On the level: some thoughts on advancement

Recently, after eighty-odd hours of play spaced out over more than a year of real time, the Team Tsathogga PCs hit level 5. Partly because half the players were new to D&D when we started, I didn't use an experience point mechanic, going instead with a 'level when it makes sense' set-up; and, so far, it seems to have made sense approximately once every sixteen hours of actual play. (They started at level 0.) The time between level-ups has been getting longer, though, and I'm sure the average will increase the longer that the campaign continues.

Of all D&D's innovations, the levelling system is one of the oddest, and one of the most influential. Like most of the game's other features, its origins can be found in the historical wargames that D&D evolved from, which sometimes featured rules to model how a unit of troops might go from raw recruits to hardened veterans over the course of a long campaign: but D&D took this simple concept and stretched it so far that it became almost unrecognisable. A D&D character advancing from level 1 to level 20 isn't really like a wargame unit advancing from green to veteran: it's more like a unit starting out as a regular WWII infantryman and gradually evolving into a Sherman tank.

As far as I know, this advancement paradigm - in which characters begin as more-or-less ordinary people and gradually transform into mythic heroes - was a D&D innovation. It's since gone on to become deeply embedded in the structure of both fantasy RPGs and computer games, to the point where it's easy to overlook how utterly weird the idea actually is, especially in its more extreme implementations. It's clearly not rooted in any kind of realism, but it also doesn't appear in any of D&D's source material: Conan, Elric, Aragorn, et al are highly capable individuals right from the start of their respective careers, and become at best only slightly more powerful over the course of their adventures. Only with D&D does the idea arise that a character can effectively change genres, metamorphosing from a grubby desperado to Conan the Barbarian to Beowulf, if only they can manage to kill enough orcs and steal enough gold along the way.

The levelling system persists largely because it satisfies what, for many players, is clearly a very basic desire: the desire to see your numbers increase, power grow, and options multiply, to have your progress and achievements measured and quantified and validated in clear numerical terms. Given that people enjoy levelling, though, it's still worth asking just how many 'experience levels' a game actually needs. Wargames usually got by with just two or three, but D&D's innovation was to add many, many more. Most D&D editions and variants assume a 20-level structure, but it's often been noted that the higher levels tend to get very little actual play: the original B/X rules provided no rules for characters over level 14, which I believe was also the highest level reached by any character in Gygax's original campaign. Early D&D 'endgame' adventure modules, like Queen of the Demonweb Pits, Tomb of Horrors, Dragons of Triumph, and Temple of the Frog, were written for PCs of levels 10-14, which further reinforces the impression that level 14 was the highest level that real PCs were actually expected to reach. (Interestingly, most modern Pathfinder adventure paths top out at level 15 as well, which suggests that the level 14-15 ceiling has held remarkably constant across different eras and editions, even though rules for going much higher have been around for decades.) TSR was publishing ultra-high-level modules as early as 1985 - M1 Into the Maelstrom was for characters of levels 25-30! - but no-one ever seems to have liked them very much, and the question of how to write good adventures for very high-level characters never seems to have been adequately solved. Look at the early adventures that people still talk about today, and you'll find they're all written for level 1-14.

So there are strong grounds for suspecting that the top quarter of the standard 20-level progression has never seen much real use. But I think one can go further: in practise, even going much higher than level 10 seems to be pretty rare. In the original game, 'name level' - the point at which your character had 'made it', and could settle down as a lord or a high priest or an archmage somewhere, was level 9, 10, or 11, depending on your class. The highest level a PC has ever reached in one of my games was level 12. The 5th edition campaign books which WotC has been bringing out over the last few years are mostly designed to take a party from level 1 to level 10, which makes them very similar to the old B-X module range of 1978-87, which theoretically covered levels 1-14 but in practise very seldom went higher than 10. Some recent D&D spin-offs, such as The Black Hack, Shadow of the Demon Lord and 13th Age (I think), even set level 10 as the maximum level achievable.

So there seems to be a second milestone, which has again remained surprisingly consistent across eras and editions, which sees level 10 as the end-point of a 'normal' campaign: levels 11-15 are for those rare campaigns which go the extra mile, and levels 16+ are barely used at all. There's clearly a third milestone around level 6-7: the original campaign-in-a-module, X1 Isle of Dread, topped out at level 7, and level 6 is used as the maximum level by several D&D spin-off systems, including Dungeon Crawl Classics, Hulks and Horrors, and the E6 hack of D&D 3.5. Levels 1-7 is where the majority of famous adventure modules tend to cluster, and it also accounts for the vast majority of my own gaming experience, in which campaigns going beyond level 7 have been a distinct minority. Not coincidentally, the 1-7 level range - especially the level 3-6 sub-range - are also the ones which are most likely to give you the 'classic D&D experience', before the easy availability of game-changing magic like Raise Dead and Teleport starts pushing the game away the default fantasy adventure paradigm. The games I ran for my level 10-12 AD&D 2nd edition group back in the 1990s were great - but they were also weird as fuck, and bore very little resemblance to traditional D&D adventure scenarios, simply because by that stage the PCs had so many tools available to them for bypassing or trivialising the kind of obstacles which form the building-blocks of lower-level adventures. I'm sure they'd have become even stranger if we'd gone higher still.

The boundaries, then, have remained fairly constant: levels 1-7 for fairly grounded fantasy adventure, levels 8-11 for high-powered heroic fantasy, levels 12-15 for fantasy superheroes, and levels 16+ for a largely theoretical end-game which very few people actually use. But what hasn't remained constant is the rate of advancement. TSR edition D&D assumed you'd need to play for years to reach name level, whereas I seem to recall that 3rd edition was built around the assumption that you'd level about once every ten hours of play - more than twice as fast as seems to have been common in 'the old days'. Shadow of the Demon Lord goes further still, recommending a structure in which one session = one adventure = one level, which would mean characters advancing twice as fast again. Personally, I find the rapid levelling of more recent editions strains my credulity: even Team Tsathogga's advancement from level 0 to level 5 over the course of two years of game time seems rather on the fast side to me. But many adventures are clearly written with the assumption that no-one will be surprised if a band of peasant irregulars transform themselves into mighty wizards and warriors after a few orc-stabbing excursions into the woods. That's what 'experience' does to people, right?

So there are two independent variables, here: both how high levels go (either in the form of a hard limit, or just a vague shared assumption that the levelling rules probably won't actually be used beyond a certain point), and how quick or easy it is to move up the scale. In conjunction, they can be used to generate four very different environments:

• Low level cap, slow advancement: The most 'realistic' option. People get more powerful, but not that much more powerful, and it takes ages. Everyone is vulnerable - no-one is ever so strong that they can afford to simply ignore low-level characters - but tearing down the powerful is much easier than rising to their level yourselves. Suitable for gritty or tragic games, in which destroying things (and people) is much easier than replacing them. This is the Lamentations of the Flame Princess model.
• Low level cap, fast advancement: The most dynamic option. The power available is limited, but it comes quickly to those who seek it. The power gap between the weak and the strong is never all that big, and can be rapidly closed by someone sufficiently determined, meaning that it's never safe to rest on your laurels: there's always the risk of some ambitious young punk bursting up out of nowhere and tearing down all your achievements. Suitable for short, fast-moving games which feature rapid shifts in the status quo, especially as surviving characters will rapidly hit the level cap. This is the Shadow of the Demon Lord model.
• High level cap, slow advancement: The most hierarchical option. There are people out there who are much, much more powerful than you are, and you will probably never be able to rise to their level, so you'll probably be spending your whole life living in the shadow of their power. Could lend itself to a revolutionary narrative about underdogs banding together to defeat the powerful through intelligence and guile, but much more likely to turn into a nightmare of being the archmage's errand boys, forever. This is the AD&D Forgotten Realms model, and it's my least favourite combination.
• High level cap, fast advancement: The weirdest option. There are enormously powerful people out there... but, with enough luck and determination, anyone can join their ranks, and do so fast. Likely to resemble a superhero setting more than a traditional fantasy world, with ultra-powerful individuals just bursting out of nowhere all the damn time. ('Last year, I was just a lowly farm boy... but now I am Darkaxe, Slayer of Gods!') Both Pathfinder and D&D 4th edition lean heavily in this direction, in practise if not necessarily in theory.

My instincts have always led me towards the first of these options, with characters levelling quite slowly, but with very few high level NPCs around to make them feel small by comparison. (In a world where almost everyone is level 0 or level 1, a 3rd level D&D PC is badass.) But I think any of them could potentially be fun - as long as the group knows, in advance, what they're getting into, and prepares their expectations accordingly. It's when there's a mismatch between system and expectations - and especially when the PCs seem to be weirdly out-of-kilter with the assumptions governing the rest of the setting - that problems are likely to occur...