This is a bit of a compare-and-contrast post. See, in Pendragon, the default assumption is that everyone will play knights. Some of those knights may come from far-off places (typically in search of interesting ethnic benefits - we'll gloss over the inherent racism in the fact that Danes, for example, get a bonus to Appearance), but you're still all expected to play fighters. On a very limited basis, GMs may allow wizards; we had one actual wizard out of the 19 player-characters that joined the group. (For the record, that was 19 characters over the span of 11 players.) It's also nominally possible to play a courtly (rather than knightly) woman, but no one ever pursued that idea, since it would mean not participating in huge swaths of gameplay. So we had 16 male knights, 2 female knights, and one male wizard. One of those knights was a dedicated archer, and another had a bow that he used occasionally.
Add to this the fact that there are some skills everyone needs, some Virtues everyone needs, some Passions that everyone possesses, and we all need to push the same ability scores as high as possible.
This means that during character creation and thereafter, you're working pretty hard just to figure out how your character is different from the others. In D&D, this is incredibly straightforward: your class (and possibly race, but mainly class) makes you work in a fundamentally different way in almost all situations from any other class. Early editions of D&D were pretty likely to see one or more characters of the same class in a party; later editions, much less so. (As late as 3.5, I did see parties with multiple fighters, though one typically multiclassed into rogue or something.)
I'd be interested to know how conscious Greg Stafford (link is to his Wikipedia page, not his personal page, which is not easy on the eyes) was of the effects of this design while he was writing it. When we watch other characters do things better or worse than our characters, but in the same way, it feeds competitiveness. Not to the exclusion of cooperation - we were sufficiently threatened by external (sometimes literally environmental) forces that we had to cooperate. In the first post, I mentioned the sort of leaderboards style of play that we got into. Watching the gigantic Saxon annihilate everything that stood in his way while my puny Roman knight (well, started out puny) was slap-fighting and struggling to stay standing... this was a pretty good reason to do something to tone down the huge dice pools of damage. But this isn't a post about balance issues; this is a post about character niches.
As my puny Roman ate his spinach (or whatever; for all I know it was gummiberry juice that Merlin fed us on the sly), he got better, enough to compete with the non-Saxon knights. See, the other thing that feeds competitiveness is that some (though not all) of my Glory comes at the expense of other players. If I kill an opponent all by myself, I get all of the Glory from that kill. If what we're modeling is a band of reckless glory hounds, then this is exactly the right system. Oh, there's a consolation prize for knights who showed willing, but it's just enough to point out how much you're falling behind. See, in Pendragon, you don't have experience points, just Glory. Glory is how you get better at things. So the rich get richer, and sometimes it feels like falling behind now means an ever-widening gap.
But combat isn't the only context for competition. Players also earn yearly Glory from their manors, and from having very high Passions or Virtues. High scores in certain Virtues also grants statistical bonuses, and a more substantial yearly Glory boost. Qualifying as Chivalrous grants three points of armor (that applies even to effects that otherwise ignore armor, like falling off a horse), and qualifying for a Religious Bonus grants a different bonus for each religion (some of which are objectively better than others, but I digress). This is, if anything, like D&D's tradition from Original through 2nd edition of granting a 10% experience bonus to characters with high stats in prime requisite ability scores. Not only are you better, you also get better faster. Pendragon uses this more effectively, though, as it is a carrot that one can work toward every year, not only at character creation.
Competition is one form of niche definition in Pendragon. Every number on that character sheet could go up, year by year. (It's mythic Britain, though, so even Age isn't absolutely guaranteed to go up every year.) Another is to pick a skill or two from the rather sizable list that other characters aren't pursuing, such as Siege or Compose. We had a number of aspiring bards over the course of the campaign. My characters, Gracchus son of Septimus and Gracchus the Younger, both got cultural bonuses to Intrigue, because apparently Romans are schemers. (Who knew?) Since I started with the best Intrigue score in the party and wanted to define my character as "the underhanded and politically savvy one," both characters focused on increasing Intrigue as much as possible. Still, I had to work on increasing my ability scores, Sword skill, and the like just as much as everyone else.
I suppose that's another important thing that contributes to Pendragon's capacity for both competition and niche definition: freeform character growth. A player has partial control over how her character improves, and little if any control over how the character diminishes (through age, failures of Virtue, or undesirable Passions). But then, this is Pendragon, which emphatically does not subscribe to the idea that a character should be whatever his player wants him to be, or even that you should necessarily be able to progress toward your ideal however you want.
I'm sure I'll be revisiting Pendragon in future posts. I find it interesting, in this case, to consider how the game's design choices direct players to look for ways to be different, as compared to other games in which strong character differences are immediately apparent from the moment of character creation (and it would typically be considered undesirable to have two characters in the party who were too similar).