I think it would be entirely possible for the villain to have mapped out his escape route... it's also possible that he does in fact know the shortest route to a point in the city. You're making assumptions that were not presented... we have no way of knowing the why of his knowing the shortest route... only that it had been important enough for the GM top note it in the particular encounter.
As to the more general thrust of your post above...who is arguing exact simulation vs. narrative. I'm puzzled as to how to respond to this as I don't believe I've been arguing for sim vs. narrative in the discussion.
Some things just are... there's no conflict, one thing is and one thing isn't... I certainly have things in my notes that aren't going to change because a player suggested changing it.
I'm advising against a playstyle thought process.
The GM technically decides in all cases. If you want more variance in the solutions the PCs can pursue, the LESS you document, plan and take note, the more open and effectively random the possible solutions.
Thus, of all my NPCs that will EVER run away from the PCs, only a small % of them should have the perfectest route across a complex city map. Mostly reserved for the super-genius PCs (so as to simulate an NPC smarter than me or the players). Everybody else, "tries" to take the best route and it is assumed they do, until challenged by the PCs and they have to prove it. Just like the PCs would have to prove it if the roles were reversed.
In a way, I am challenging absolutism of the GM's notes. There's no way his notes actually cover every detail. Nor, that just because the GM can declare the NPC does something successfully, does it mean that is actually fair resolution. Up until it makes player contact, it certainly expedites things.
But once it is directly opposed, maybe the GM does NOT have that right.
Change the question to "Is there an Apple cart along the street?" The player might want to throw an apple at the NPC or steal one. Or simply buy one so he can eat it while staring at the direction the NPC went.
In this case, there's no contest. The player has a valid question that the cart might exist. But there's no real debate on whether the AppleMan could have known to park his cart right there or not by GM fiat or player whim. It's there or it isn't and up until the PC asked, it wasn't even important.
In the case of the chase, as a player, at the moment I'm pursuing him, if I have to make checks to navigate the city efficiently, so should the NPC. The GM by-passing that better have a good reason.
JC said he prefers a more simulationist than narrativist. He's certain open to "creating content" when players ask about things not on the notes.
I would propose that certain other things should NOT be in the notes. thus, the notes should not say:
NPC takes the optimal path from point A to point B while the PCs pursue him. The PCs can only stop him if they out-run him or magic.
they could say:
NPC takes the best path he knows from point A to point B. If the PCs pursue him, make such-and-such checks as needed.
For me, a lot of times, I'm hesitant to put that much. I may detail NPCs and places, but I may avoid putting them in one specific spot (unless they are mostly stationary). thus, I will put NPCs in-game, where it makes sense. As such, I don't know or plan on NPC being at point A, for me to know when I need him to run to point B (because he may never have need to run, running is the "new" idea that occurs to the NPC).
I don't know what this concept should be called. But using the system to determine actual events seems a simulationist thing to do. using notes to declare absolute events seems what a narrator would do (hence being narativist). Note, I used "events" and not "facts".
the apple cart is a fact. It's either on 5th street or on this street.
the villain carrying a rod of evil-stuff is a fact.
The players may not like it, but generally, the GM does get to freely determine where stuff is (the apple cart is at 6th street today, and the villain owns a Rod of Evil-Stuf and carries it on his person)
the villain racing acros town is an event. It is someting that he intends to do well, but may be contested by the players and the rules.
My reasoning on handling events has plenty of holes. I don't expect the GM to roll every attack and skill check that happens off-camera from the PCs. So when the villain murders the mayor, sure, that's an event, and the PCs sure didn't want it to happen. But they weren't around in anyway that would make me have to prove how it went down because they might interfere.
I think timing would be a cue as to when this principle might apply. If the PCs are sleeping, and the villain goes to the mayor for a secret meeting, and then kills the mayor. There's nothing the PCs could do to interject into round 3 when the mayor actually dies. They weren't even around to be part of initiative. So they couldn't be racing to th spot to get there by round 2 to change the outcome. Therefore, there's no need to run the murder as a combat.
But in the race from A to B (presumably hot on the tail of the NPC), there's a lot of factors at stake. are both parties using the "city" map that does not show alleys as known pathways? Are both parties abstracting the city to a couple die rolls?
it certainly wouldn't be fair if the GM said the NPC was using alleys and backways that weren't on the map to run the perfect route, but the PCs have to roll for it, or worse, stick to the map, which doesn't show those. especially with a PC who "knows" the city.
Now we've kind of looped in topic to the beginning of the thread.
I would propose for GMs:
allow for both NPC and PC to use the same mechanisms to determine sucess
be cautious of over-detailing events
consider that some GM info is possibilities rather than fact, thus opening options for alternate player solutions.
