Your analysis is mathematically correct, but meaningless. What was happening in Iraq four years ago bears absolutely no resemblance or relevance to what is going on now. Force levels, weapons, training, strategy, tactics, and operational realities have all shifted significantly, as have counterforce levels and operations. Casualties, being a product of all of these things cannot be understood solely by quantitative analysis.
The reason that US casualties is down between May of this year and today has little, if any, correlation with past history and everything to do with the military elements of the surge. A historical example will suffice: US military casualties were far higher in the European Theatre of Operations in 1944 than they were in 1943, and much higher than in 1942, trending upward beginning in June '44 and continuing upward through January/February of 1945. We must have been losing the war, right?
Excuse me. You were the one who said the numbers were trending down. To say that the past has “absolutely no resemblance or relevance to what is going on now”, is incorrect; today’s decisions are a consequence of yesterday’s actions. That said, I still do not see how you can support that statement that there is a downward trend. Comparing months from 06 and 07, May 07 had 1.8x more deaths, June had 1.7x, and July had 1.9x as many. Hardly a down trend.