Game Over. Insert Coin.
The balance between carrot and stick, reward and punishment, in game design was so much simpler back in the arcade.
Take the gamer's money and give them a limited number of chances to progress, usually called "lives" since failure nearly always means death. When the player runs out of lives, they can pay to keep playing if they agree within a given time period. If not, the game resets itself to the start.
In some ways, it's a magnificently simple and beautiful state of affairs compared to what PC and console gaming has become, where the entire price of a game, hardware included, is bought and paid for in advance, and "pay for play" means online access fees and MMO subscriptions.
How, in an environment where you can't hit the gamer in the pocketbook for failing to demonstrate the requisite skills, can you punish them? Should you even try? Arcade games were designed to be "finished" only by the best of the best, but today's story-driven, cinematic AAA titles cost millions to make-- is it wise to reveal the entirety of one's design only to a select few? Might that not tempt designers to leave the ending out (I'm glancing in your direction, Halo 2, and yours, too, Indigo Prophecy) and focus energies on the beginning-- the part that most reviewers will see?
Is death in games supposed to be punitive, or is it there only to prevent the player from progressing through the game until they've demonstrated a certain minimum level of proficiency? If it is supposed to be punitive, what does it say about designers' opinions of their own game if the worst punishment they can come up with is playing the game more? Isn't the idea of dying, the message of failure, more important than the actual consequences? Or is it? Can a game design aspire to have replayability and still consider repeat play as a punishment for dying? What other punishments can there be? Should there be any punishments at all? Can any punishment be as useful or effective as requiring the player to insert another quarter, and if not, should gaming return to the arcade model, or should it abandon player punishment altogether?
Death Is Overrated
Certainly some PC games, especially in certain genres, don't include death or anything even approaching death-like consequences. Adventure games like the Monkey Island series don't include dying. The worst punishment you can receive is simply the failure to solve the puzzles required to advance, which restricts your access to more content. The content itself is the reward; seeing new screens and new levels. Simply being deprived of it is punishment enough.
Other events in the course of the game that seem like punishments-- blowing up, being captured by cannibals, falling out of a tree and hitting your head-- are all simply events that occur and advance the plot. Some puzzles might have longer or shorter solutions, and some have shortcuts, but when executed properly this kind of design never gets the player into a position where the game can't be completed, and the save game functionality is there to allow you to stop playing and come back later-- not to protect you from making a bad decision or to recover after a failed attempt to play a portion of the game.
Creep And Save
When you combine a game that still keeps player death as a punishment, with the PC (and now console) gamer's expectation that they be able to save their progress, you get creep and save. Every potentially dangerous situation that might result in player death could result in loss of progress, so the logical reaction is to save the game before that situation to avoid the loss of progress, and repeat only the minimum portion of the game absolutely required.
Of course, that leads designers to prohibit saving in certain areas, and furthermore players can never be sure what rooms or areas are dangerous and which ones aren't, so that leads to saving everywhere, saving all the time, saving every time an opportunity presents itself.
Limitiing saving to certain special areas is only a partial answer to this problem, as this often encourages players to backtrack, sometimes considerable distances, in order to reach a save location, and replaces the tedium of repeating the same gameplay sequences over with the tedium of traversing empty game space over again. However, players are willing to do it because while the former feels punitive, because it results from the player's failure to successfully complete a gameplay sequence, the latter does not. Furthermore, the latter is within the player's control, while the former is not.
Checkpoint, One, Two, Three
A further evolution is the checkpoint system. Halo certainly didn't invent the system, but in terms of console shooters it certainly popularized it. With control over saving taken out of the control of the player, it now happens where and when the game decides, according to rules that are only visible to the designers. Depending on where death occurs, the player may have a few seconds or many minutes of gameplay to repeat. Again-- is this supposed to be punitive? It certainly feels punitive. Certain portions of games get reputations for being difficult based on how often death occurs, and how much of the sequence needs to be handled perfectly before triggering such a checkpoint.
Hangar Bay 2 of Halo 2 and the lift room on Truth and Reconciliation immediately spring to mind as long sequences that had to be completed in their entirety without fatal mistakes.
The very least one can expect from designers in such a system is not to link cinematic elements to checkpoints unless you really want players to hate your story. The pumping station at the end of the underground sequence in Gears of War is a great example of this. Dying during that sequence meant being sent back a long walking distance from the encounter, and enduring yet another repetition of a typically corny conversation with Dom. Once or twice, the scene is fine. The tenth repetition lends it little more depth, and by the twentieth the player could be forgiven for giving Dom a lobotomy with the business end of a torque bow bolt.
Send In The Clones
Bioshock took an interesting take on the checkpoint save system by taking the player and resetting them to a previous location in the physical game world, but not returning the state of that game world to a previous state.
So in Gears or Halo when you encounter that massive group of enemies yet again, only to be killed (yet again) by the last one, you know you always have to fight through the entire group again until you get it right.
Not so in Bioshock. Kill the first four in a group of five only to die, and when you resurrect you'll be at less than full health and some distance away from the battle, but you'll only have to face that final enemy. Of course, to make fights challenging, that means making some enemies considerably stronger than they might have been otherwise, which is the role played by the Big Daddies.
It does give the player some interesting choices to make. They can choose to try and prepare carefully in advance for encounters, and complete them without dying. Or, they can approach them aggressively and attempt to defeat enemies by attrition, whittling them down life by life.
In my own playthroughs of Bioshock I tended to deploy both approaches, depending on context. Is the mere fact of death enough to discourage reckless play?
Damn The Torpedoes
In my opinion it is. I tended to want to try and defeat enough tough enemies like Big Daddies without dying, and would carefully collect the best possible weapons for the task, and choose advantageous positions-- usually near hacked missile and gun emplacements. If, however, the attempt failed, the likelihood of my simply rushing in and pouring more firepower into the target, without serious advance planning, would increase with each death.
In actuality this plays out in games with the more traditional checkpoint system anyway. Just recently, while playing Halo 3's The Ark level on Legendary with a number of skulls on, I spent a great deal of time preparing for the very first encounter-- keeping marines alive, giving them sniper weapons to make them more effective, trying to spare ammunition by always going for head shots, and making good use of enemy weapons by killing tough units with plasma pistol combos.
Invariably a long unbroken sequence of successful encounters would end in unexpected death from an unseen carbine jackal or a lucky grunt grenade throw, and each time I began the sequence over, I had less patience for the planning and preparation-- which only led to quicker player death. The end result is encouraging a style of play that is conservative without being too time consuming, and aggressive without being reckless, to minimize the time spent-- because after awhile you want to be rewarded, and reward means seeing the next encounter. Often the first casualties of such an approach are your AI allies. Although the ammunition in their sniper weapons is unlimited, they simply aren't as effective with it as the player can be, and without a vehicle it is hard to control their actions and thus protect them. This is a shame, as keeping marines alive with you is the closest one can get to the best experiences Halo has to offer without network cooperative play.
Home Of The Brief
Because of the increasing price of games and game hardware, the increasing fidelity of audio and visual assets, and the increasing time required to develop games, it has become fashionable to accuse games of being too short.
It is easy to see how the traditional checkpoint system helps protect developers from this problem; gamers conservative enough to want to complete a game with a minimum of deaths will most likely proceed slowly and take more time. Gamers who rush in will likely die and have to repeat sections.
Bioshock's modified system doesn't do much to actively reward players for playing it safe, nor severely punishes them for dying, beyond the bare knowledge of player death itself. As such, it doesn't significantly impact the amount of time required to play the game, and places the pace more into the control of the player than the designer.
Is there a way to control player progression in shooters without death? Should death be punitive, or is failure to progress punishment enough?
Living On The Edge
I'm of the opinion that the optimal design for an encounter, or indeed the sum total of all encounters within a game, is for the player constantly to feel as if they are about to die, but never actually dying.
One may wonder if such a threat is credible if death is not merely possible, but actually happens. This leads to another question-- should the player be able to successfully complete the game the first time through, without dying?
Most games, especially shooters, are simply not designed that way. Players are given very little in the way of battlefield intelligence before being thrown into the fray. The first encounters are to gather the information needed to actually defeat the encounter-- where enemies will spawn, how many, and when. Without that information, player death is an inevitable consequence of player surprise. There's little chance of completing most encounters until you know in advance what will happen. The sole tool for the improvement of one's performance becomes trial and error.
I think it's open to argument, though, whether the actual reality of player death effectively establishes the risk of future player death in order to create dramatic tension. In fact, it may actually desensitize the player to it, as it does in the above example from the Ark level of Halo 3. Player death increases player frustration and impatience, making future deaths more, rather than less, likely.
The Illusion Of Danger
For creating dramatic tension it may be enough just to create the illusion of danger-- situations in which either the player cannot actually die, or is actually unlikely to die, but that feel dangerous to the player.
A good example of this comparison, and how I think designers can sometimes be tempted to mess with a good thing, is the change in the design of Halo's hunters between the first game and the following two.
Hunters in the first Halo game, as in the sequels, were nearly impervious to all damage to areas covered by their armor. However, their exposed points, the largest of which was an inexplicably large patch on their back, was so sensitive to damage that a single pistol shot would do the trick.
This was deemed a poor design as it made dealing with Hunters too quick, and the spot was toughened (and the pistol removed) in later iterations.
But while hunters became tougher to fight in Halo 2 and Halo 3, the encounters themselves became less interesting. Hunters could be destroyed at long range by rockets, or by a sufficient number of grenades, but the satisfaction of the one-shot, close range kill tended to tempt the player into proximity. At that range, the hunter can kill the player with one melee swipe. The swipe itself, however, was predictable, and the payoff so great that it was not a deterrent. Once learned and properly timed, a good player can easily take out a pair of hunters without serious risk.
However, I'd argue that it always felt like a serious risk, even when it wasn't. There is something that feels dangerous about tackling those enemies at close range, about hearing the clinking armor almost feeling the motion of air against your cheek as the melee swipe misses and you pivot around the charging Hunter to get the one shot kill.
By contrast, Hunters in Halo 2 and 3 are harder to kill but much less fun. That they can deliver a follow-up melee hit behind them if they miss the first time means that there's more risk in getting close. That you can't reliably kill them with one shot from any weapon means there's no incentive to get close. That they are tougher means it takes more time and more ammunition to kill them, and the changes to the hunter's weapon makes it also simultaneously more deadly and less dangerous-seeming.
The fuel rod gun bolts of the first Halo game, with their arcing trajectory and massive-sounding explosions, gave further incentive to the player to close within melee distance-- where the hunter would stop using the weapon-- and try for the one shot kill.
By contrast, the continuous beams of Halo 2 and Halo 3 hunters are more deadly, but the lack of an explosion makes them seem less so. They encourage players to stay in cover at a distance, rather than come out into the open and close the gap to the target.
Halo 1 hunters were formidable-looking, dangerous-sounding beasts best killed up close and personal, that bestowed personal satisfaction on the player out of proportion with the actual difficulty involved, but perfectly proportional to the level of danger perceived by the player.
By contrast, hunters in the sequels are pains in the butt best engaged from long distance (like almost everything else in the sequels) that one is glad to be rid of, but that produce little exultation. The exciting, satisfying kills for me in Halo 3, for instance, are usually vehicles, like the Phantom dropship, the Scarab, or even the Prowler and the Wraith. These units not only are dangerous, they feel dangerous-- sometimes more dangerous than they actually are, and player perception is actually more important than reality. Using the audiovisual means at their disposal, those units actually present the perceived risk of death better than other units that are actually more capable and likely of producing it. They better achieve the goal of making the player constantly feel as if they are on the verge of dying, without actually killing them.
Which is good, because I'm all out of quarters.