First the graph:
Again the caution: The mid-month numbers are somewhat useful since we are dealing with a daily average. However, the number could still drop down near .5 deaths per day (if no more occurred this month) for August or could increase substantially if a great number of troop deaths occur later in the month.
What do the numbers mean?
Generally, the graph suggests that Iraq was slightly more dangerous for coalition troops in August (so far) than in July--about as dangerous as June.
On one hand, a mild increase in the numbers could perhaps be expected because of the upcoming provincial elections. The enemy may be expected to try to disrupt that process. Even though it might make more sense for the enemy to try to foment sectarian violence rather than attack foreign troops, our troops will tend to have more engagements when the enemy attempts a greater number of attacks.
On the other hand, the numbers have grown so small at this point that what might look initially like a trend may well be simply fluctuation around a fairly stable level of risk. Separating non-hostile troop deaths from those resulting from hostile action will help make the numbers more meaningful as well. I will begin to keep track of separate columns for those differing stats and provide an update at a later time.