Missing the obvious?

Gamasutra has an article up about the data mining Valve has been doing through their Steam service.

While the article is called "Steam Stats Surprise Over Episodic Content", there's something else I found more curious.

The stats they've collected show a few interesting things:

• The average amount of total playtime players give Half Life 2 is 2 hours and 48 minutes.
• The average play session lasts 37 minutes (most of which is probably just loading the levels).
• Almost 75% of the players play at Medium difficulty.
• For those that do complete the game, the average time it took them was 4 hours and 56 minutes.

  So they were surprised to learn that only 22.87% of the people who downloaded the Episode One expansion actually completed it. Further, only about half the people who reach the final map actually complete it.

  Their theory?

  This leads us to believe that either players are quitting before they see the credits, or there is a bug in how we collect this data.

  (presumably the data is collected during the Credits sequence).

  What I find surprising is the more obvious questions they could me missing. Like:

  1. Maybe the last level is boring?
  2. Maybe it's too difficult?
  3. Maybe it's too difficult in the wrong ways, like a mob that acts stupid but has insanely-high HP regen rate (sound familiar?)

  Granted, it's possible that there's an error in either how or when they collect this data too. But for them to miss the obvious implication that perhaps there's something wrong with the game play makes me wonder who's actually reading the numbers.

  It wouldn't surprise me if the designers of Episode One first learned of these stats from the Gamasutra article. I've seen and heard of many cases where the people who do the work are never really told the results of them officially. In some places, even doing post-mortems is unheardof, so focused the companies are on the next big thing. Sometimes it's innocent, this belief that everyone automatically learned something and automatically seeks to improve themselves. Other times people just don't want to discuss failure at all because it makes them look bad.

  That's what I wonder in this case. Are they only focusing on their one theory because they assume the game was perfect?