I've been hearing people cite a "77% drop" in American casualties in Iraq during reporting today. 77% since when, exactly? The helpful people at icasualties.org have compiled this chart showing American soldiers killed in Iraq from the invasion up until a little over a week ago:
You can see an expanded version (that may be more up to date, depending on when you check) by clicking here. They've labeled major events in the war, including Fallujah, the "surge," and so forth.
The big conclusion I draw from this chart, viewing the war effort as a whole, is that I am unwilling to cite any immediate drop in casualties as an indicator of success, or even hope that things will stay that way. It's hard to draw a clear correlation between our strategy such as we know it and American casualty numbers. I can point to a sharp drop-off in deaths from June of this year until now, but I can point to the exact same drop-off between May and July of 2004...and that was followed by a return to May 2004 death rates in roughly December of that same year...and then followed by another substantial drop-off, and so forth.
In short, the current trendline is by no means predictive. It's as if Bush supporters pointed to an off-season cold snap as proof that global warming had been turned around by voluntary emission controls.