Afghanistan: decision time for Obama
I have a question for you: where have more US military personnel died this year - in Afghanistan, or Iraq?
Afghanistan, of course, is the right answer: 297 deaths so far this year, compared to 144 in Iraq. (There have so far been 98 UK military deaths in Afghanistan.)
But it's also the wrong answer. Because more US men and women in uniform have committed suicide this year - at least 334 - than have died in either Afghanistan or Iraq.
I mention it because it's worth taking into account as we prepare for President Obama's announcement next Tuesday evening (Wednesday morning if you're in the UK) on his plans for future military deployments in Afghanistan.
He knows that for tens of thousands of American military families - and for many, many more who live in their communities - what matters is not only how many men and women are killed in action, but how deep are the scars, both physical and mental, that they bear long after they have returned home.
So my hunch is that the President will present his decision next week as a strategy for getting out of Afghanistan. This, he will say, is what we intend to do so that we can leave the place to its own people, knowing that we have given them a decent chance of running it themselves.
I suggest that you look not so much at how many extra troops he's decided to send (32-35,000 seems to be the current best estimate), but where he's sending them and what he's asking them to do. Because according to many analysts, there's now a growing realisation in Washington that killing Taliban fighters doesn't get you very far.
One of the most common questions that policy-makers get asked when they're making decisions about military deployments is: "How will we know when we've won?" After all, no one expects the Taliban to sign a formal surrender document.
So, the usual answer is: when the people of Afghanistan can be relatively confident that they and their families are secure, and when there is a degree of political stability that looks likely to last.
Take a look at how other insurgencies in the region have been tackled. According to Paul Staniland, writing on the website ForeignPolicy.com, the usual deal involves "messy and ambiguous bargains that states make with armed groups and local political actors combining accommodation, coercion, bribery, and coexistence."
He calls it "ugly stability". "The government accepts that insurgents will continue to control parts of their own community, but insurgents know that pushing the state too hard can trigger a crackdown. Governments flip over some former insurgents to act as pro-state militias, insurgents and warlords sponsor normal politicians, and both sides become linked to peripheral war economies. A strange but often enduring quasi-stability can persist, whether in Karachi, the Bodo hills, or Nagaland."
In other words, it's not anything like what you'll find in Westminster or Washington, but in a way, it works. And it's an approach that closely resembles what a US army special operations officer, Major Jim Grant, is reported to have outlined in a paper called "One Tribe at a Time: A strategy for Success in Afghanistan."
According to Fred Kaplan, of Slate.com, Grant's premise is that Afghanistan "has never had a strong central government and never will. Its society and power structure are, and always will be, built around tribes - and any U.S. or NATO effort to defeat the Taliban must be built around tribes, as well."
So is this the picture that's emerging? Forget all that stuff about democracy and women's rights - what this is about now is getting out as quickly as possible without leaving behind too much of a mess. According to an increasing number of analysts, that's likely to be the best offer available.