The I Want To Be a Wizard Dillemma

This will be a D&D-centric post, but it will address what I call the “I Want to Be a Wizard” dilemma. Very often, I will approach a game as a player with a certain loose character concept in mind, i.e., a class if we’re going to play D&D or a similar game. Some people like to play fighters or magic-users exclusively, whereas I just want the freedom to choose what I want to be. All the more interesting character details tend to fill themselves in as the games progress, but I like to be able to choose my starting point.

So, right up front, one of the game systems I really enjoy is GURPS. It is a remarkably simple system to play in once the game is going, because all of the requisite work is front-loaded into character creation. You start with a strong character concept and build the game mechanics according to it, such that you have an accurate representation of the character. The downside is, there’s often little room to let those characteristics come into existence organically through play. It’s also nigh impossible to have a “pick-up” game of GURPS, even if your GM has gone to the trouble of making Templates to replicate classes. Often, a clear campaign idea has to be established from the get-go before you can start generating characters.

Therefore, as we dismiss GURPS and instead focus on systems that are akin to D&D, we notice the familiar trend of player attribute scores ranging from 3-18 (as a result of a 3d6 roll). Each possible class has an associated principal attribute that will matter above all others for the majority of class-based actions (Fighters need Strength, Wizards need Intelligence, and so forth). So of all the six ability scores, you’re going to want one to be higher than the other if you want to play a particular class.

Now, the most straightforward way to get around this issue is to simply arrange your six results to taste, rather than roll each attribute in order. This is probably the way most people ended up playing, but with the “old school” movement, there seems to be some impetus to stay consistent to the original letter of the law. However, there have been some interesting ways around the 3d6-in-order rule that allow for some character freedom.

The earliest days probably saw the rise of the super-attribute: roll 4d6 and drop the lowest die. This is referenced, I believe, in the 1st-Edition DMG itself. (Someone correct me if I'm wrong.) It tends to promote much more competent characters, less susceptible to death in an exceptionally lethal game system. For a detailed look at the stats, check out AnyDice and their analysis: anydice.com/articles/4d6-drop-lowest/ ...But beware, for down the road of the super-character lies the realm of PC preciousness and invulnerability. Death should be commonplace, shouldn't it?

Perhaps the most fun variant on character generation exists in Dungeon Crawl Classics with its idea of the "character funnel." Essentially, each player starts with multiple (usually four) 3d6-in-order characters, with the last one standing being able to graduate to a first-level adventurer. There's more rules and opportunities for advancement there, but it's a novel approach that exposes a new truth about 3d6 character generation. D&D's random character generation engine has a key flaw in that it only has a sample size of one. That is, while there are exceptional failures and exceptional successes for every attribute, the mean should be 10.5 for each. However, a single sample could fall anywhere on the curve. With four samples, you get a better chance of an average or better-than-average starting point for the type of character you'd like to play. As an example, you may very well end up having two "normal" characters, one "weak," and one "strong" that you'd prefer to play over the others. Depending on the game, that character may survive, adding an extra dimension of randomness.

I think this is an idea that should be explored in more depth in other systems. While I love DCC, it does gravitate toward being a rather off-the-wall game, and it could be a hard sell for a group who wants something more down to earth. But the idea of multiple instances of character creation may be gaining traction in other systems. Here is the optional "Multiple Characters" rule from Adventurer Conqueror King System, which is:

As an optional rule, at the start of play, each player generates five characters and selects one primary and two back-up characters. The remaining two characters are given to the Judge to use as NPCs to populate the campaign setting. If the primary and backups are all killed, the player generates five additional characters, again picking three and giving the Judge two. This ensures that the player always has a variety of characters to choose from to find one he likes, and gives the Judge some additional NPCs to populate the world.

This offers an interesting variation that helps mitigate the problem of limited sample size. The only issue is that rolling 3d6 thirty times can get tedious. I might end up tinkering with some other ways to throw around 3d6 rolls to see if I could keep the high sample count while lowering the overall number of die rolls.

Have you run into problems with "3d6 in order?" How do you typically handle attribute generation in your games?