An Iraqi policeman sits on rubble after a bomb attack in Baghdad. (Ahmad al-Rubaye/AFP/GettyImages)

There are two pieces of prevailing conventional wisdom about the Iraq War: first, that the U.S.-led invasion was a mistake and, second, that things started turning around in 2007, when the U.S. changed strategies, Iraqi politics sort of came together and the death toll went down dramatically. That, in the most general terms, is the basic American narrative of the Iraq War: from very bad to less bad.

An extensive new study into Iraq's war deaths makes that storyline look a bit different. The study, conducted by four American and Canadian universities along with the Iraqi Ministry of Health, estimates that 461,000 Iraqi deaths can be attributed to the war, of which about two-thirds were direct violence; the other third is from indirect causes such as breakdowns of health-care systems that are attributable to the war. Those numbers are much higher than earlier estimated. Worse, they tick back up in the final years of the war – when things were supposedly getting better.

The results, charted below, show the number of war-related Iraqi deaths over time. It's grim – and a direct challenge to our understanding of the war as having improved after 2007:

(Source: University Collaborative Iraq Mortality Study)

This study confirms two components of the Iraq War narrative: that fighting dropped sharply after 2007 and that 2008 was a relatively successful year in reducing combat deaths. You can see that the dark-red bars get smaller, meaning there are fewer deaths from direct violence. But this chart still contradicts our overall understanding: It turns out that deaths picked back up in 2009 and then even further in 2010, to 2005 levels. The gain was temporary and in the process of reversing by the time we left.

The study cuts off at the war's formal end in mid-2011, which is why that bar appears so short, although fighting has continued.

The research is more rigorous than previous assessments of Iraq war deaths; it relies on randomized household surveys, meaning they asked regular Iraqis to recount if, when and how members of their household died during the war. It's also different because it looks at indirect deaths that, the authors argue, can be attributed to the war. We do forget that the invasion, occupation and subsequent sectarian violence led basic services to collapse. One of the authors pointed out to NBC News that neighborhood-to-neighborhood fighting often prevented Iraqis from seeking medical treatment.

This chart also complicates our understanding of the war. The y-axis, which is labeled with that confusing "45q15" notation, indicates the probability of dying between age 15 and age 60. In other words, this chart shows the chance that an Iraqi man or woman, age 15 to 60, had of being killed due to the war.

(Source: University Collaborative Iraq Mortality Study)

Two trends should jump out at you right away: The war got a lot less dangerous for Iraqi men, but it stayed about consistently dangerous for women. There seem to be two most-likely explanations for that. First, a lot of men who might have fought were killed in the war's first five years, so by 2008 there were fewer of them around to fight and die. Second and perhaps more plausible is that the nature of the war changed. For the first several years, violent deaths tended to come from combat, which disproportionately kills men. But over time, more of it came from bombings and other acts of terrorism, which target civilians and thus have a high probability of killing women. This sort of violence has been getting much worse since 2011.

Iraqis who were contacted for the survey blamed U.S.-led coalition forces for 35 percent of violent Iraqi deaths, blamed Iraqi militias for 32 percent of deaths and "criminals" for 11 percent. Another 21 percent said the killer was unknown.