I'm running behind schedule with my recap of the 49ers' defensive stats this season, so in the meantime I'll wet your whistle with a post I came across by football probability guru Brian Burke of Advanced NFL Stats. For those not familiar with Burke's work - rhyme not indended there - he's the probability yin to Football Outsiders' (FO) yardage yang. The underlying principles are the same (i.e., down, distance, & game situation are indispensible; as is having a massive sample size), but, whereas FO focuses indirectly on winning by measuring the efficiency of yards gained on a given play, Burke focuses directly on winning by measuring the impact yards gained on a given play have on a team's in-game probability of winning.
In the post I've linked to above, Burke compiles 2009 rankings for all QBs in relation to
- Win Probability Added (WPA) - the extent to which a QB increased (or decreased) his team's chances of winning, which is based on the historical probability of winning a game given a team's down, distance, and game situation.
- Expected Points Added (EPA) - the extent to which a QB increased (or decreased) his team's expected point total, which is based on the historical number of points a team can be expected to score given the down, distance, and situation.
He links to much-more-detailed explanations of these stats just in case you're not following at the moment. The bottom line for our purposes is that, regarding the 49ers' QBs in 2009
- Shaun Hill increased the Niners' win probability by about 50%, whereas Alex Smith decreased the Niners' win probability by about 150%. So, in 2009, the 49ers were about 200% more likely, or twice as likely, to win with Hill at QB.
- Hill increased the Niners point total by about 1 point this season, whereas Smith decreased the Niners' expected point total by about 17 points. So, in 2009, the 49ers were about 18 expected points better with Hill at QB.
- The Niners' win probability increased an average of 1% for every 17 plays involving Hill rather than Smith.
- The Niners could be expected to score an average of 1 more point for every 25 plays involving Hill rather than Smith.
- In 2009, Hill was about 9 times more "clutch" than was Smith.
So, give Burk's post a read. It's not that long, and it's very informative if you're into stats in general or advanced football stats in particular. I'll have my review of the 49ers' DEF this season up tomorrow morning. In the meantime, let the Smith vs. Hill melee begin.