How important is failure to the enjoyment of digital games? I contend it is the central issue in designing for an audience, since players who want to strive against impossible odds and eventually triumph must fail in order to enjoy their success, while the vast majority of mass market players will tolerate only a modest degree of failure as part of their play experience.

For many years I have advocated attention to the issue of whether gameplay should be fail-repeat – as the old school arcade games always were – or fail-continue, allowing the player to proceed even if they can’t master a particular challenge. It’s taken a long time, but the industry is finally catching up to my argument that if you want to reach a wide audience, you need to offer them tools to prevent the bottlenecking associated with fail-repeat. But there’s a cost – because fail-repeat is also important for players who are triumph-seekers, and who need to strive against impossible odds to get their eventual reward.

In this piece, I review the issues of balancing fail-repeat against fail-continue, specifically in the context of the Air Conflicts games, where I have been lucky enough to be able to experiment with new game structures for better management of frustration.

Infuriating Failures

Recently, I was playing Nitrome games’ Canary (at the suggestion of longtime ihobo stalward, Roman Age) and thinking about the role of repetition in gameplay, as I often do. Canary has a great central mechanic involving using a laser to cut through rock, which then drops down and can be pushed about. But it has a fixed rate of scrolling, so failure often happens as a result of bad screen positioning. This in itself would be tolerable, but alas failure means you start the entire level at the beginning, without any check-pointing, and this caused me to rapidly lose interest in the game.

Fail-repeat has long been a topic of interest for me – see this old post of mine from 2005 on Ratcheted Progress, for instance, which although rough around the edges is still quite pertinent. I have long advocated a fail-continue structure in which the player is not required to repeat gameplay sections as a vital tool for reaching out to a mass market audience – and I have often been met with incredulity and derision by publishers and developers stuck inside gamer hobbyist thinking. I wrote about this in 2008 under the title Freedom to Fail. Last year, I was thrilled to find no lesser a games industry celebrity than Miyamoto-san endorsing fail-continue in the design of New Super Mario Bros. Wii, implementing features I’ve been experimenting for many years.

It is important to recognise that for an important minority of gamer hobbyists, failure is a vital element of their play. Such players are not usually conscious of this circumstance – although they may recognise that what they enjoy is the intense states of emotional arousal games can illicit in them, characterised by heart-pounding excitement and furious anger that motivates continuing play. I have linked the role of anger in perseverance to testosterone, on the basis of recent neurobiological research, and am confident this stance will be validated. But I have written considerably less about the intensity of excitement that often (but not always) goes hand-in-hand with the anger.

It is well known that there is a relationship between risk and reward – gambling thrives on this coupling. Real time digital games also thrive on risk-reward relationships; to generate strong emotional responses from players it is not strictly necessary that they fail provided the cost of failure is sufficiently high. This is why “permadeath” (when failure has permanent consequences) creates such an extreme play experience – the risk is perceived as gigantic. It is also why permadeath will never be an especially popular game mechanic, since these extreme costs for failure radically winnow the possible players to a very tiny pool of diehards.

The main cost of failure that players are “threatened” with is repetition – complete this challenge, or you will have to do it all again. There is a fine line to be walked here, because failure is always frustrating to some degree, and frustration (being an experience of anger, connected with the neurotransmitter norepinephrine) is a cumulative emotion – the more frustration, the angrier you get. Become too angry, and you stop playing. This is where testosterone comes in, since one important effect of this chemical is to increase tolerance to anger and foster perseverance, allowing testosterone-infused players to endure frustrations that seem incomprehensible to many observers.

Because the risk of repetition is used to enhance the reward of success – to create the experience of fiero, triumph over adversity – there is an important role for failure in play. But because perseverance in the face of frustrations is a minority interest, related partly to testosterone, games that leverage these kind of mechanics narrow their audience quite significantly. There’s a reason that Nintendogs can sell 24 million units – four times as many units as Gears of War – and testosterone-moderated tolerance for frustration is an important part of the story of why the latter game appeals to a much smaller audience. The fail-repeat gameplay that violent action games depend upon is both the reason for their success, and the reason that the ceiling for their success is always radically less than it is for a true mass market game.

Freedom from Failure

The alternative to fail-repeat is what I have termed fail-continue structures: failure does not end the game, the game simply continues. There may be rewards for succeeding that are not won, but with fail-continue the player’s progression through the game is not linked to succeeding at specific tasks. This has become an absolutely vital force in commercial game design for the mass market, as the lineage of games that passes through Harvest Moon and Animal Crossing right into the heart of the Facebook “farming” games all attest to the increasing commercial relevance of game structures where failure is either a trivial or a non-existent element of play.

One of the main places I’ve been experimenting with fail-continue is in the Air Conflicts games, which are made by the development team 3 Division, a part of Slovakian developer Games Farm. The first of these games, called simply Air Conflicts, was my first successful experiment with fail-continue, using an approach similar to early games such as Defender of the Crown: failing in a mission is part of the story, and the narrative continues regardless (achieved in Air Conflicts by having a squadron of planes that is restocked as the player succeeds, so losing a plane is meaningful). In the second game, Attack on Pearl Harbor, a compromise was reached between the triumph-seeking player’s desire to strive to overcome and the mass market desire never to be bottlenecked, and the fail-continue feature was simplified to something very simple: the choice, after each mission, of whether to replay it. Want to beat it, and earn the emotional reward of fiero? Try as often as you like. Don’t want to? Just move onto the next mission.

The beauty of this mechanic is that it leaves it entirely up to the player how to approach their play. If you are what I have termed a Conqueror style player, you will want to repeat the challenge until you win. But if you aren’t, you might prefer to move on rather than get stuck. Offering a choice to replay or continue after each mission failure provides a simple and elegant choice, one that we have used again in the third game in the trilogy, Air Conflicts: Secret Wars (pictured above). For this one, the idea of a hanger-full of planes has returned, so accepting failure means you can’t fly the plane you just crashed until its repaired, a mechanic that works surprisingly well to mediate replay of specific missions. There is also a limit on the number of skips per chapter, a requirement since if you don’t complete enough mission objectives you don’t earn new planes and thus find yourself at a radical disadvantage.

A new style of fail-continue was added to this new game. The story of Secret Wars involves fighting alongside the resistance movements of World War II, using real battles and events – often with very depressing outcomes. At the end of each chapter, a flashback reveals some of the backstory by using some equally depressing battles set in World War I. During the flashbacks, a narrator character – one of several pilots who flew in the Great War – tells their story, and the events are echoed in the play of the mission. Because it’s a flashback, it makes no sense to fail, so if you crash or are shot down the narrator remarks that “it didn’t happen like that…” and you take to the skies again. You can’t fail these flashbacks – they serve the narrative role of a cut scene, but the player is in control of their plane throughout. I see this as a really effective storytelling technique, that also happens to be fail-continue.

Alas, when all is said and done, the Air Conflicts games aren’t going to enjoy astronomical commercial success because they are at heart arcade flight sims – I hope that more players will discover these great little games, but I know that there is only a limited audience for any game requiring three-dimensional control skills (such as a flight sim – even an arcade-style flying game). However, I hope that my experiments with fail-continue structure here will serve as an inspiration for new possibilities in other genres – and even if no-one else pays attention to what we’ve been doing with these games

Conclusion

Players who want to be driven to the edge of their limits will always be an important part of the market for digital games – and for the most part, this audience will always be comprised primarily of teenage boys, not uncoincidentally, those who are in the grip of significant swings in testosterone levels. But as the rising cost of development on the power consoles has sky-rocketed, the number of franchises that can compete for this audience has been radically reduced. There is almost no point in entering this arena now unless you have the resources to compete at the level of, say, Call of Duty. Not making the grade (for instance, Brink), is a costly business, and can even result in studios closing (although I believe Splash Damage will live to fight again). For new companies, targeting this market is suicide.

Where there are still opportunities for new game developers is in the mass market for games, and one of the key ways in which game design can help open up new markets is by finding new solutions to the problem of failure. The wider audience for games is not interested in failing, or at least, is not willing to tolerate a high cost of failure (of course, some failures are very low cost – fail at Bejewelled, and you lose very little, and generally want to play again right away). Because of this aversion to repetition, frustration and costly mistakes, exploring new solutions to the problem of failure in games has the potential to be highly profitable.

Fail-continue structures are one way of reaching out to this wider audience – and there are many different ways to mount this kind of  mechanism, most of which have never been tried. Some may prove to be the foundation of entirely new game genres, and those developers that discover these untapped niches stand to make astronomical profits. Perhaps I’ll be lucky and be involved in one of these games, but even if I’m not, I’m still glad to have been ahead of the curve on the role of failure in games, and, over the course of the last five years, to have been proved correct in respect of the commercial importance of freedom from failure.