Wednesday, November 10, 2004

Figures Don't Lie, but...

In the email mentioned in the previous post, I also made a prediction that Andrew Sullivan would take heart in a decrease in casualties from 85 to 65:

"Either way, Glenn Reynolds and Andrew Sullivan take heart over the fact that the number of coalition troops killed in hostile action is down from 85 in November to 65 in December. "Obviously," writes Andrew, "we are wearing the resistance down. By the way, did I mention that I am gay?" Glenn Reynolds posts Andrew Sullivan's entire article on his blog, and then adds his own eloquent and descriptive observation:

"'Indeed. Interesting. Read the whole thing.'"

For those of you who didn't get the sarcasm, remember that this was written in July (2004). At this point, other than the unusually high casualty months of November 2003 (94 hostile fatalities), April 2004 (131) and May 2004 (65), which was really in a sense just the tail end of April, there had been 45 or fewer hostile fatalities per month since the end of major combat operations (May 1, 2003), and fewer than 40 from March 2004 backwards.
The point was that the overall casualty trends, p[articularly in regards to hostile-fire fatalities were going up, but that a temporary one-month decrease would be seen as significant.

Andrew, to his credit, does not seem to have been fooled in this way, however, The Belmont Club appears to be quite enamored with this sort of analysis, where the time periods under review are specifically chosen so as to obscure long-term trends.

No comments: