Storms on Cloud 9

Storms on Cloud Nine #7:

The Puzzle Enigma

by Scott Holliday
April 4, 2003

If you've been reading the SOC9 forum, you would have seen that I'm planning much of the up-front action in Orphan Crown as secrets, riddles, and puzzles. Naturally, this raises a lot of red flags for anyone with experience in gaming. So, for today's column, I decided to talk about the much abused role of the explorer in online games.

The explorer: I'm sure many of you are thinking of the player who goes out and investigates every nook and cranny, finds and reports all the secret passages, bangs on each puzzle until it breaks, and knows all of the game's stories and riddles by memory. Unfortunately, this is not how it usually works. Instead, the first explorers go out and explore everything. They then show it off or post it on the web, and now everybody knows. Every stone is turned and the world becomes a more boring place. Later explorers may find tidbits here and there, but typically nothing to really make their focus exciting. Worse, if they do find something nifty and show it off, then they've just ruined it for the next explorer in line.

As an online game, this presents a significant problem. Developers can only create a finite amount of content. Even if they are constantly adding new puzzles and tricks, they are only serving whichever explorer happens to get there first. This gives new game content an extremely low return for the time spent. So, as usual, I'll go with a list of ideas, some good, some maybe not-so-good...

Player generated content: It's an online game, and player interaction is supposed to be key. Unfortunately, most games do not have a way for players to change their environment significantly. Likewise, the incentive to participate is greatly reduced. The potential explorer wants to find uncharted vistas... not find your lost pocket-watch. Besides, I bet you know where you put it and won't tell me! On the other hand, the presentation of verbal riddles and games between players can be a lot of fun. This isn't exactly the realm of the stereotypical explorer; however this sort of activity often falls into the same venue. Since explorers often reach self-actualization by proving themselves, being able to answer another player's riddle is often just as satisfying as finding an unexplored nook. And other personalities enjoy the creation of these same puzzles. The goal for the designer should be to find a way to encourage this sort of behavior. When the players serve themselves, it's a win-win situation.
Computer generated content: One of the joys of the computer age is having a machine to do the work for you. Other columns I've seen here at Skotos have talked about using a computer to generate landscapes or terrains. The same concept could be used to put in secret passages, hidden rooms, and nifty tricks and traps. One important issue here is that puzzles shouldn't be difficult to locate. If there is a puzzle in the room or area, there should be hints to at least look at it. Likewise, players often need an incentive. Secrets and puzzles of this sort should present some sort of tangible reward. If it becomes, "blah! another useless secret room," the strategy has failed. In fact, the best incentives are those that encourage the explorer to go get their friends and show them the spot. In game terms, this could be a room where you regenerate faster, or can work on a certain skill, or spy on other locations.
Computer randomized content: Some puzzles are too complex to easily program a computer to generate. For instance, computers (at least currently) aren't very good at making riddles. My guess is that this is because the nature of a riddle usually involves imagination, very fuzzy logic, or even a certain interpretive key. In any case, there are already plenty of riddles out there to be used, and people are always making more. In fact, perhaps the incentive for making new verbal puzzles (as mentioned in part one) could be the honor of having them added into the games repertoire. With a large enough set of puzzles, the designers can spread them out across the game world to satisfy the wanderer. Note that spreading can be done across game distance, time, or even pure randomness. For instance, the sphinx may only give a riddle "when she's not hungry." Furthermore, the riddle could be from a large random set that she knows. In addition, if players have a supervised method to submit their own riddles, the sphinx can become more and more crafty.
Player specific content: No matter how many mazes, puzzles, secret doors, and riddles that are available, it is human nature to show, quantify, and distribute information. As soon as the players hit, everything becomes old news. One idea to stem the tide of information leak is to provide certain puzzles to certain players at certain places. In other words, each puzzle is available to every player, but not in the same place. Finding that same puzzle isn't just a simple look at someone else's web page. Likewise, if you run into a puzzle you haven't seen before, you can't just find someone else's exploration log of the area, because it wasn't here for them. Alternately, simple puzzles such as mazes could be dynamically generated based on set characteristics of the player. Perhaps a seed is generated when the character is created. This way, when two players enter the same maze, they will see two different mazes. However, if they leave and return, it won't change for either of them. Clearly, player specific content also presents several disadvantages. The main one being that the explorer can't show other players his discovery. One player cannot lead another through a dynamically generated maze. My assumption is that if a designer chooses this method, they will have to provide a way for player to 'lead' other players to the same spot and perception. In one solution, the player following might see the leader pass through walls and take impossible actions. In another scenario, maybe the leader's generation seed is passed to the follower... as long as they are still following.

As a conclusion, I'll use an example of one of the systems that I've generated for Orphan Crown. Orphan Crown is a fairy-tale world, which presents many problems, one of them the nature of magic. Early on in the process, we quickly encountered the problem with describing fairy-tale magic. It has several very odd properties and common themes. After thinking about it for a while, I had a somewhat related idea. What if each magician's spells were completely different? Immediately, I began drawing up systems involving spell mechanics dependent upon character randomized seeds. In the end, the system didn't end up looking at all fairy tale like, so it was scrapped. However, it did make for a nifty internal game... which I then began work on.

To cast a spell, a magician would simply input:

Spell [target] 'you banana head!

The room would then see:

[Actor] begins to cast, "you banana head!"

Then based on the character's random seed number, the spell would do... something. Always the same thing, but until you've tried it out, you never know what. In fact, pretty quickly we realized it was fun when about half of the spells backfired - even dangerously. For dueling purposes it was a lot of fun. Going into a duel blind (with no spells researched) was a hoot. You were just as likely to blow yourself up as your opponent. Like I said, since it ended up being so blatantly opposed to the fairy tale theme, we haven't quite decided what to do with it yet, but it's a lot of fun. In any case, an analysis:

Player generated content: A little bit. Since the player gets to choose their evocation, the player chooses what phrases to try. If "eat my shorts!" turns out to be a death spell, everyone involved gets a laugh.
Computer generated content: None. All of the spell effects have to be programmed in manually. Although mix and matching is possible, inventing new effects requires imagination instead of logic.
Computer randomized content: Lots. There are tons of different spell effects, types, and interactions. In fact, some of the more unusual spells are quite difficult to find. In addition, the same seed that generates the spell description and effect, also generates length of casting, the strength and stability properties.
Player specific content: Totally. Each player's spells are generated from a different random key. You can't learn spells from other players, instead you must experiment for yourself. And without the proper precautions, spell research can be deadly...