Replayability is an oft-debated concept in game development. Developers value it because we have the impression that players want it. The reasoning goes that highly replayable games represent greater value, and thus we should aim to make our games as replayable as possible. At a time when the cost to develop and market a triple-A blockbuster can reach $100 million, one of the best ways to make sure your money is well spent is to make sure the customer feels theirs is as well.
But what does ‘replayability’ even mean? The word itself implies an obvious definition: that the game can sustain player interest over the course of multiple playthroughs. Yet in a practical sense, data shows that players rarely finish our biggest games, never mind play them multiple times.
I think the above definition of replayability is an oversimplification of a couple of concepts that deserve closer scrutiny. A better understanding of these concepts will help developers increase the value they provide to players and avoid the kind of feature creep that can inflate development costs in the face of pressure to ‘add replayability’.
First, I think we need to dismiss the idea that replayability is a measure of how far players progress, or how many times they replay a game all the way through. Red Dead Redemption was highly replayable, but I’d be shocked to learn that more than 40 per cent of players got to the end, or that over 10 per cent got all the achievements. When faced with such numbers, it’s hard to refute the idea that the game isn’t replayable enough. So we need a more robust definition, lest we get strong-armed into making a worse game by someone with $100 million and a playtest report who says players aren’t finishing the game.
In a theoretical sense, I think the best predictor of replayability is the depth-to-exhaustibility ratio. By ‘depth’, I mean specifically the degree of rich interconnection between well-balanced systems. And by ‘exhaustibility’, I mean the degree to which a game relies on static content to deliver a message. The inherent claim of this simple ratio is that more systemic depth and less static content corresponds to greater replayability.
By this measure, we can look at classic games such as chess or go and predict that since they are entirely richly interconnected, well-balanced systems with virtually no content, they should exhibit almost infinite replayability. And they do. By contrast, a game such as Dragon’s Lair – virtually all content and no systems – should exhibit almost no replayability. And once you’ve exhausted the content and had your kiss from Princess Daphne, replaying is almost pointless.
In the modern market, where the depth-to-content ratio tends to be more middling, we can look at examples such as Civilization, which contains a great deal of content, but principally delivers it in the service of many extremely rich systems. As a consequence, it approaches the degree of replayability of chess or go. In the multiplayer domain, Civ (or any game) offers even higher replayability, because human players themselves become elements of the system space. With all other factors being equal, and barring degenerate strategies in the design, multiplayer games will always have higher replayability than singleplayer games. This explains the unfortunate trend towards tacked-on multiplayer as an attempt to mitigate the risk of low replayability.
So if all we need to do to increase replayability is to improve the ratio of richly interconnected, well-balanced systems to static content, why are we still floundering? This is particularly puzzling in an era when the cost to produce static content is rising non-linearly, and the cost to design robust dynamic systems remains comparatively fixed. But I feel there are two major hurdles to embracing this obvious solution.
First, even non-linear increases in the costs of producing static content are predictable. More people with more experience and better production methodology allow for predictable management of ever-increasing budgets. An easily exhausted, content-centric thrill ride that has a billion-dollar launch weekend and then fades from memory in time to start marketing the sequel is increasingly attractive when compared with the requirement to iterate through productive failures in order to create deep, interconnected systems. This is especially true when the methods for monetising years of continuous replay are poorly understood.
But developers also have a hurdle to acknowledge: our tools for delivering messages via static authored content are increasingly powerful – and seductive. After decades of yearning for cultural relevance, it often seems the easiest path to fame and wealth is to deliver authored messages via static content the way musicians or authors do. After all, there is no doubt you are culturally relevant when five million fans are arguing over whether you’re an ‘artist’. The question that arises as the replayability ratio approaches infinity is whether we’re willing and able to recognise the legitimacy of the culturally relevant act of abdicating the generation of meaning to players – which I believe is the highest value proposition that has ever been made.