Iteration is a huge topic.  Like most things you can view it at a "micro" level, and a "macro" level.  You can think of "micro" iteration as the internal, organic and exploratory design process used by studios all over the world.  You take an idea and you experiment and try different things with it until you get something you're happy with.  At the "macro" level, you can think of iteration as genres, franchises, sequels and clones.  As a designer I'm much more invested in the "micro" level, but today I wanted to write down some ideas about iteration on that larger scale.  These are going to be pretty subjective observations, but I hope they can at least start an interesting conversation.

Why do we iterate?

Lots of reasons.  In some ways, we can't not iterate.  Indie auteur Jim Jarmusch once wrote:

Nothing is original. Steal from anywhere that resonates with inspiration or fuels your imagination...And don’t bother concealing your thievery—celebrate it if you feel like it. In any case, always remember what Jean-Luc Godard said: “It’s not where you take things from—it’s where you take them to."

Most artists agree that on a basic level everyone steals from everyone else.  It just happens; you see things, you hear things, you process them and re-interpret them.  That's how we work!  And obviously, it's nice to get a chance to go back and improve and refine an idea that just didn't get long enough in the oven.  The thing that interests me most, though, is the powerful and pervasive idea that iteration on this macro scale reduces risk.  By revisiting a popular property or idea, you're more likely to succeed.  This is obviously the backbone of our commercial game industry, and it's worked pretty good for a long time.

What's this article about, if iteration works so well?  I think this practice of iterating on a proven game concept is so widely accepted and relied upon that we forgot about how many drawbacks it has.  We all know what happens when you make a copy of a copy of a copy, right?  For a lot of studios the good still outweighs the bad, but that doesn't mean the bad magically disappears.

There are a lot of risks involved with iterating on an established game style, but they primarily fall into two camps: production & content problems, and design & audience problems.  If this strikes you as ironic, that's good.  These are precisely the areas that iteration is intended to help.  By working off an existing template, you don't have to worry about a long exploratory design process, or mismanaging your team.  So why doesn't that work?  Why do most games ship late, or with bugs, or after crunching for months?

Production Trouble

I'm going to keep this part brief, as its a bit more pragmatic and clear than the next section.  Games based on iteration require more content than games with a more original premise or control scheme.  Games that do something new gain value from novelty.  Now, novelty doesn't come cheap, and isn't without its own risks, but if you're going to follow in someone else's footsteps then you need to provide value in some other way.  Usually this means content, and a lot of it.  This plays well with the big studios, but even they struggle to keep up with evolving technology.

The point here is that deciding to iterate on a proven design may solve some pipeline problems, but it creates others.  These problems might be more familiar, but I'm not sure they're any easier to solve.  And for small studios, having to compete with the big boys in raw content production can be a death march, followed by a death knell, followed by DEATH.

Design Trouble

This is the part of macro iteration that is more interesting to consider, and has been consuming some of my spare brain cycles for the last few months.  Iterative design, regardless of production pipeline concerns, is a double-edged knife.  Inevitably, when you design a sequel or a followup or a clone or an homage, you are going to focus on some subset of the original piece.  Rarely does a followup keep every single aspect of its predecessor intact; it is simply impossible in many cases.

This is not a blind or random process, of course.  The developer or designer has their own opinions, as well as the opinions of their friends and fans.  But one way or another, facets of the successor's design will differ.  While obvious, this idea fascinates me.

I love looking at my favorite game series to see what aspects survive from title to title, and which parts become vestigial and are lost to digital evolution forever.  These are not scientific or analytical, and there are lots of ways to look at these things.  Also, my goal here is not to argue that this is the right way to view these things; just that it can be interesting or enlightening!

The Legend of Zelda

One way of thinking about any game series or genre is as a tree, rather than a straight line.  The Legend of Zelda tree has a funny structure.  At the root is of course The Legend of Zelda, for the NES.  The tree grows upward and a new leaf appears: Zelda 2!  Some neat things here, but some pretty big problems too.  A new branch grows up from there: A Link to the Past, for SNES.

Link to the Past is a pretty big deal for the Zelda series.  It took the towns and villagers from Zelda 2, but returned the world to a top-down view.  LttP brings a lot of new items, new enemies and new music to the series.  Every other Zelda game ever made will reference this game, rather than its ancestors.

The LttP branch grows a lot of games: Link's Awakening, Minish Cap, The Wind Waker, and Ocarina of Time.  The Ocarina and Windwaker branches are both pretty fruitful, creating some Wii games and some DS games.  Majora's Mask is kind of hard to place in the tree (is it its own tree?) but we can see some interesting patterns anyways.  The thing I always wonder is, "why is the bottom of the tree so bare?"  Or is something like Infamous a new trunk of this tree?

Metroid

Metroid has a pretty similar tree.  At the root is Metroid (NES), a slow and moody exploration game.  Up from there we grow Metroid 2 (Game Boy), which is fairly slow and moody still.  We grow upward a bit more until we find Super Metroid for the SNES, which is in many ways the Metroid equivalent of Link to the Past.  The developer's goal was explicitly to create a "good action game", not to continue investigating the moody, exploration space.

Most people love and prefer Super Metroid over its ancestors, but for me and a few other people the hostility and atmosphere and satisfaction of the NES original was completely lost when it was turned into an action game.  That's not to say it wasn't the right decision; just that it was a decision.  Nintendo decided to focus on specific aspects, and fans of the original are left with a disappointing sequel.

Like the Zelda tree, the Metroid trunk is bare until Super Metroid, when it suddenly flowers into Metroid Fusion and Metroid Prime.  Fusion branched into an action-remake of the original (Zero Mission), which is funny when you prefer the original over Super.  Prime spawned sequels of its own, as well as Hunters.  Not going to cover Pinball here, but a lot of people see Shadow Complex or something like Arkham Asylum as a branch of Super Metroid (though it is obviously cross-pollinated from the Metal Gear Solid and Resident Evil trees).

God of War

I could type about this stuff all day and I'm sure it's already pretty boring, but the God of War series is an interesting one to look at as well.  I'm a big fan of the first two games, but the second game definitely has different strengths.  While combat was occasionally a little rough or repetitive in the first game, the story was small but thoughtful for an action game.  Most levels were very dark, but intentionally so.

The sequel did a great job finding and polishing out the rough spots in the gameplay.  God of War 2 is a remarkably smooth game, with a variety of both light and dark environments, and beautiful, varied art direction.  The end-game sequence was so polished and satisfying that it's honestly hard to find any room to complain.  However, the story devolved from "you accidentally murdered your family and must atone for it" into "blargh, the gods are bad I guess."  But you can see how it evolved - a little more attention to gameplay and visuals, a little less attention to story.

I have only played the demo for the third game, but the trend is a little disappointing to me.  While the visuals got a detail upgrade, the palette has been reduced to brown on brown.  Some nice new brawler abilities are offset by needlessly elaborate execution QTEs, which poorly suit the brutal, angry Spartan.  I've heard similar concerns from other developer friends.  So while God of War III was a given, I definitely don't think it was without risk.

Conclusion

In the case of God of War there is another iterative risk at work, which is the risk of failing to out-do your predecessor.  Darksiders' gorgeous, detailed artwork simply couldn't make up for its lackluster, empty God of War/Devil May Cry gameplay.  The point of this whole rambling mess of words, of course, is that no design choice is "easy," or "guaranteed."  No matter how you approach your design you will have hard problems to solve, and you'll have to put some love into your project, whether it's innovative or not.  I think it would do everyone some good to stop assuming that just because it's "like this other game," that it will somehow succeed by association.  The more publishers understand that, and the more we really actually understand that ourselves, the better off we will be.

Additional Reading

Jean-Paul LeBreton did a fantastic writeup of DooM, often considered the father of the FPS genre, where he explores what changed and what stayed the same.  It's a lot more level-headed and insightful than my own juvenile treatment here.

Wikipedia has some nice notes about Super Metroid's development, which cites a bunch of cool interviews.

I wrote about Infamous's similarity to Zelda in a previous article.

My idol writes a bit about Metroid's design in this crazy design doc.