The question seems to have come up with how simulationist mechanics can support actual play. I thought I'd have a stab at illustrating this with examples of actual games I've run using different sim systems and why I chose each system.

Ok, firstly, around 2-3 years ago I decided I wanted to run a gritty fantasy game set in an alternate history Roman Britain with magic and all. Up till then my group had played mostly Gurps. However, one of my explicit campaign design goals was to have the characters start out as ordinary people and chart their rise to heroic stature.

Gurps isn't actually so good at this. It is good at creating competent individuals but the interaction of the heavily stat based skill system and the default skill mechanics make it quite tricky to create characters with great potential but poor current ability.

Looking around, I noticed that Runequest supported this well. Give characters low starting skills but decent to high stats and you easily create characters with heroic potential who are presently pretty useless.

Runequest explicitly supported my campaign goals in a way that Gurps did not. Different sim mechanics, actively assisting my sim priorities in play.

Game 2. I wanted a sword and sorcery game based on Clark Ashton Smith with a bit of Robert E Howard. Gurps was not quite right, characters had an everyday flavour and I wanted heroes. Also, Gurps with high power levels begins to break down a bit. Runequest wasn't so good as the fiddly hit locations and poor starting skills jarred with my campaign concepts. It was nearer though.

In the end, I chose the Elric! system. Characters in Elric! typically have starting weapon skills in the 90s percentile wise, example soldiers and guards have between 40 and 60. New characters are clearly heroic. However, one hit in Elric is likely to leave you badly hurt, possibly with lasting permanent scars and disablements. This fit the Ashtonian mood well, characters could battle many foes because their high skills made parrying extremely probable and meant they rarely missed their opponents. Characters were still vulnerable however as a single lucky hit could kill or cripple them. Heroic, but dark.

Note, my priorities in play were actively supported by the system. The mechanics in Elric! supported this game concept in a way that Runequest or Gurps would not.

Thirdly, my current game. Characters began as part of a peasant levy. A great battle has just taken place which they lost badly, made outlaw they will have to take to brigandage or become mercenaries to survive.

Gurps might work fine for this except that like my Roman Britain game the characters are meant to be pretty useless at the start. Elric! would make them far too powerful. Runequest would be ok, but I want the focus of the game to be on the community they build around themselves of fellow flotsam, the bandit/mercenary band they come to lead. Also, their increasing understanding of the national politics which uprooted their lives and led them to this life is a key feature, the transition from peasant pawn to active participant in national life is important to the campaign concept.

So, I use Pendragon. Pendragon has several explicit game mechanics designed to assist in the running of a game where characters routinely lead bodies of men in battle. There are developed and streamlined skirmish rules. The feel of the mechanics fit the downbeat setting well. Now, after four weeks, the characters lead a band of 20 men and the game mechanics I have chosen actively assist in the running of a game where the characters are responsible for a body of followers and where battles rather than melees are the norm. Pendragon also of course supports gaming involving politics, raising of armies and feudal economics.

Again, system supporting priorities in play. I have to head off right now, but later on I'll give one more example. Why I used OtE for a sim game revolving around the machinations of the Illuminati in modern day Vegas and why Gurps would not have worked for it.

My OtE addendum.

I wanted to run a game of Fortean weirdness, involving classic Illuminati elements. Characters would ideally be normal people, some perhaps with a tinge of oddness to them.

Using Gurps, it is laughably easy just to put half a point into a guns skill or brawling or somesuch. Characters again just tend to be too competent in too many things. You can restrict the starting points but then everyone has lousy stats and they become a bit too average.

CoC works fine for normal people, but I wanted play to progress quickly and it's not always the quickest game engine in resolution terms. Also, the mere fact of using it would predispose players to wondering if I was going to bait and switch and suddenly introduce Cthulhoid elements.

OtE supported my campaign goals though. Quick in play and with a freeform chargen process which admirably supported "real" people. I ended up with a Capoeira instructor, a presidential press aide, a British police seargeant and a rave dj. A great mix of people one could actually encounter in real life.

Now, OtE has many narrativist elements. It supports sim play poorly in many respects. But for this particular sim game it worked well as it supported the kinds of characters who were key to my game concept.

Put another way, system does matter, even for sim.