Friday, September 29, 2006

Shadow Content

Shadow content is content which improves a player's gameplay experience while not in direct contact with the player. This doesn't mean the player can't directly experience this content. It means that the content has a beneficial effect even when not immediately present. A possibility of contact is often required, but not always. Shadow content feeds player hopes, anxieties and expectations.

Too often, game developers seem to underestimate the returns on investing in shadow content. Perhaps this is because the concept has been poorly implemented in past games.

Some considerations in implementation:

  • Awareness. It's easy to create such content and forget the need to relay knowledge of its existence adequately to the player. Depending on the precise content, this may be done mostly through out-of-game advertisement or through in-game lore and mechanisms.
  • Veracity. The player must consider his or her hope, anxiety or expectation to be a reasonable belief. Playing the lotto is only fun to those who trust that their hopes are reasonable.
  • Audience. Is the possible contact with the foreshadowed content a pleasant prospect to all players? The appeal of such content doesn't always need to be universal, but its effectual audience is an important piece of information. The intended audience helps define both the content's nature and how awareness of it may best be achieved.
  • Balance. Contact with the foreshadowed content should not disrupt the greater design. Star Wars: Galaxies failed in this way, in my opinion, when they made jedi slots regularly accessible.

Some examples of shadow content:

A gambler is someone who risks something in hope of an uncertain reward. It's important for the designer to know where and when the stakes should be high or low. Some people prefer big bets, others small ones. Misplacement can alienate audiences.

As Tom Abernathy stressed at the AGC, timing is very important to comedy. Sometimes, comedy works best as a shock, but it can also be springboarded by expectation. Those who are already laughing are generally better able to appreciate a good joke. A designer must exercise some control over experiences that might disrupt a player's feelings of expectation.

Horror games thrive off the belief that the player could encounter danger at any moment. Like with comedy, timing is often very important to creating scares and an unsettling atmosphere.

A handful of employees solely devoted to “playing” NPC creatures, just one or two per server, can have a profound impact on player expectations. I'm referring to unscheduled and unadvertised player experiences.

Imagine you're walking through the woods. You see some wolves off to the side. They're out of hostile range and you've got other fish to fry right now, so you continue on. A few minutes later, you stop and look around to get your bearings. A wolf is standing off in the distance behind you, and it seems to be looking at you. You're curious, but you assume it's just a mob you didn't notice as you passed by. You continue on...but curious, you look back. The wolf is not far off. It's following you! You run to attack it, but it runs off and disappears. A few minutes later, it's there again.

In that scenario, a dev-controlled creature has made a player's game experience extremely unique and interesting. The player has been given an interesting and unique story to tell...and didn't even fight anything. If all players are aware of such sporadic, direct developer involvement in creature behavior, then the gameplay experience of non-participants is augmented by their merely having heard the tale from their fellow player. Of course, some players would make up stories, so there would be debates about which stories are actually true; but that helps to spread awareness (or perception) of the shadow content.

NPCs can be occasionally possessed by developers to create unexpected behavior, within the confines of that NPC's history and personality. Maybe an NPC in town that you see 20 times a day is acting a little weird today. Maggie May, the local librarian, is in the tavern and drunk as a skunk. Odd.

The dev-controller could take control of NPBs (non-player-beings...not necessarily persons) all over the map on that server. Nobody would know where the "smart" NPB would show up next. The controller would keep a record of every being controlled in every area, to spread the love to as many players as possible (that includes players like crafters, who spend all of their time in town). Plus, when feasible, the employee would keep track of which PCs he or she interacted with. In addition to spreading contact with such encounters, giving one player 2 or 3 such experiences would create the illusion that smart NPBs are more common than they truly are. It would also ensure players who have such an experience don't give up hope of experiencing it again.

Shadow content can be immensely effectual. In the first 6 months of Star Wars: Galaxies, everybody knew that the odds of unlocking a jedi slot were astronomical. But most players still couldn't help but hope that he or she would be the one to do it. The desire to play a jedi overshadowed our understanding of the possibility's unlikelihood. It was an addictive feeling, and it was similar to this possessed NPB scenario, because nobody had any clue how to unlock a jedi slot. The reward of that particular shadow content was two-fold: (1) it inspired player imaginations, thereby encouraging a general excitement that raised our spirits and augmented gameplay; (2) it provided a strong goal which contributed significantly to player retention.


  1. I like the term, though you could have been a lot more specific in defining it in the first paragraph. A lot of what you're talking about could be implemented with character AI instead of employees. Its a lot cheaper that way!

  2. AI might be cheaper, but not as cost-efficient. With what I'm talking about, you get a lot more bang for your buck...especially since the people you hire to do it don't exactly need to be brain surgeons (i.e., you don't have to pay them a lot). And AI can't respond to player actions creatively and with individual-player consideration, so this opens up experiences not possible through current AI.

  3. Well, it depends on the algorithm. Currently there are several threads of invention in terms of character AI. For instance, this is an engine design my company's R&D is pursueing. The ruse might not last on deep interaction, but it could work for the shadowy purposes you describe.


Note: Only a member of this blog may post a comment.