So I was watching the Hobbit. Though it is not the only example of this type of combat. Fellowship also had a running fight or a fighting retreat. Lots of movies do.
So how would you other GMs handle your adventurers running afoul of a large number of enemies that are willing to try to run them down. Generally when an enemy or a PC runs it boils down to who can move faster and get into a position to attack.
In films the formula is to throw many enemies at the heroes who facing huge numbers flee even if the general skill of the enemies is not on par with them. They opperate on the real world assumption that anyone can get a lucky hit and kill you. So the heroes flee.
In PF there is a point at which low level enemies virturally scease to be a threat at all. So there is little reason to flee.
So as I see it there is two problems:
1- Action economy- PF action economy favors generally one room or general space (a field or clearing) so running full tilt and fighting off attackers is not something normally you are able to do.
2- Space- If you start one side of the battle map and everyone is running really fast they will run out of space quick.
I suppose it would be possible to do a kind of chase mechanic hybrid with attack actions thrown in.
If you get into a situation where one group is fleeing from another and they aren't necessarily "in combat" anymore, you can always resort to skill or ability checks to see who "wins" the pursuit.
Besides the obvious ability (character movement speed) you can introduce additional elements by having terrain involved, where higher dex gives a bonus in moving through the difficult terrain, or climb checks if the pursuit is through mountains or hills.
In 4e they do this with 'skill challenges' where the GM is supposed to identify something like five potential skill checks and then whoever scores three "wins" first wins the pursuit/retreat.
I am not a fan of the specific implementation of skill challenges in 4e, but I think the basic idea has merit and have used it in my PF games.
The boring way to do this is by staying in rounds, and having one player move, then one enemy move, then the next...
While it's by the rules, it's also entirely based on movement speed. Faster group wins, every time. And if both groups move the same speed, it goes on for infinity. Not super fun.
My advice is to find some chase rules you like. I've used the Chase deck, made my own custom chase cards, and used the Classic Spycraft rules for vehicle chases, all with success.
Most important, share your plans with the players BEFORE the chase begins. That way, everyone knows the rules, and players can prepare for it. That ranges from how to spend skill points to what spells to prep to what equipment to buy.
If you don't, you may find that you have a rules subsystem that some players can't participate in. Not fun.
First of all ignore the grid, unless you want to use it show who is ahead of whom. No tactical movement. Think of yourself as the director of cinematography. Describe the action and make it exciting. Like AD mentioned if you are familiar with skill challenges then you can use them thematically, but never mind the rules as written. Reward fast characters of course (hello fast movement), but also reward teamwork (paladin pushes bard up to a ledge), cunning (the party quickly hides then doubles back), daring (swing across a chasm), and creative uses of existing abilities (wizard casts telekinesis to drop rubble onto path behind the party). You might have to resolve a couple of attacks when someone falls behind but don't let it bog down the action.
Even if the party doesn't get away, they could make it to a location where the terrain works more to their advantage.
The first step is to find players who are actually willing to run away from monsters. That's a rare thing it seems...
Here is how I try to do it. Once the situation becomes a chase, I start trying to set up checkpoints where activities can occur that might affect the outcome. For example, if the terrain changes, that might give the chased party a chance to create a delaying tactic, or the chasing party to try to exploit a new path.
If there are multiple types of terrain available, I will try to let the party know what sorts of terrain they can potentially exploit. A party with a druid or ranger fleeing into forest might get an advantage.
I might have the fleeing party stumble across another party with mounts to give them an opportunity to decide whether to steal the mounts. There might be a place to hide, or a more defensible location where the fleeing party can prepare an ambush. If the chase is going into the hills, I might say "There are loose rocks and boulders all around you." Giving the party a chance to decide whether to attempt to start an avalanche.
What I try not to do is make it all about rolling dice. I do my best to give the party a chance to make a conscious decision that would have a definitive impact on the chase.
The first step is to find players who are actually willing to run away from monsters. That's a rare thing it seems...
In my experience players running from monsters is more common than monsters running from players. But both of those can occur.
I find that the knowledge that all encounters are NOT tailored to the party ecl does the trick of making running away aka the tactical retreat a necessary option for survival.
A sandbox with set encounter pieces and random encounters based on location type and not ECL makes knowledge skills a survival necessity.
The first step is to find players who are actually willing to run away from monsters. That's a rare thing it seems...
Ha! A fair point! I believe that this has to do with game balance, and the belief that all encounters are designed upon the level of the group, such that every encounter should be possible to win. As for myself and my group, we rarely run away. There are occasions where it is the best thing to do. It is usually preceeded by a couple of rounds of nervously tapping our fingers and glancing at each other around the table.