His ideas about decision points (as in thinking about them, ensuring your encounters have them, and ending encounters when your players are out) are sound.
However, he resorts to abstract distance. While a common technique, in the context of D&D (especially 3e or 4e) it's kind of a cheat.
Why?
Because it is the combat system of D&D that sets strict limits on the fleeing assassin's chances of escape.
What I mean by that is, sure, if you run the encounter like Angry DM where assassins are "five steps ahead", then her chances of escape is however high (or low) as you the DM feel like it.
But you are now
not using the rules of D&D. You are, in fact, using made-up rules that the players won't be masters of, much less know of at all. And much more importantly, you have made it impossible to translate certain character abilities. (Such as that thief of yours who move twice as fast)
And by doing so, you've just eliminated the vast majority of pitfalls that the thread starter is actually asking about.
For some groups, this is a good thing. But to actually answer the question, that is "does chases work in the context of D&D, specifically its combat/movement framework?" we need to
actually use that.
---
Actually answering the post needs to be out of scope for this particular post, but let me start off the discussion by saying that any party not of the lowest levels will cover much more ground than you expect.
So, unless you want the assassin to be caught immediately and trivially, start off the encounter at a way longer distance than you think you need. And add many more obstacles and challenges for the party to overcome than you think.
And still, if you really need the assassin to escape, or even escape for a while, you need to play dirty, with hired thugs, lookalikes meant to divert the heroes. And don't forget about magic - once play hits tier II, the assassin could and should have some sort of magical assistance to make it just a little bit harder to catch her.