to offer as a result. Since it costs an exponentially higher amount to produce games like this with every passing millisecond, the execs who bankroll them want to ensure the broadest possible appeal. And if they’re not careful - which often they’re not, of course - they end up shipping a product so generalised and inoffensive that it appeals to nobody.

Starfield isn’t that. The first wave of reviews contains some effusive praise for its depth and richness, the kind of world-building and storytelling that BGS built its name on, and the delivery of its promises about epic scale and spaceship battles. There’s a winning blend of traditional BGS RPG and grand space exploration in there that can’t be accused of playing it too safe.

But the early reviews also contain some considerable dissent. Some critics have actually called it too ambitious, arguing that the execution of its thousand explorable planets doesn’t live up to the concept. Others lament the emphasis on quantity over quality, an uninspiring stretch in its opening hours, and the way fast-travel robs space traversal of its wonder.

It’s not a failure, by any measure. In fact, against the backdrop of commonplace disastrous triple-A releases, it’s remarkably robust for an RPG with such scale. But right now as the first reviews trickle in and Starfield sits on a PC metascore of 88, you wonder if it’s going to have the same craterous cultural impact that Skyrim or Fallout 3 had. There’s a sense that, even in the hands of the absolute masters, maybe this is simply as good as you can make a traditional triple-A game in the current climate, with all the challenges such a project faces.

Because if BGS can’t replicate that Skyrim kind of success in 2023 with this game, it’s hard to see who else can. Baldur’s Gate 3 has been a tremendous success for Larian this summer, but it’s not really the same kind of RPG. It’s all about systems and agency, malleability and rolling with the player’s whims. It’s not going for Hollywood sheen. 

Maybe that’s why Baldur’s Gate 3 has been received with warm hugs and teary eyes - because it doesn’t fall into the triple-A production values trap. It was in Early Access for an extended period, and its studio spent time and money on refining systems, rather than pushing the pixel fidelity higher. It spent a long time ensuring that all its component parts actually worked, that they were fun, and that the player understood them, all in a live testing environment where a huge player base could offer constant feedback. 

Might this be a turning point for triple-A, then? Starfield and Baldur’s Gate 3 offer two contrasting models of how to achieve a great RPG that seeps into wider culture and that people end up playing for weeks, months, and maybe years. And right now it’s Larian’s nu-school thinking that seems to have worked best. 

Imagine this: what if Starfield’s 6 September launch actually marked the beginning of a year-long Early Access phase? What kind of game might we get in 2024 afterwards?