Recently I watched the season opener of Arrow and while I found it enjoyable one of the things that bugged me was the final scene, which involved the foreshadowing of a future death. That in and of itself didn’t bother me (beyond the usual “great, which of my favorite characters is going to die?”); what did bother me was the official commentary afterward that indicated that even the show writers don’t know who is going to die.

Say what?

I’ve had similar issues with other TV shows in the past: Enterprise’s Future Guy and what was really happening on the island in Lost, just to name a couple. It just burns me up when I’m trying to make connections and piece clues together only to find out that the writers are just throwing spaghetti against a wall and running with what sticks.

But then again, mea culpa.

In the past, this was exactly my GMing style. I’d set up interesting hooks and situations without a strong sense of where things were going; I just made it up as I went along based on what the players were doing. For the most part I kept this to myself and the players were often amazed at how things turned out, but if they looked too closely at what went before they’d find plot holes big enough to drive a truck through.

I once made the mistake of sharing this with a player and I was taken aback by her reaction. She loved investigative adventures, but knowing that I was “making it up as I go” ruined them for her. From that point forward, I tried to run “pure” mysteries with everything worked out in advance. I found this a bit constraining, especially when I’d made things a bit too easy and the players jumped to the end. That made for a satisfying Sherlock Holmes short story, but when I hoped that the mystery would last a whole session or three it was taxing to have it wrapped up in two hours.

As time went on I settled on a compromise. I work my plot threads out in advance but leave enough room to change details as I go along. If I do decide to change things (usually because I’ve figured out how to shore up a shaky plot element or – more often – the players come up with a better idea), I make sure that nothing I change invalidates what’s already occurred at the table.

So how about you? Are you a spaghetti GM or do you set things in stone? Are you somewhere in the middle? Has something ever gone terribly because of your mystery style? If you tend to swing one way, would you consider swinging towards the other?