Now some modern fairy tales have been created which either have plenty of statistical play or, at least, can be easily adapted into games that have statistical play. Prime examples of this include Quest for Glory (a computer game) and Thieves and Kings (a comic/illustrated story). There are plenty of others.
These systems do pretty well at one- or perhaps two-character statistical play. But they are extremely difficult to adapt into four, five, ten, a thousand character play. They still have the memory of the original fairy tale structure, and therefore they always orbit a Hero, and anyone else he or she meets is likely to just be an Accomplice or Background Flavor. A good example of this are the classic Tolkien novels, with their scads of characters, only three or four of which are even worth remembering the names of. Think of this back before the movies came out, not since.
There are systems where many characters - five, six, even seven - can all get along well enough and do cool things in a statistical setting. As far as I can tell, these were all pioneered by the dark and sinister crossover of WWII wargames and
Fundamentally, they took all the elements of war stories and mixed them with fairy tales to come up with a statistical variety. Magicians have a different way of playing than warriors than thieves than elves than... well, you get the idea.
By introducing this highly varied statistical system, these games allowed players to all play together as a functioning team, each player having different capabilities which are more or less useful in various situations and don't require a main hero that dwarfs the other members of the party. (Although early balancing being what it was, this often happened anyway.)
If you notice, this is what the Lord of the Rings movies adopted to distinguish the ten billion minor characters those books contained. While seeing exactly how an elf or dwarf fights as opposed to a human has no particular story merit, it looks cool and is very well known. So now we can remember Legolas as The Elf That Surfs Wooden Planks Down Stairs While Firing Fistfuls Of Arrows And Looking Intense. Before the movies, we would remember him as, "Oh, was there an elf in that party? Yeah, I guess there was." Not that there's any real story difference between those two memories, because Legolas, along with just about everyone else, is there only to serve as background noise until their single plot event.
But there is a gameplay difference.
This gameplay difference is extremely good at allowing for small parties like this, and is the backbone of nearly every modern RPG, both tabletop and computer. How large and diverse your parties can be depends almost entirely on how distinguishable your characters are from each other statistically. So some games, like Disgaea, allow you to have many characters in any given battle and dozens of characters in your roster. They do this because their statistical play is so complex and nuanced that there is an almost infinite variation available. Of course, they are also almost unapproachably complex to many gamers for the same reason, so there's a tradeoff.
MMORPGs have a fun time of this, too, trying to balance variation with complexity and... uh... balance.
Anyhow, that's all getting off the track. These statistical variations are a trick to make you able to have more than just one or two main characters like old games and fairy tales. However, they make it difficult to tell a fairy tale.
FFVI (FF3 as it was called when I played it) gives us a good example of an approach that might help. Like many other games of its time, it had many characters, but it took a somewhat unusual stance that you might remember: the party got split up and you had to play through their story segments separately. This isn't unique to FFVI, but FFVI is the game I bet everyone's played.
This method requires the same diversity of stats to keep play fresh, but by creating multiple independent sub-fairy-tales it gets around the hopelessly dreary march from A to B that most of these games have. Of course, this solution is probably not ideal, as it does involve quite literally making multiple games. It also doesn't adapt well to tabletop games.
Radical diversity of gameplay is another option. This is when the various character classes have SUCH different capabilities that they are essentially playing a different game. This is a subject all its own, but it doesn't work very well in tabletop games due to the way it pulls the GM apart.
A few theoretical methods I've come up with - some of which I've used to a bit of success, some still untested - might be worth mentioning.
One is the Worldbuilder method. In this method, the game is not about dungeon crawling (although there is plenty of it): it's about worldbuilding. For example, players might build or obtain a castle, they might recover or enchant a magic sword, or discover a long-lost spell, or open up a cave full of obedient golems... and, of course, the Enemy might do all of those same things.
In this case the players probably start statistically fairly similar to each other - somewhat distinct, but not tremendously. As they proceed, they define themselves by what they have (and where they are). Someone with a magic crystal sword will have statistical characteristics that revolve around it, while the same character with their own castle will have different characteristics (such as guards). The very act of defining and obtaining these things can also set side quests in motion, allowing for loosely connected MMO plots.
The basic play results in kind of a magic-the-gathering sort of feel, but it also results in an a very fairy-tale feel as well. These games tend to be fairly short - five or six sessions each - but they lend themselves well to multiple games played consecutively in the same world.
The best method I've found for allowing this kind of variation is to allow players to generate two fairy tale elements out of components, and then the GM chooses his favorite for the Enemy. If the players choose to contest it and try to obtain it before the Enemy does, then they have a tough fight ahead of them and the Enemy gets the second object instantly.
Of course, it's not just defining and obtaining components, because these things need to wrap back in on themselves, but you get the basic idea.
Another method of doing fairy-tale games without being simple fantasy games is to use less character-centric play. If the players don't play individual characters but instead play, say, concepts, then the players can use the characters like chess pieces to build the game into a meaningful and deep story.
This is also not a new idea, although it's still fairly rare. It's extremely rare to give the players an actual goal, too: most of these systems simply use a bidding system to make players form conflicts. That doesn't give players enough of a goal for me, so I strongly recommend giving players a concept or moral-of-the-story or something that they can strive for that isn't directly related to in-game assets. IE, don't make them want to revive a kingdom or save a princess. Those are character goals. Player goals would be more like "show crime doesn't pay" or "you get a point for every round of a duel (2-character fight)".
Can you think of any other methods to make fairy-tale-like games? Have you ever used any of these methods, or played in a fairy-tale-like game?