Death is Imminent

Why did I enjoy that game? I really have no idea. Almost no multiplayer (hey, leaderboards matter) and not much sense of achievement. A platformer full of raw punishment for any deviation from perfection. Back then, my parents likely wished I didn’t waste so much time and money in the damn thing. But somehow the family finances survived. The investment might even be considered to have paid off, in the fullness of time.

Anyway, unless you’re one of the tiny number of people making or porting coin-up games, your revenue doesn’t dribble in with every two bits fed into a slot. Today, we don’t have any financial incentive to kill our players. In the abstract, a player wants an enjoyable experience for his or her fifty or sixty bucks, and death can only disturb that. Even if we take it for granted, it’s the most obvious break in the connection between player and character. And yet… our games rely on conflict. And thus most obvious form of conflict: fighting. So, death, death everywhere. I wonder just how many avatars my failures have sent tumbling down to the virtual underworld. It has to number in the tens of thousands, at least.

Now, I know there are some good reasons why we slaughter our virtual player personas. One, habit. Two, the fundamental logic of combat-focused games. More meaningfully, we need an instrument to measure failure, and to allow for game difficulty. Players want a challenge, or else they will become bored (FarmVille notwithstanding, apparently). Death is simple and easy to understand, and so death is our default penalty for failure.

It doesn’t have to be. Historically, games don’t simulate death. Consider: sports and sport games don’t. Card games don’t. Strategy and puzzle games don’t. These are the real origins of what we call games, even if we can’t ask their designers any questions, if we could even identify who they were. Losing just meant losing: someone else got more points, or got to the finish line before you. “You” didn’t die. We can track the prevalence of death in games to the relatively recent phenomena of controlling a virtual character.  We’ve developed this habit in less than thirty years, since Space Invaders, etc. (Senet is over five thousand years old by comparison.)  Logically enough, I think the simulation of death in games can be traced to the widest permutation of the term “roleplaying” – any situation during which the player becomes embodied.

Most players haven’t gone aboard the death train that you (likely reader) and I accept. Truly popular games – stuff on the web, on Facebook, the Sims, etc. – don’t really involve player death. And certainly doesn’t simulate it. As much as I love Modern Warfare, Battlefield, and Splinter Cell, given the opportunity to think about it, I’d like to find a way out of the death trap. Many games of the day no longer bother to explain death and respawning. The player’s character dies from too much damage, and we restart him at a prior checkpoint. It’s a game, and we’re so used to the mechanic that the designer doesn’t deign to explain the logic of reincarnation or time travel.

Speaking of which, another option is to adopt a mechanic to “avert” death. In Sands of Time, our protagonist announces “that’s not how that happened”, and instead of suffering a mortal wound, we rewind back in time. Similarly, in the more recent Prince of Persia, Elika intervenes when failure would result in the hero’s demise. There’s no gameplay difference in either case between that and killing the player and respawning, but instead of complicating or disturbing my relationship to the game and its characters, the system behind those two games reinforces my connection. Planescape: Torment, Assassin’s Creed, and Bioshock have their own answers to evade the reaper, though they weren’t without complication.

The impossible design challenge would be to take a core player experience -  a “realistic” shooter, most obviously – and then turn around the game’s failure system to remove death from the equation. The obvious thing is to hack the system with magic, imaginary worlds, simulations within simulations, or other tricks. And I think we’ll continue to do so, to take away the sting of the player’s failure. Hopefully, without annoying the player so much that he hates the innovation so much that he’d rather resort to a fast reload to checkpoint, logic be damned. Stepping outside that box would probably mean ditching a health-based model completely. Stealth games do that, but not that many of our action genre games step outside of health as a core feature.

Neither Boethius nor I have an unambiguous conclusion, much less a prescription. The closest I’ll come is: Don’t take death for granted. Just because it’s the default failure condition of the day, doesn’t mean it has to be.

The player has paid you money to play the game. Let him pick what difficulty he wants, when he wants. If this is so hard to alter dynamically, let him restart at the last checkpoint with the settings he or she wants.

A lot of design decisions from the original God of War have survived. The "You Have Died" screen is one. Another is the fact that you can't change the difficulty post-start, except by means of the most insulting means possible. Why?

I like success. It's like cake. But I can't learn anything from cake. I can learn from the stack of broken crates, though, and the other failures.

Painful lessons seem to be the ones that teach the most, whatever the subject. Pain provides motivation to find a path that leads to success. I remember the first time I waved a bat as a baseball sailed on by. It was an ugly swing. And my coach let me hear it, in front of all my friends. In fact, my coach would say my swing never was pretty. But his public hazing did make me work hard at improving it.

When it comes to video games, difficulty will always be a charged issue. Pressure comes from all sides: to appeal to a wide audience, to appeal to the hardcore, to appeal to casual player. Whatever those classifications mean in the real world (maybe nothing), game difficulty is something everyone has an opinion about.

I don’t pretend to know how hard you should make your game. Recently, though, I’ve come to think that as game designers, we should be playing with our dials turned up.

How did I get there? When playing Mass Effect 2, I started with a medium setting (“Veteran”), and played through the game rather happily and easily. I predictably found that I was playing the game for its story content, not for its gameplay. Shooting just filled the time until the next cinematic or dialogue moment. I wasn’t enjoying it especially, and I certainly wasn’t learning anything from it. Enemies could be defeated with just about any weapon, tactic, or character ability, and I could engage them head on, ignoring the cover mechanic and succeed. Of course, when you can defeat your foes any way you want to, the decision of how to win becomes empty of meaning. Empty of lessons to learn.

After finishing the game, I started a second playthrough. This was my “evil bitch” experience, and wow, is the female Shepherd a superior voice actress. This time, I set the difficulty to its highest setting — “Insanity.” And in terms of the moment-to-moment, I discovered a whole new game. Enemies, even stock ones, posed a major threat. My allies needed to be managed and controlled lest they die — repeatedly. In order to survive myself, I had to find weaknesses in my opponents or vulnerabilities in the level design. Each combat formed a battle combining ally management, refresh management, and a search for anything that could give me an edge. Some of these elements (vulnerability to ammo types or to certain powers) were intended by the game designers. Some (an inability to deal with player traps in pathing) were not. My tactics bordered on exploits. In any event, I learned what my character could do, what each AI was capable of, and how I could win in different encounters. In short, I knew the game far better than I did before.

The question of “is that an exploit or a good tactic?” is a good one, especially in online games where your play can affect others. I remember early days of MMO play, when we wondered if reverse kiting, sending tells to determine spawns, or FD-camping monsters represented an exploit. EverQuest supported lots of unforeseen play, as a result of the game’s steep difficulty curve. Soloing in specific stood out as ridiculously hard. Players had to find a way to avoid the hard combat math the game’s designers had forged: they had to find a way not to get hit. Without finding gaps in the system or the vulnerabilities of the AI, the player couldn’t succeed on his own. Back then, of course, the game’s harsh penalties of failure and death didn’t stick out like the sore thumbs they would today.

More recently, on a dare I took to playing Darksiders on its hardest setting, appropriately named Apocalyptic. Twenty-odd hours later, more than three thousand gallons of blood has been spilled beneath my blade and scythe (according to this). And because I played the game at such an unforgiving level, the average hit from a demon shed half of my characters health. Learning the game’s rhythms, the pace of when to attack and when to evade, was the difference between life and death.

Losing fifty percent of your character hit points with a single blow sounds painful, doesn’t it? What I found though, was thanks to the mild penalties (fast reloads, frequent checkpoints), I relished the challenge. More importantly, I came to know what defense I had to employ against each AI type, who to kill first in group encounters, and when in each encounter I needed to fire off my panic and powerup buttons.

I’m not sure that “hard is the new good” when it comes to player experience – certainly not for all games or all players. But for game designers, I think it’s invaluable. If you want to play a game at its most pure — not for the story, not even for a good time — tune the difficulty up. Make yourself defeat the game where the margin for error is razor thin. You may learn something.

There’s a lot to like. Collecting and side missions slowly capture map sections and generate the feeling of real progress. At times, the navigation play on city wires can drop into a enjoyable rthyhm. Then there’s some game systems that need help.

• Hivemind AI. Once an enemy has spotted you, they all know where you are, pretty much permanently. Run away from one, then come around a corner, and another enemy will shoot your eyebrows off just as you become visible. The game should be hiding more information from its own AI.
• Hyperaccurate AI. Never seen a game that tried so hard to kill me. Enemy snipers on every rooftop. If you’re not tumbling, they will hit you.
  
  

  I said hyperaccurate!

• High Durability AI. The punks on the street take a lot of hits to kill. After they get up off the ground the third time, I was rather annoyed. For a superhero game, I expect the character to demonstrate the power to chew up the default thug-level bad guys like so much popcorn.
• The Targeting UI. It lies to us. If you put a targeting ring in the middle of your game screen, and it lights up when over an enemy, and the player pulls the trigger…. guess what should happen? Hit the bad guy! Sadly, not so in inFamous. Given how much action and movement goes on, I wish I could play the game with lock-on targeting akin to that found in its predecessor Crackdown.

Combat is challenging, though not in a good way.  The game ramps difficulty by turning the knob to the right on monster hit points. Fine, if rather pointless, if the character damage scales similiarly. It doesn’t. This combat design belongs in a different game. This is a superhero game, ultimately, but the emotion most fights leave me with is relief. I survived. Fine in a realistic game. In a superhero game, I should end battles pumped full of adrenaline. confident in my ass-kicking power.

What else deserves some attention in the inevitable sequel? The repetition. inFamous isn’t alone in the open world genre with the issue of repetitive objective (hello, Assassin’s Creed!). In fact, it’s better than many, which is sad. Still, the gameplay and side missions spread a 6 hour game over a period twice that length. Finishing the game was an act of wanting to see how the story ended, not a desire to play more of the game.

My best session today was ten sessions long. The problem with GDC is that, even when you know the speaker and the topic, sometimes you don’t get you what you’re looking for. Or the speaker has a bad day, whatever. What’s one answer to that? Shorter sessions. For the second year in a row, I attended the the microtalks panel: Ten speakers, five minutes each.

Tracy Fullerton’s talk on difficulty was interesting. As was Clint Hocking’s rock-throwing at reviews in games. But for information density per second, I enjoyed N’gai’s examination of player controlled difficulty: sometimes overt and sometimes as a dynamic response by the game. There’s obvious value and applicability in letting the player choose difficulty, even in multiplayer (Gears 2), but what interested me was how else you could systematically layer in difficulty without just tweaking numbers and AI reaction times.

Best example? Ratchet and Clank. Die in the game, and you respawn, of course, but so have the power-up bolts. And you still keep the bolts you’ve already earned. Unlike our usual expectation where the hard core player is the only one to embrace repeated content and grab collectibles, here the power-up wee factor means that the guy who dies doesn’t feel so bad, and the extra power-ups mean an easier experience as he goes.

PS: I love Metal Gear, including MGS4. And what was my favorite takeaway from the Kojima keynote? The stealth genre was born out of a sprite limitation on the MSX.