Larian's Swen Vincke Talks Choices & Consequences

Larian Studios' Swen Vincke has spent some of this time to blog on the subject of choices and consequences in role-playing games. As always his style is pretty frank, making for a pretty interesting read. Here's a snippet:
Some time ago I asked Octaaf, our lead QA to give me a rapid overview of the state of the art in choice and consequence gameplay, the idea behind my request being that I wanted to compare what we had planned for project E and Dragon Commander to what our competition was doing. Here's what he told me:
    '¢ Skyrim 244 quests, 300 points of interests, with about 150 dungeons. Most quests are of the (go do this) type and while there are different methods of doing things, the end result is usually the same. The game is divided in two major factions and your opening quest opens one faction tree and closes the other.
    '¢ Dragon Age 2 49 main quests spread over 3 acts, 110 side quests categorized according to importance (secondary quests, side quests and companion quests). Companion mechanic allows for some non-linearity. Branching quest trees. When you finish an act, most quests of the previous act aren't available anymore.
    '¢ The Witcher 2 8 possible endings, subdivided in 2 main categories depending on whether or not a main character survivives, with further variation achieved via the survival or death of two lesser characters. 54 steps in main quest spread over three chapters. 42 side quests. Most quests are linear, with quest failure not having significant impact.

I don't know how he got the numbers, nor did I check if they're 100% accurate, but they were sufficient for me to do the comparison I wanted to do and more importantly, it told me the one thing I really needed to know nihil novi sub sole, nothing newer under the sun.

The numbers of quests for these three games are low enough that you can immediately see that it's impossible for them to have real choice and consequence in them, and that any claim to the contrary is just marketing.

That doesn't demean them, because it'd be productionally insane to offer real branching storylines for every single choice you encounter in a game. Including just ten choices like that would lead to 1024 different endings, which in addition to be very expensive to make, would also be nearly impossible to QA.

So what to do if, like me, you want to make a game that presents you with plenty of real choices ? The answer of course is to make them indepedent of one another. After all, the branching problem only occurs if the option of making one particular choice is dependent on a previous decision being made.