Sunday, November 12, 2006

Killing the elder game

I had a chance to sit down with Andrew Tepper, the designer for A Tale in the Desert. His game is a good example of a village game, one of those small MMO that flourish in the dark corners of the internet. They’ve got a small group of dedicated developers making a wonderful, profitable game that is invisible to much of the world. They bootstrapped themselves into existence without selling their souls and are planning for an optimistic future. I believe the market could eventually sustain a couple hundred more such carefully targeted titles without undue competition. Just don’t make a traditional fantasy MMOG

One of the fascinating things about A Tale in the Desert is that it ends. Every 18 or so months, the game starts over again with a new ‘Telling’. This is highly uncommon. Most online games are actually two very distinct game designs mashed uncomfortably together.
  • The advancement game: The newbie grind
  • The elder game: The eternal end game

The two games
In the advancement game, the player gain levels, learn skills and build in-game friendships. The grind up to level 60 in World of Warcraft is a good example of the advancement game.

Once you've completed the advancement portion of an MMOG, you begin the elder game. Now that you’ve got hundreds of hardcore max-level players, you need to keep them entertained in order to keep the subscriptions flowing. Typical mechanics that you find here include PvP, Quests for items that give social standing but not practical benefit and lots of guild politics.

The elder game is all about maintaining a steady state. Brian Green describes it as an episode of Star Trek. No matter what crazy things happen in an episode, everything is back to the same situation by the end. Sure, you spent two months planning a raid to capture that enemy castle. The balance is carefully tweaked however so that eventually, you will lose the castle. Use of heavy negative feedback means that all the PVP, legendary questing and more is ultimately and intentionally for naught.

There is obvious tension here.
  • Game balancing. Building and balancing two distinct play systems cost a lot of money, time and effort. The shear scope of the game systems is often a problem.
  • Audience Expectations: During the advancement game, you’ve attracted and trained a population of players that loves advancement. Then you plunk them into a steady state system. The whole process provides a schizophrenic value proposition to the customer
  • Cultural shifts over time: Steady state systems are even more difficult to balance in the face of shifting cultural dynamics. When the game launches, you want a lot of investment in the advancement game. As the game evolves, you need to start improving the elder game to keep the players who have maxed their character. Finally, you need to start streamlining the advancement game since the newbie zones are often desolate after the initial rush of players. So over the life cycle of the game, you end up with radically different design requirements that are constantly shifting.
Cyclic games
A Tale in the Desert is a great example of how you can run the game in 'cycles' or 'Tellings' in Andrew's terminology. Instead of heading directly into the elder game once the advancement game is played out, you end the game and start it over again. This gives you the ability to start out all the players from scratch and let them replay the game that they originally fell in love with. Think of it as the repetitive play of Tetris, but on the scale of 3 to 18 months instead of 15 minutes.

You end up with some great benefits:
  • More focused design: Instead of multiple game designs, you have only one, the advancement game. This can reduce costs and improve balancing.
  • Built in opportunity to refresh the game play. Suppose you want to add a completely new system into the game to address some flaw that was uncovered in a previous Telling. In a traditional MMO, you’ll often run into massive resistance from the player base if you try to alter an existing system. They seek to protect their investment. When a new Telling begins, you have license to add new mechanics. Players may even expect it; there is a thrill in discovering new systems.
  • Clearer player expectations: It is easier to align player expectations with the design. You aren’t forced to constantly shift the design as the player demographics age.
  • Marketing relaunch opportunity: You get the opportunity to relaunch the game in the market. Every Telling is in effect a new version of the game that can be promoted and sold as a bright and shiny new thing.
Not all is rosy
You introduce some interesting new problems, but these are not insurmountable.
  • How do you get players to go happily into the light? Andrew’s answer is group goals. The group chooses to complete goals that win the game and start the next cycle. By empowering the player to make the choice, there are far fewer complaints. The designer engineers the Seldon-esque dynamics of the system to encourage player choice to flow, on average, in a particular direction, but the players still make the final choice.
  • How do you prevent increased player churn at the end of the telling? Currently, the churn doesn’t seem so bad from the numbers that were bandied about. A small meta-game involving tokens that indicate social hierarchy might be enough to stem end of Telling churn even further.
  • Wither the player content? When you reset the world, you wipe clean vast quantities of player content. I don’t honestly know how bad of a problem this ends up being. This social force could easily contribute to the dramatic momentum of the game. Imagine one group trying to prevent the end of the world and their clashes against those who want to advance to the next round. Setting player expectations regarding the purpose and life span of their creations is likely the biggest challenge.
  • Gaps in the advancement game turn away new players: In a typical advancement game with lots of grinding and leveling, there quickly emerges a gap between players that join only weeks apart. In a traditional MMOG, you know that if you grind long enough, you can ultimately join your friends at the elder level. Cyclic games that only have the advancement game always have to deal with players that exist in classes based off when they joined. One such cyclic title: Earth 2025, mitigates this somewhat with the use of allegiance systems, where new players can join existing groups. In other games, setting up a healthy flow of economic value between newbies and elders can also increase the collisions between players of different levels.
All in all, these cyclic online games are a fascinating variation on the typical MMOG. It only goes to show that the dominant game design model is not the only model and that there is still an enormous space to explore in online games before we've completely tapped out their possibilities. Thanks to Andrew and Brian for the great conversations.

take care

A Tale in the Desert
Andrew Tepper's game.

Brian’s web site. He mentioned the elder game concept to me.

Earth 2025
One of my favorite examples of a cyclic game.

Village Games
My article on village games from 2005


  1. Note that most content in online games is set up like a Star Trek episode. You can rescue the fair princess, uncover the dastardly plot, or even save the universe as we know it. The next person in line is ready to do the same. The content has to reset so that multiple people can go through the quests. It's expensive to create content that will only be used once! On the other hand, it makes the world feel a bit more artificial. So, it's interesting to note that while people claim they want a more immersive world, these types of quests have been becoming more and more dominant; try to imagine World of Warcraft without any quests.

    What ATITD has done is very interesting. A great example of how a smaller-scale online game can be run with some measure of success. I think it's niche status has allowed Andy to try out something that's fairly innovative without risking millions of dollars to do so. I wish there were more small-scale online games to try out other new ideas.

  2. Something you said made me think of Guild Wars, which contains both of those games, but they are very distinctly seperate games.

    If you want to play the 'elder' PvP game, then you can create a full level PvP character from scratch.

    Or you can play the 'grind', which to me has always felt like a massively co-op RPG. Where things don't follow the Star Trek reset-switch model as everything is instanced and you and your party can actually proceed through a plot, never feeling that it's being done by hundreds of other people at the same time.

    And of course you don't have to pay montly, but instead new campaigns (with additional PvP elements) are released episodically. I never could understand how that didn't appeal more than World of Warcraft.

  3. Anyone here know of any real efforts in creating a persistent game world, where the player-character themselves are NOT persistent. Also known as perma-death, or realistic life cycle.

    I don't know of any since early UO. Which is in its own right the best MMORPG design ever. Naivé maybe. But still the best.

    I think the cyclic-world design is a step in the right direction, but not really taking the real plunge (again).

    It would be really interesting to see if anyone can come up with REAL solutions to the problems perma-death poses. Of which there are many. But having persistent characters do too, in which I don't see a real solution yet. It might be simpler by having perma-death actually.

    Any takers? Or links to pages that do?

  4. "Seldon-esque". Nice reference. This is an otherwise useless comment.