Right now I'm thinking about improvisation. In my play-by-blog campaign, I improvised a LOT. I created maps and NPCs, and had a rough idea of what was in each location up front, but little else was planned. Virtually every time the PCs encountered a random NPC a new plot thread was created on the fly. Dozens of times, players did things I hadn't thought of. Other times they simply came up with really cool ideas that I decided to run with (yeah, I'm an idea-stealing bastard).
The latter, I suspect, was way more common than the players probably realize. Many times the changes were small, in the details. Other times the player-inspired changes were gigantic.
For example, there was a large dungeon in the world – Blackwell Keep – that the players opted to explore. As they entered the above-ground ruins of the keep, I gave them a layout of the upper works just for flavor. These "flavor ruins" happened to have a church near the keep. One of the players suggested that the group first examine the church to see if there was some secret way into the dungeons beneath the keep.
There wasn't. Not when I planned it. But when a player suggested it, I thought "Heck, that's a great idea." So I improvised a small set of catacombs giving access to a secret tunnel that led into the keep's dungeons, and which even the keep's denizens were not aware of.
The point being, I improvised a lot in that campaign, and often riffing off (ripping off?) player ideas. The "Blackwell Church" improvisation alone was responsible for the final 42 posts of the campaign, out of 178 posts total – a full 23.5% of the entire campaign.
Of course, in any postal format, this kind of improvisational "oh-cool-idea-I'll-just-plunk-that down-here" is very easy to do. Whenever a player did or suggested something unexpected, I had forty-eight hours to mull it over and "improvise" something (I put those quotes there because when you take forty-eight hours to do it, I'm not sure it's technically improvisation anymore).
But I like to improvise. I do it every day in my teaching. I make my lesson plan. Then I get to class, set my lesson plan aside, and let the discussion go anywhere the students want to take it, until they run out of ways to go, and then I have the lesson plan to fall back on. The best class sessions are the days when the students do all the talking and my lesson plan just collects chalk dust.
The question is really to what extent I can convert what I enjoyed doing so much in a play-by-post setting, and what I do in a non-gaming context in my job, into (virtual) face-to-face play. And just how much prep – in terms of creating a minimal framework to facilitate improvisation – I might need to do to make it work.