Cheating Death: Accommodating player failure and recovery

At its inception, player death was something of a necessity: a way for the arcade cabinet to enforce the collection of quarters. Life was a commodity, and if you wanted more of it, you had to pay. As games moved out of the cabinet and into the living room, that idea persisted, first through the direct ports of the arcade games, but then as an established convention. The “Game Over” screen was a staple of video games, and it’s been tenacious.

As with most conventions, it’s not questioned nearly enough. When accepted as essential, player death robs the player of the chance to recover from failure, to push through an unwanted or unfavorable situation to reach a satisfying conclusion further down the line. As an industry, we’ve attempted this narratively, with player choice being increasingly pushed to the forefront of story-focused games, where bad decisions aren’t punished, but supported just as much as the good. However, mechanically speaking, we’ve been far more stagnant -- if only because conceptually the alternative is that much more difficult to comprehend.

“In a game where killing is a systemic verb, I feel like death is certainly [important],” Clint Hocking answers when I ask him if player death is essential. “You can certainly create a contrivance around it; you can have an infinite number of buddies who rescue you in Far Cry, or you can be possessed by an elf wraith from a thousand years before, and you can increment forward on failure.”

With that latter example he’s referring to the recent example of an alternative to player death being a fail state in Shadow of Mordor, where, narratively, the player dies at the beginning of the game. Any subsequent mechanical deaths are accepted by the game, and in some ways encouraged, as it progresses time by three days and allows Mordor’s procedurally generated orc armies to reshuffle and regroup.

Both Far Cry 2, of which Clint Hocking was the creative director, and Shadow of Mordor offer an attempt to subvert the mechanical convention of player death resulting in that disruptive “Game Over” screen. The former responds to the first player death after a save point by having your "buddy" (a fellow morally ambiguous mercenary) come in and rescue you, dragging you away from a firefight and plunging a medical syringe into your chest, getting you back on your feet and back into the fight. The caveat there is that your buddy is now in the field of combat, and suddenly an actor in it; they fight, and they can die, resulting in them being systemically removed from the game.

“Now you’re playing for real stakes.” Hocking explains. “When you’re playing a game and you die, sure, you have lost some time, but with most games today you don’t even get reset more than 10 feet back. With this game it was more to say that now there’s a real stake. Now this buddy rescue has happened, now your ante is on the table. Now, if you don’t deal with that situation… you can just run away, but if you do that your buddy is sacrificed, they’re gone. You have skin in the game at that point, because now it’s a gameplay asset, and hopefully it’s more than that, hopefully it’s a friend and ally, that is going to be lost if you don’t deal with the consequence of your failure.”

And that idea of consequence is at the heart of what’s lost when the player-death-as-failstate is followed. While it flirts with the borders of the concept of ludonarrative dissonance, more importantly it interferes with the flow and impact of the interaction between mechanical systems and narrative context. It’s telling that Mordor struggled itself during a few of its more story-driven missions, taking away the immortality of the player in favor of more actively controlling the way it told its story.

 “I’m certainly not saying we found the ideal solution for [player death],” says Michael De Plater, director of design on Shadow of Mordor. “It’s something we’re thinking about a lot moving forward. The way in which death is treated is a key part of the style and tone of the game. We had a goal which was to keep time moving forward and create new opportunities and a motivation for revenge against an enemy who took you down. I love the trend for roguelikes now, and how they handle death and failure.”

Roguelikes, for their part, have certainly popularized the idea of death and consequence, albeit it in a way that is less engaged in telling an authored story, but rather allowing one to emerge out of systemic interactions. The likes of Spelunky, Rogue Legacy, and The Binding of Isaac all are fully aware that death will be frequent and necessary for progress, and build the games around that concept.

Fundamentally, this is where the crux of the problem lies, in the tension between authored and unauthored game time. The more control the game tries to exert on any one moment, the more rigid it becomes in regards to the possibility space. The more systemic and unauthored, the more it can support and cushion failure and failure states. One of Far Cry 2’s greatest strengths was in building failure into the experience, from the jamming weapons, the malaria attacks and the propagating fire, to the buddy system and the way it embraces consequence.

According to Hocking, this was present in his design ethos going all the way back to the original Splinter Cell. “I fought as a level designer really, really hard to say 'I think all of these “Game Overs” you get from being detected by a guy, or failing to hack into a computer, are all terrible, and really frustrating experiences.'

"If I’m Sam Fisher and someone detects me I don’t just throw up my hands and say, 'Oh well, I guess we can’t prevent the war in Georgia, the nuke’s going to blow up and everyone is going to die.' I probably try to kill that person and adapt to the situation. The reason they were there was because as designers and developers we couldn’t support the failure cases. We didn’t have workarounds for what happens if the player gets detected in that situation. So we just had to gate them with Game Overs.”

The difference between recovering from being detected and recovering from being "killed," though, would seem to be the reason one has been consistently rectified and the other hasn’t. We’ve reached a point where instant failure in stealth games is mostly a thing of the past, most exemplified by Dishonoured’s generous possibility-scape when it comes to raising alarms and adapting to situations. But importantly, the difference mechanically between one and the other really isn’t so large. Instead it’s a conceptual problem; we can imagine, instantly, how Sam Fisher would react to and recover from being discovered, whereas we can’t do the same when he’s shot and killed.

One of the oldest and most effective attempts to solve this problem can be found in Planescape: Torment, where the player character begins the game in a morgue, apparently having recovered from being dead and cold. Throughout the game your immortality is emphasized and even used as a narrative device, memories and facts from your past bubbling up to the surface in the desperate recollection in the moments preceding death, life literally flashing before your eyes.

“With Torment, there was the added complication that one of our goals was to “tell the story of what happens after the death screen,” Chris Avellone says. He was lead designer on Torment. "And in the context of the story and the arcane physics that caused the Nameless One to be immortal, a perma-death didn’t make any sense either -- instead, it seemed more appropriate, narratively, that the Nameless One would wake up somewhere else (technically a time cost).

"But then, I got excited about the possibility that it could be used as fast travel ('Hey, I need to go back to the Mortuary fast, or back to the merchant area') and even a puzzle element, and could take you to new locations or allow you to get through areas mortals couldn’t, as long as you allowed yourself to die (which we didn’t use very much in the game, unfortunately).”

The result of this creative approach to player death meant that, once you accepted the conceits and  contrivances of Planescape: Torment, there was very little to get in the way of you experiencing the game. While there were still a few "hard" fail states, they were few and far between, meaning that the majority of the time the experience was unadulterated by the "gates" that Hocking mentioned.

However, that doesn’t mean that Avellone believes that death and failure states need to be removed from games. “I don’t consider the divide between player death and story to be a bad one, and in fact, adds adrenaline and urgency into a situation that may lack drama when you know that there’s no real fail state. Furthermore, it’s an understood mechanism of the genre, especially for RPGs. I do feel there’s more interesting things to do with death that sometimes narrative can give context to (it sure did for Torment, and I believe it will for inXile’s Torment: Tides of Numenera as well).

“There needs to be a consequence and a fail state -- if there’s nothing to lose, then a lot of narrative tension dissolves as a result. I think there are other consequences you can levy on a player (and many games do) but you have to be careful in how they are presented or else they will simple cause a reload as a result.”

One interesting case of "other consequences" is actually found in the Mass Effect series, at the end of the second game. While combat in the Mass Effect games isn’t trivial, it can be fairly unthreatening, especially on the lower difficulties. But at the end of Mass Effect 2, BioWare instead made it about your narrative and tactical decisions, rather than your performance in any one fight. Suddenly, companions that you had spent dozens of hours getting to know where exposed and vulnerable based on decisions you made -- decisions which were far enough back that a simple reload would revert hours of progress. They were essentially leveraging your patience for narrative strength, and it paid off extremely well -- even if it was hard to swallow when your favorite character bit the bullet.

“I think a lot of players are raised on games that are pandering to the power fantasy.” Hocking says, when I ask why we haven’t seen more experimentation in regards to player death. “I think when you try to do things that are more challenging than that, or more questioning of that -- or undermining it, playing it against itself -- you will get negative reactions.

"And you get negative reactions almost universally at first, so the question is, what percentage of the players are willing to suck it up for half an hour and realize what is going on? I think once you get over that hump, though, the possibilities of what your game can be and what your game can say is astronomically larger. It means you can be more than just a power fantasy, which is great.”

The rise of the Souls games, currently culminating with Bloodborne, is certainly encouraging when it comes to the idea of a developing player understanding of what death can be in games, as well as a deconstruction of the power fantasy so often extolled by them. While the attitude towards death in those games is certainly one that won’t work for the vast majority, it is an alternative that’s taken root in the minds of the audience.

Far Cry 2, for its part, could certainly be argued as a dis-empowerment fantasy, with the increasing moral ambivalence of the player character and the world around them echoed systemically as your plans are driven awry by the constant interference of its systems, from out of control blazes to weapon jams, to a range of other interactions.

After so much discussion with the people behind some of the most successful attempts to alter the convention of player death being a fail state, I realised that death is more of a symptom that is caused by an inability, whether financial or creative, to create the systems necessary to accommodate player failure and recovery. It’s no easy task to fundamentally change what is expected when the health bar goes down to zero and it all goes dark, not least because of what has been the case for so long.

But to at the very least question it, so that checkpoints and “Game Over” screens aren’t in your game before there’s even been a discussion, is the way that more interesting and compelling interactions will be pursued and realised. Even something as simple as Prince of Persia: Sands of Time’s rewind mechanic was a revelation at the time, allowing players to manipulate a system to reattempt a difficult jump or encounter without the brusque interruption of a “Game Over” screen.

The thing is, the game is going to go ahead, whether the player dies or not. No matter how many times they see “Game Over”, the game isn’t over unless they decide to turn it off. Otherwise they will just keep pushing ahead, wrestling with that screen until it doesn’t appear, and the "canon" version of the story is cemented by a save point. Now victory is assured, always was assured, and any failure is relegated to the narrative beats that were set in stone before the player even started. Letting go of that adherence to victory, to success, is going to, and has, made things a lot more interesting. Failure, and recovery from that failure, is fun.