17. Awakenings

While the maneuverings over COIN and the surge played out on the home front, something was happening at the epicenter of violence in Iraq.

Seventy miles west of Baghdad, in Ramadi, the capital of Anbar Province, a city of nearly a half million people where all pretenses of order had long since been abandoned—where Sunni insurgents ran free and al Qaeda gunmen enjoyed unchecked control—suddenly, a mere six thousand American troops were turning the war around through the classic methods of counterinsurgency.

The reversal came to be called the Anbar Awakening, but it rose and rallied so abruptly and rapidly in the fall of 2006 that the most alert observers, back in the States and even elsewhere in Iraq, barely detected the rumbles.

And the Awakening’s chief strategist, Colonel Sean MacFarland, a brigade commander with the 1st Armored Division, was equally oblivious to the political struggles going on back in Washington.

Yet the two breakthroughs—the rise of COIN doctrine in American policy circles and the success of COIN practice on Iraq’s most dangerous landscape—arose from the same set of ideas, networks, and influences: the classic books, the West Point nexus, the lessons learned or relearned from the small wars of history, David Petraeus’s experiment in Mosul three years earlier, and the mediation of all these factors through Colonel H. R. McMaster.

McMaster’s role in this shift was indirect but pivotal: when his 3rd Armored Cavalry Regiment pulled out of Tal Afar the previous January, it was Sean MacFarland’s brigade that moved in as the replacement.

During the transition, McMaster gave MacFarland a complete tour of the area, introduced him to the players, briefed him on every aspect of the clear-hold-build campaign that had pacified what was once an insurgent stronghold: the outreach to tribal leaders, the isolation of extremist Islamists from merely devout Muslims, the swift construction of American outposts in the heart of the city, and the generous outlay of cash to buy off fence-sitters and hire workers for reconstruction projects.

MacFarland grasped the concept quickly. Ten years earlier, he’d been one of the first squadron leaders to separate the warring sectarian militias in Bosnia as part of the effort to enforce the Dayton peace accords—a task that had forced him to figure out who was in charge, arrange power-sharing schemes with local factions, and help reconstitute a local government.

He’d also long been an avid reader of military history, first as a cadet at West Point (class of 1981, three years ahead of McMaster), then as one of the elite junior officers at Leavenworth’s School of Advanced Military Studies, where he pored over T. E. Lawrence’s insurgent memoir and every chronicle of the war in Vietnam—books that well disposed him to John Nagl’s Learning to Eat Soup with a Knife when it was published a few years later.

In short, COIN was not at all a new concept to MacFarland.

During his command at Tal Afar, he built on the foundations that McMaster had laid, getting a feel for what this kind of warfare involved: the balancing act of offense, defense, and stability operations. In late May, when his brigade was transferred to Ramadi, he set out to apply the same approach.

Initially, it was unclear whether these principles would be relevant. When McMaster came to Tal Afar, there was at least a functioning city government; not so when MacFarland arrived in Ramadi. Tal Afar’s population was a mix of Sunnis, Shiites, Turkomen, and Kurds, so McMaster could play them off one another; Ramadi consisted almost entirely of Sunnis, many of them well-armed Baathists still angry at Saddam’s ouster and their own loss of power. Tal Afar was small and surrounded by desert, so McMaster could keep out foreign jihadists by building a wall around the town; Ramadi was much larger, it was surrounded by suburbs, and jihadists were already dug in and allied with the more militant tribal elders.

MacFarland’s brigade was part of a multiservice force that included three other Army battalions (which, together, were roughly the size of another brigade) as well as a Marine regiment and a few teams of Navy SEALs. Still, they didn’t add up to much. To amass a force with any potency, he would have to recruit local police and soldiers to join the Americans’ side in a very hostile environment. In Tal Afar, MacFarland had benefited from a thick binder that McMaster left behind, listing the tribes, the key players, and all of his contacts. In Ramadi, he had nothing; he and the other units would have to start from scratch, one neighborhood at a time.

As a first step, the “clear” phase of clear-hold-build, MacFarland launched an offensive against some of the heaviest al Qaeda strongholds. Once he secured those areas, he set up combat outposts, like those McMaster had set up in Tal Afar: small bases inside the neighborhood, where he and his troops would live, patrol, and protect the population from any returning jihadists. From there, he established contact with tribal leaders and swayed them to help screen and recruit a police force, whose members he armed and trained. MacFarland also started reconstruction projects right away, in some cases while the fighting was still raging, again using the tribal sheikhs as the legitimizing funnel to direct the spending. (In COIN, as Petraeus had said in Mosul, cash was a form of ammunition.)

In establishing these contacts, MacFarland was fortunate to have on his staff an outgoing former Special Forces soldier, Captain Travis Patriquin, who was conversant in Arabic. Patriquin served as interpreter, negotiated the payment schemes, set up the meetings with sheikhs, got to know their families, and earned their trust.

One thing Patriquin learned quickly was that the townspeople, even some of the most militant Baathists, were turning against al Qaeda; the jihadists, it seemed, had overstepped their welcome, robbing businesses, stealing away the daughters of prominent families, and assassinating the parents who refused to comply. A year earlier, some sheikhs and Sunni nationalists had formed a resistance group called the al Anbar People’s Council; but, outnumbered and racked by internal disputes, they were crushed, many of the group’s leaders brutally murdered. The survivors were primed to try again, but only if the Americans promised to stay for the fight till the finish.

MacFarland gave the guarantee, though he had no authority to do so. He was freelancing a fair bit of this whole campaign, as Petraeus had in Mosul and McMaster had in Tal Afar. General Casey paid a visit once to see how things were going and was clearly skeptical. Casey’s larger campaign plan leaned toward pulling out, and here was MacFarland burrowing in. Casey also worried about the idea of a neighborhood watch program on its own terms. It was one thing to encourage reconciliation with Sunni tribes; but MacFarland was arming militiamen who had been killing American troops just a few months earlier and—who knows?—might resume killing them, this time with American-supplied weapons, a few months hence.

Casey’s doubts were fueled by an August 16 report by Colonel Peter Devlin, the US military’s intelligence chief for the province, declaring that the American forces were “no longer capable of militarily defeating the insurgency in al Anbar.” The previous few months, he wrote, had seen “the near complete collapse of social order.” Al Qaeda had emerged as “the dominant organization of influence” and was viewed by many local Sunnis as “their only hope for protection” against fears of “ethnic cleansing” by the Shia-dominated central government—fears that had some basis, as General Chiarelli was witnessing in the bloody streets of Baghdad around this time. Devlin concluded that “there is nothing” the US-led coalition could do to change this dynamic.

Yet this dynamic started to change less than a week later, on August 21, when al Qaeda fighters attacked a new police station with overwhelming force, and the police fought back. Hours later, the jihadists upped the stakes by murdering Abu Ali-Jassim, a particularly beloved sheikh who had encouraged his tribesmen to join the resistance, and then—in a blatant violation of Muslim law—dumping his body in a field to rot rather than bringing it home for a proper burial. The jihadists had gone too far, and several sheikhs, who until then had been reluctant to ally with “infidel occupiers,” flipped to the Americans’ side.

A new figure emerged in the backlash, Sheikh Abdul Sattar Abu Risha. Sattar was a shady character who had earned much of his wealth through smuggling. But one of MacFarland’s battalion commanders, a jovial, cigar-chomping lieutenant colonel named Tony Deane, happened to run across Sattar, struck up a conversation, and came back to the base, telling MacFarland that this was a great guy—he loves America, loves George Bush, and wants to meet the commander. MacFarland went to meet him. It turned out that the jihadists had killed Sattar’s father and two of his brothers; he wanted to sign up for the resistance.

In the aftermath of the August 21 attacks, Sattar became MacFarland’s main agent in the Ramadi Street, the charismatic figure who persuaded other sheikhs not only to rise up against al Qaeda but to do so in open alliance with the Americans. On September 9 he organized a tribal council attended by MacFarland, his staff, and over fifty sheikhs. Sattar called it the Awakening Council. They approved a resolution to expel al Qaeda from all of Anbar Province, form a local government, and reestablish the rule of law.

With the legitimacy of these local leaders, the police force swelled from roughly one hundred officers to over four thousand. The number of neighborhood combat outposts, which were now manned jointly by American and Iraqi soldiers, expanded from four to twenty-four. By November, nearly every tribe, not just in Ramadi but also on the city’s northern and western outskirts, declared support for the Awakening.

The American troops in Ramadi were technically under the purview of a Marine expeditionary force, whose commander, Major General Richard Zilmer, had given MacFarland remarkably free rein in his unconventional approaches. At the end of the year, Zilmer was replaced by another marine, Brigadier General John Allen, who, after getting briefed by MacFarland, endorsed the idea explicitly, built on the Awakening Council’s momentum, and extended the operations, first to Ramadi’s eastern outskirts, where the jihadists retained control, then ultimately across Anbar Province. Allen stepped up the pressure, calling in tank assaults and air strikes to clear al Qaeda’s remaining strongholds. Then, emulating MacFarland, he set up neighborhood outposts to hold the areas, build up local government, and fund reconstruction projects—which, like MacFarland’s, were administered through the sheikhs.

The fighting continued for several months. In early December, a roadside bomb killed Captain Patriquin; more than a dozen sheikhs attended his memorial service. Sheikh Sattar would be assassinated the following September. But by the end of 2006, the tide had clearly turned: violence was cut in half; the incidence of insurgent attacks, which had numbered fifty a day when Colonel Devlin wrote his gloomy intel report, plummeted to a couple of dozen and, by the following spring, to nearly zero. The city was crawling back to somewhat civilized life.

When General Ray Odierno took his predeployment tour of Iraq in early December, he paid a visit to Ramadi. He’d seen the city briefly three years earlier, when he commanded the 4th Infantry Division; now he could hardly believe the change. This was clearly the new model of what COIN could accomplish in Iraq.

David Petraeus traveled to Ramadi soon after assuming his post as top commander in early February 2007. He asked MacFarland if he’d read the COIN field manual. MacFarland replied that he hadn’t. “That’s all right, you don’t have to,” Petraeus said. “You’ve been doing it.”

•  •  •

Odierno took over from Pete Chiarelli as deputy commander on December 14, 2006. But George Casey still held the reins, and his latest campaign plan, issued on December 5, bore no sign of retreat from his previous plan, no acknowledgment of the new developments either in Anbar or back in Washington. In fact, his plan, called the “Transition Bridging Strategy,” ordered a “speed up” in the handoff to the Iraqis. Specifically, American troops were to “move outside all major cities” over the next three months and, after that, do little more than guard the borders and provide Iraqi troops with logistical support.

Odierno was in a bit of a bind. Casey was his boss, but only for another two months, and he knew from various contacts, especially Jack Keane, that his new boss was likely to be Petraeus. Not only was Petraeus certain to take a different direction; his very appointment would indicate that the president of the United States had decided to change course as well. So Odierno and his staff would have to write their own campaign plan, something that Petraeus could adopt, modify, and build upon the moment he arrived. The two couldn’t communicate directly during this time, lest they be accused of subverting the chain of command. But they kept in touch indirectly—Odierno informing Petraeus of the plan’s progress and Petraeus sending Odierno his feedback—through Jack Keane in his well-honed role as middleman.

Drawing up some sort of plan would not have been out of place for Odierno. His formal title, corps commander, meant that he was not only deputy to the force commander (Casey) but also in charge of day-to-day military operations. The force commander’s campaign plan outlined the basic strategy and guidance; one of the corps commander’s jobs was to turn that guidance into a detailed plan of action. The difference, in this case, was that Odierno, anticipating Petraeus’s arrival, was also quietly outlining a new strategy.

The first thing Odierno did was to list the campaign’s goals, the various obstacles to those goals, and some ways to surmount—to defeat, preempt, or co-opt—those obstacles. This was basic stuff, the sort of three-column outline that junior officers learned to do in strategy courses at the Command and General Staff College. Casey’s strategy relied on a similar method. But Casey’s premise was that the occupation had run its course; that there was little the American military could do, or should do, to improve Iraq’s security; that, in fact, the mere presence of foreign troops was making things worse.

Odierno started with a different premise: that there was still one more chance to get the war right.

At the top of a yellow legal pad, Odierno jotted down the main problem: a “failing state.” Below that heading, he wrote that a “large gap” was separating the Iraqi government from its people and that “malign actors” were exploiting this gap. The task was to figure out how US and coalition forces could step in and help fill the gap. He titled the memo “Closing the Gap.”

In the first column, he listed the “needs and desires” of “the Iraqi individual” that the Iraqi government was not fulfilling. Among them: “physical security,” “essential services,” “justice under the rule of law,” and fair “political representation.” He broke down “security”—which, as he’d been learning, was the precondition for everything else—into three parts: “Protect the population, priority to Baghdad”; “Engagement [with the tribes] leading to local cease-fires and popular rejection of extremists”; and “Develop capable, credible ISF [Iraqi security forces].”

In the second column, he listed the groups that were “exploiting the gap”: Shia extremists, Sunni rejectionists, organized crime, al Qaeda, and external forces such as Iran and Syria.

In the third column, he listed several ways that American troops could fill the gap. The main thing was to rebuild the failing state. This would mean helping the government protect the population, reform its ministries to make them less corrupt, provide basic services (including sewage, clean water, and electricity—the elements of Pete Chiarelli’s aborted SWET plan), and reconcile with sectarian foes (as MacFarland was doing in Anbar), so that people would see their government not as “Shia-based” but as a ruling body “advancing the interests of all Iraqis.”

Odierno’s memo was a complete repudiation of Casey’s latest campaign plan. It resembled, in certain respects, the plan that Casey had briefly put in place in the summer and fall of 2005, under the tutelage of Bill Hix and Kalev Sepp, who’d also encouraged him to create the COIN Academy. In short, Odierno was spearheading a campaign to restore a counterinsurgency strategy.

To help him translate these principles into a military plan, he brought with him to Baghdad a staff well suited to the task. He’d been assembling much of this staff through the fall while he was still in Fort Hood, before the surge briefings—and before it was clear that Petraeus would soon be replacing Casey. As a result, back then the two could still talk directly with each other, and they often did. And among the topics they discussed was the staff’s composition.

For his chief of staff, Odierno picked Brigadier General Joe Anderson, who as a colonel had been Petraeus’s 2nd brigade commander in Mosul, the one who had most actively helped him make contact with the city’s tribal chiefs, merchants, and prospective political leaders.

Also at Petraeus’s suggestion, Odierno picked as his executive officer Colonel Mike Meese, the chairman of West Point’s Social Science Department, who’d served Petraeus in that role in Bosnia and Mosul. (After Petraeus came to Baghdad, Meese moved his office to work for him once again.)

And there was Odierno’s own wild card, a short, slight, thirty-eight-year-old British development worker and self-described pacifist named Emma Sky.

Odierno had met Sky in 2003, when she was a civilian adviser in Kirkuk, the city in northern Iraq with the thickest mix of Sunni Arabs, Kurds, Turkomen, and a few Shiites. Sky had volunteered for the assignment from her post at the British Foreign Office. She’d opposed the war but spoke Arabic, had experience mediating conflicts between Israelis and Palestinians (she spoke Hebrew as well), and thought she could help keep Kirkuk from going up in flames.

When Sky started out, she was formally working for Jerry Bremer’s Coalition Provisional Authority, but she was attached to the US Army’s 173rd Airborne Brigade, which itself had links to the 4th Infantry Division, so Odierno would see her when he visited Kirkuk. Her presence puzzled him at first, but the more he watched her in action, the more she impressed him. She seemed in control of everything in her domain: reaching out to all the sectarian groups; coordinating the brigade’s civil-affairs soldiers with AID officials and nonprofit volunteers, and, all the while speaking to any and all military officers, of whatever rank, with brutal frankness, telling them they didn’t know what they were talking about (usually they didn’t), asking why they were doing things that made no sense (most of them had no explanation), and spelling out local political realities with a stark clarity that jarred them into changing course. Some officers resented her, but Odierno liked her style.

When he knew that he’d be heading back to Iraq as corps commander in the fall of 2006, he sent Sky an email asking her to come be his political adviser. Sky was reluctant. She’d gone back to Jerusalem to work on Middle East peace talks, which were going nowhere, but she’d had it with the war in Iraq; she was impatient with the narrowness of American military culture and saw the whole enterprise as hopeless.

In a return email, she inveighed against the war that he’d asked her to rejoin, calling it the most colossal strategic failure in American history.

Odierno wrote back with a question: What should we do about it?

She liked that, liked the fact that he didn’t deny her indictment or get defensive about it, as many officers she’d met would have done. She’d always thought that Odierno asked good questions; that for all his flaws as commander of the 4th Infantry Division, he seemed genuinely curious about how to do things better and about what she was doing at her job. She accepted the offer.

When Sky first returned, she was appalled that so little had changed. During one briefing, she noticed several slides referring to the enemy as “AIF.”

“What does AIF mean?” she asked.

“Anti-Iraq Forces, ma’am,” the briefer answered.

She chewed him out. These people are Iraqis, she told him. How can they be anti-Iraq? They’re Jaysh al Rashideen, Jaysh Muhammad, Jaysh Rijal al-Tariq al-Naqshabandi, she said, listing the names of just some of the groups. If you don’t understand that, she went on, you don’t know who they are, what they want, or why they’re resorting to violence. And if you don’t know that, you can’t come up with ways of getting them to stop.

Odierno took the point to heart. The second slide of his campaign-plan briefing, right after the slide on closing the gap, read:

•  •  •

The term AIF is not adequate to describe:

1. Shia vs. Sunni Sectarian Violence.

2. AQ and AQI [al Qaeda and al Qaeda in Iraq].

3. Sunni Insurgency.

4. Kurdish Expansionism.

5. Shia on Shia Violence.

6. External Influences (Iranian, Syrian, and Turkish).

He was acknowledging that there were multiple threats, each distinct, some of which could be co-opted or played off one another, some of which couldn’t. You had to know which were which before you could figure out what to do with each.

Sky knew little and cared less about counterinsurgency theory. But her knowledge of local political cultures, and her broader background in development, placed her on the same track as the COIN strategists, made her look at the war in a similar way—as a contest for the people’s loyalty, a contest that, therefore, had to be fought in a way that honored their interests, offered them livelihoods, and ensured their security.

In the first few weeks of 2007, Sky used her facility with the language and familiarity with the culture to open doors all over Baghdad and to build some degree of trust, not just with Maliki’s ministries but also with the Shiite militias. And because they trusted her, they trusted Odierno.

Over the coming months, wherever Odierno went, whether to explain the mission to American troops, to meet with Iraqi officials, or to explore deals with militants, he brought Sky along. The pair made an odd sight, the general towering over his aide by more than a foot. Odierno was once asked by a fellow general officer what he got from her. He replied, “She helps me with the why.”

•  •  •

The month before Odierno came back to Iraq, the seventeen thousand soldiers of the 1st Cavalry Division returned as well, for their second tour of duty, as the main US force in Baghdad. Since securing Baghdad was Odierno’s top priority, his staff worked in tandem with the 1st Cav’s staff, especially with its planning chief, a lieutenant colonel named Doug Ollivant.

Ollivant came out of the West Point Social Science Department, having taught there from 1999 to 2002, overlapping with John Nagl, Ike Wilson, and the others who thought and talked about the future of the Army in the age of the “moot-wah wars” in Bosnia, Somalia, and other third-world hot spots. Afterward, he studied at the School of Advanced Military Studies, in Leavenworth, during which time he missed the invasion of Iraq, but he was deployed during the first troop-rotation as operations chief for one of the battalions in the 5th Cavalry Regiment.

In the early fall of 2004, as the resistance was intensifying, Ollivant’s battalion was rounding up insurgents at an impressive clip, owing to a trick that he’d figured out along with the unit’s assistant intelligence officer, a University of Chicago graduate named Eric Chewning, who had quit his Wall Street job to enlist in an Army officers’ program after some of his friends were killed in the September 11 attacks. On his own initiative, Chewning was making pals with Shia and Sunni militiamen, some of whom had joined the armed wing of the Supreme Council for the Islamic Revolution in Iraq. As such, they were bad guys in the eyes of the US-led coalition, but there were varying degrees of bad: some were shooting at American soldiers, some weren’t for the moment. In any case, between one group and another, Chewning’s sources knew who the shooters were and where they were hiding. So Chewning would put together the intel, and Ollivant would form the raiding party. At their peak, they were running three-day targeting cycles: making an arrest, getting new intel, making another arrest—all in a three-day period. After a while, Chewning’s cell phone contained the numbers for dozens of friendly bad guys from a wide array of sects and militias, and those of several ordinary Iraqis as well. When gunfire sounded, he could make a few calls and learn right away who was shooting.

Neither he nor Ollivant knew much about counterinsurgency. Chewning was essentially doing the same thing that he’d learned as a mergers and acquisitions broker at Merrill Lynch—he was networking. The difference was that in Baghdad, the local power brokers in his network were Shia militiamen.

Ollivant’s tour ended in February 2005. Back in the States, he talked with some of his Army friends about what he’d been doing. He was eager to put the experience in perspective, to see if it held any lessons for the future. John Nagl told him that he should get hold of Galula’s Counterinsurgency Warfare. Ollivant found a copy and read it with astonishment. He and Chewning had been following several of Galula’s principles, especially the idea that intelligence was the springboard for operations and that acquiring intelligence required living among the people in order to earn their trust. They’d been doing COIN without knowing it.

That fall, Ollivant learned that David Petraeus, the new commanding general of the Combined Arms Center at Fort Leavenworth, was posting a writing contest: whichever officer wrote the best essay on countering insurgencies would win $1,000 and get the piece published in Military Review.

Ollivant and Chewning had already started jotting down a list of what they’d done, what worked, what didn’t, and how it all fit into Galula’s theories of counterinsurgency, which, they now believed, should guide the Army’s strategy in Iraq more broadly. When Ollivant heard about the contest, the pair sharpened their prose, threw in a lot more references to “counterinsurgency,” and submitted it under the title “Producing Victory: Rethinking Conventional Forces in COIN Operations.”

Their essay won.

The contest’s judges included Ollivant’s old colleague from West Point John Nagl, as well as Conrad Crane, Kalev Sepp, and Jan Horvath—who, at the time, were putting together the first full draft of the COIN field manual. Petraeus, who’d handpicked the judges, liked the essay, too. The contest was but one piece of Petraeus’s War of Information campaign to spread COIN throughout the officer corps, and the Ollivant-Chewning piece most closely matched the message that he was promoting. It was telling that the runner-up prize of $500 went to a major named Paul Stanton, who had written an essay about Petraeus’s stability operations in Mosul; though the official announcement of the contest’s winners didn’t say so, Stanton had participated in those operations as one of Petraeus’s company commanders.

A half year later, in early January 2007, just a few days after being named the next commander of US forces in Iraq, Petraeus learned that Ollivant was working as the planning chief of the 1st Cavalry Division, known in Army circles as First Team. He sent Ollivant an email: “Doug, if this reaches you, pls let me know if you still believe in the thesis in your article, ‘Producing Victory.’ If so, can you execute it with First Team? Thx.”

Ollivant wrote back, assuring Petraeus that he still did believe in what he’d written and that he was executing it at that moment.

Through their varying backgrounds and experiences, Ollivant’s team and Odierno’s staff were coming to similar conclusions about how to secure Baghdad. Ever since Odierno’s return, members of the two staffs had been getting together two or three times a week. On Christmas Eve, Odierno met in a conference room with the 1st Cav’s commander, Major General Joseph Fil, and their respective staffs, to lay down the guidance for a common campaign plan.

They all agreed on the basic challenge: how to break the cycle of sectarian violence, especially in Baghdad, while helping to bolster the legitimacy of the Iraqi government. Odierno laid out the main points of his “Closing the Gap” memo, which his staff had by now refined to a single PowerPoint slide: protect the population, live where the people live, build trust, use that trust to gather intelligence on extremists, and go after them relentlessly while also reaching out to reconcilable elements.

For Odierno, these ideas stemmed from everything he’d learned in the previous three years—from his time in the State Department with Eliot Cohen and Phil Zelikow, his command training sessions with David Petraeus, his reading of the COIN field manual (and every other book on the subject he could lay his hands on), the brainstorming with fellow officers, and, most recently, his trip to Ramadi, where for the first time he actually saw how this sort of operation worked. For Ollivant, the ideas resonated with his earlier time in country and the lessons he and Eric Chewning had drawn in their contest-winning essay.

The biggest decision that came out of both headquarters, Odierno’s and the 1st Cav’s, was to move the troops out of the superbases on the outskirts of town and install them in “joint security stations,” which would rapidly be built—eventually, seventy-seven of them—in Baghdad’s neighborhoods. As with McMaster’s combat outposts in Tal Afar and MacFarland’s in Ramadi, the troops would eat, sleep, and patrol in and around these stations twenty-four hours a day, seven days a week, to keep the insurgents from coming back after they’d been pushed out. This was the clear-hold-build strategy put in motion, only now Odierno called it “clear-control-retain”—in part to play down “build” (which seemed overly ambitious, given the depths of the crisis) and in part to emphasize “retain” (the idea being that he would not hand over an area to Iraqi troops until they were ready to control it by themselves).

Odierno and the 1st Cav did differ on one aspect of the new campaign plan. Ollivant assumed, in drawing up his part of it, that all five surge brigades would be deployed in Baghdad; that was what President Bush had said when he announced the surge on January 10.

In the meantime, though, new intelligence had come to light that pushed Odierno and his staff in a different direction. During a recent sweep of a jihadist hideout, American Special Ops forces had come across a hand-drawn map that appeared to illustrate al Qaeda’s plan for an assault on Baghdad. There were markings near various cities and villages encircling the capital—the Baghdad Belts, they came to be called—as well as lines drawn along roadways from the belts into Baghdad, like the spokes on a wheel, suggesting that those would be the assault routes.

The map only added credence to a massive analysis of insurgent activity along the belts that had already been produced by a team of three Army majors on Odierno’s intelligence staff: Nichoel Brooks, Ketti Davison, and Monica Miller—all three, strikingly, women in their late thirties, who were known collectively to the officers around them (for the most part, good-naturedly) as “General O’s Girls,” “Odierno’s Chicks,” and (the nickname the women themselves liked best) “the Coven.”

None of the three were West Pointers; they were, in fact, all autodidacts. Brooks, tall and athletic, had gone to the University of Nevada on a basketball scholarship before earning a master’s degree in information systems at the University of Texas and joining the Army through an officer candidates’ school program with a specialty in tactical intelligence. She’d first deployed to Iraq at the end of 2003 and acquired deep expertise in Shia militias, tribes, and families, tracking their movements, alliances, gains, losses, and shifts. When Odierno returned to Baghdad, she was well into her tenure as chief of the command’s analytic center, fusing and analyzing data from more than three hundred intel officers, about fifty of them embedded with US combat brigades.

Miller had taken ROTC to fulfill her phys ed requirement at Texas A&I University, liked it, won her officer’s commission, fell into the intelligence ranks, impressed her commanders, and wound up in Iraq, immersing herself in the data about Saddam’s former entourage and Sunni militias, in tandem with Brooks’s work on the Shia.

Davison was the one who took the thousands of scattered data points and put them all together in a series of interactive graphs, charts, and tables. As a junior officer at Fort Leavenworth’s School of Advanced Military Studies, she’d taken a course on systems design and written her thesis on applying its principles to intelligence analysis. Her work in Iraq put the thesis into practice.

With Brooks’s and Miller’s analysis plugged into Davison’s models, the team could brief Odierno on the daily shifts in Iraqi sectarian politics: who was doing what where, to what effect, and with what consequences for American strategy, tactics, and policy.

To the Coven, the al Qaeda map was nothing new. They’d pieced together a much more detailed map, from their own analyses, which pinpointed the bomb-making sites of the various militias as well as the supply routes that they followed into Baghdad. The Sunnis’ sites were to the north, along a trail that began in Fallujah. The Shiites’ were to the south, centered in the cities of Karbala’ and Najaf. The road to Baghdad from al Kut was a favored route for criminal smugglers. The most violent areas along the Baghdad Belts—in Baqubah to the northeast and Yusifiyah to the southwest—were “friction points,” where the various support zones and supply routes overlapped and the various factions fought one another for dominance.

This discovery wasn’t merely interesting: it uncovered a major flaw in the impending plan for President Bush’s troop surge. Putting all five of the extra Army brigades in Baghdad wouldn’t solve the problem, because the bombs were being built—and the militias inside Baghdad were being supplied—by extremist leaders in the belts outside the capital. At least some of the extra brigades had to attack the belts and interdict the supply routes.

The Coven’s intelligence analysis provided daily updates on precisely where the bomb sites were along the belts—and precisely who was controlling the routes. In the end, Odierno decided to split the five brigades between Baghdad and the belts. (The two Marine battalions were sent, as had always been the plan, to shore up Anbar Province, which was already well into the hold-and-build phase of clear-hold-build.) And he used the Coven’s analysis to pinpoint exactly where to send the brigades along the belts.

•  •  •

On February 10, 2007, General David Petraeus assumed command of US and coalition forces in Iraq. The same day, Odierno briefed him on the campaign plan. Petraeus approved it—no surprise, since its overarching ideas had come from Petraeus, directly or indirectly, and the operational details that hadn’t, including the innovative intelligence on the belts, struck him as sensible.

Odierno would remain in charge of day-to-day military operations. Petraeus would fit those operations into a broader strategic and political framework, a Joint Campaign Plan, which he coordinated with the US ambassador, Ryan Crocker, to ensure that every aspect of politico-military policy—strategy, tactics, intelligence gathering, and all the pieces of nation-building too—flowed in alignment with the overall plan. There had long been a brigade here, a battalion there, whose commanders understood what counterinsurgency was about and who tried to apply its principles to the narrow areas they covered. Petraeus would now give them “top cover,” legitimizing what they’d been doing and actively making sure that other colonels and captains who hadn’t been doing COIN—whether because they didn’t get it, resisted it, or had just been following contrary orders—now joined the program.

Finally, Petraeus would be the point man for visiting legislators, journalists, and other opinion-makers—a role that he’d always performed shrewdly. This was classic information operations, which he saw as an integral part of war, especially counterinsurgency war. The “center of gravity” in such a war was the people—not just the people in the country where the war was being fought but also the people on the home front, whose children and spouses were fighting the war, whose tax dollars were funding it, and who needed some hope, some reason, for remaining patient a while longer.

The war was unpopular back home, and with good reason; the chances of a turnaround seemed dim. Yet Petraeus was widely seen as a different kind of general. He was still basking in the plaudits for his campaign in Mosul; his COIN field manual was touted as the work of a creative strategist; if he could replicate what he’d done in Mosul, if he could turn the field manual’s ideas into reality, maybe the war wasn’t such a losing proposition; maybe he could turn it around. That was why Bush ordered the surge and chose Petraeus to install a new strategy: this was the last chance for redemption; Petraeus was typecast as the redeemer.

Petraeus brought along his full entourage, the officers he’d been grooming and encouraging, whose views he’d sought out and found compatible or promising, for several years. There was Mike Meese, the West Point Sosh chairman, serving as his special assistant; H. R. McMaster, called in just as the council of colonels was breaking up (he never would resume that fellowship in London); Pete Mansoor, also from the colonels’ council and, before then, Petraeus’s handpicked director of Leavenworth’s COIN Center; Joel Rayburn, McMaster’s political adviser at Tal Afar, who, for the past year, had been an internal dissident on General Abizaid’s staff, grinding out memo after unread memo (though some outsiders, including Petraeus, had been reading them) on why Casey’s transition strategy wouldn’t work. And there was Dave Kilcullen, the Australian COIN soldier-scholar who’d drafted the irregular-warfare section of the Pentagon’s Quadrennial Defense Review, written the “Twenty-eight Articles” essay that so many junior officers had read on how to perform COIN in their daily patrols, and spoken at so many of the culture-changing COIN conferences since. (John Nagl was not among the chosen; he had recently been named commander of a battalion in Fort Riley, Kansas, that was being reorganized into a dedicated cadre of Army officers who would train other countries’ security forces, a mission that Petraeus saw as too important—possibly for the Iraq War, certainly for future insurgency wars—to interrupt.)

As part of his information operations campaign, Petraeus asked McMaster to bring together twenty experts—fellow officers, embassy officials, and prominent academics—to form what he called the Joint Strategic Assessment Team. The stated rationale of the team was to help the commander decide on a new strategy. In fact, though, Petraeus already had a new strategy; the team’s real purpose was to rally buy-in for the strategy from critics and institutions whose support Petraeus knew would be politically useful. In this sense, the JSAT was similar to the workshop that Petraeus had put together at Leavenworth a year earlier to vet the COIN field manual. The conference hadn’t been entirely a ruse; some of its participants did raise provocative questions, point out problems, and make useful suggestions. The same was true of the JSAT, especially since Petraeus hadn’t yet completely thought through the Joint Campaign Plan’s political section. But in both cases, it was Petraeus and his inner circle who did most of the work; the others were selected to widen the circle of support, to impart a sense of ownership to the diplomats who would have to carry out the political part of the plan and to the academics who might write articles about the new strategy in major magazines and newspapers back home.

The JSAT’s members met in one of Saddam’s former palaces, now an annex of the American embassy, with marble floors and thirty-foot-high ceilings. They met every day, sometimes twelve to fifteen hours a day, for three weeks, usually splitting off into plenary groups to hammer out the details of specific issues.

One issue that the JSAT fiercely debated was how far to extend what Sean MacFarland had done in Anbar. Turning and reconciling with former insurgents was a big part of counterinsurgency; Petraeus had written much about the subject in the field manual; Odierno had called for it, at least selectively, in his campaign plan. But were there limits? Should the United States try to co-opt all the Sunni insurgents into an alliance of convenience against jihadists? Or were some of them so militant, had they killed so many American soldiers, that the very idea of allying with them went beyond the pale?

Dave Kilcullen was on the side of pushing deals as far as they could go. This was a sectarian civil war, he would say. Intervening powers in a civil war have two choices: they can pick one side and annihilate the other, something that the United States wouldn’t and shouldn’t do; or they can broker a deal and set up a power-sharing arrangement.

Stephen Biddle, one of the few civilians in the group, agreed. Biddle had been among the handful of critics, along with Jack Keane and Eliot Cohen, to discuss the war with President Bush in the Oval Office the previous December. Biddle was more pessimistic about the course of the war than most of the JSAT members, which was why he favored making deals, even ugly deals. To his mind, the whole campaign should be built around devising incentives for bad guys to stop fighting and start talking.

But several of the military officers in the room resisted the idea, and H. R. McMaster was the most vehement. McMaster was all for reconciling with reconcilables; he’d shown MacFarland how to do it in Tal Afar. But some of the guys that Kilcullen and Biddle wanted to work with were brutal gangsters, monsters. He’d seen what they’d done, not just to Americans but to innocent Iraqis, to women and children. It didn’t take much for McMaster to get emotional; he started screaming, at its peak with an intimidating tone, which spurred others to scream just as loud.

Biddle, a soft-spoken academic, was alarmed by the ferocity of the argument. At one point, he looked around and realized, with a slight shiver, that he was the only one in the room who wasn’t carrying a gun.

Out of this discussion, though, came a consensus of support for one of the decisive turns in the war: the Sons of Iraq. This was the moniker that Petraeus gave to Sunni Arabs, many of them armed militiamen, who agreed to stop shooting at coalition soldiers and to start cooperating with them instead—patrolling neighborhoods, unearthing hidden caches of weapons, and joining in the fight against al Qaeda, sometimes even against other Sunni militiamen who had chosen not to change sides. In exchange for flipping, they would earn a regular salary; and unlike MacFarland’s Anbar Awakening, which paid co-opted Sunnis out of the Iraqi government’s budget, Petraeus paid the Sons of Iraq from the American military’s cash fund.

Petraeus had tapped a similar fund to great effect in Mosul. But there, the money had been spent on small local projects: sweeping streets, cleaning sewers, building schools. Buying off insurgents who had been killing American soldiers was different, and Petraeus knew it.

Looking for legal cover, he asked the command’s lawyers if he could use the funds to pay the Sons of Iraq under the rubric of “site security.” The lawyers squirmed but concluded that it would probably be all right. So he did it. And because the funds could be disbursed at the commander’s discretion, Petraeus felt no obligation to tell the president, the Congress, or anyone else in Washington exactly how he was spending the money. So he didn’t.

But before many Sunnis could be persuaded to give up the resistance and join forces with foreign infidels, some groundwork had to be laid. Not least, Petraeus had to crush the main Shiite militias; they were the ones whose death squads were killing Sunnis, compelling many survivors to support or join Sunni militias and even to turn to al Qaeda for protection against ethnic cleansing.

This was the “cycle of sectarian violence” that Petraeus and Odierno were determined to break: a “spectacular attack” by al Qaeda or another jihadist group, followed by retaliation from Shia militants, followed by counterretaliation from Sunnis, then counter-counterretaliation from Shiites, and on and on the killing spiraled.

When Pete Chiarelli had been corps commander, he understood this cycle as well. To help break the cycle, he planned targeted raids on Shia militias, especially soldiers in Muqtada al-Sadr’s Mahdi Army in Sadr City. But Prime Minister Maliki would always warn Casey to stay away from Sadr, and Casey would order Chiarelli to comply. Toward the end of his tenure, Chiarelli opposed the surge, figuring that there was no point in sending more troops if the Iraqi government blocked them from doing what they needed to do.

Petraeus got around this obstacle by simply ignoring it. He sent his troops into Sadr City without telling Maliki in advance. It was here that the Joint Strategic Assessment Team served its most useful purpose. In their briefings to the JSAT, H. R. McMaster and Joel Rayburn emphasized what they’d learned from their tenures in Tal Afar and Central Command headquarters: that the Iraqi government was a party to the civil war; that many of their police were, in effect, Shia death squads, whose killing sprees against Sunnis were ratcheting up the sectarian violence.

Rayburn had read the “end-states” briefing that Celeste Ward had written for Chiarelli, with its conclusion that the coalition and the Iraqi government possessed divergent interests and pursued divergent goals. Ward and Chiarelli had inferred from this that the war was all but hopeless; by contrast, Rayburn inferred that the United States should use its leverage to get Maliki to change course.

The embassy officials on the JSAT agreed. When Maliki called Ambassador Crocker to complain about Petraeus’s insolence, he received no sympathy. It was made clear that the days of blank checks were over; that, as long as Maliki and his ministries were pushing sectarian violence, the United States would push back.

Petraeus and his aides stepped up the pressure on all the parties. When they discovered that a national police brigade in southern Baghdad was murdering Sunnis, they told the interior ministry that the coalition was withdrawing all support for that brigade until it stopped the killing. Suddenly it had no fuel, spare parts, money, or equipment; nor were its men allowed into any area controlled by American troops. The brigade backed off within days. Some (though far from all) of the other police squads, watching the standoff play out, followed suit.

Maliki tried to regain the upper hand, giving Crocker a list of fourteen demands. Most of them called on him to resume policies that Casey had pursued but that Petraeus had ended. Among them: withdraw all US troops from Iraqi cities, release all Shia detainees, and accelerate the transfer of authority to the government. Maliki had been pushing for these moves all along, so that he could intensify the attacks on Sunnis without American interference.

Petraeus called for a meeting with Maliki’s national-security adviser, Mowaffak al-Rubaie, and told him that none of the demands would be met. If the prime minister wanted to phone President Bush and complain, Petraeus himself would take the next plane back to Washington, go directly to the Oval Office, and explain the situation.

At other times, Petraeus would cozy up to Maliki, play good cop to his own bad cop, pump the prime minister with pep talks about his destiny as a great national leader, and brief him—share sensitive intelligence with him—on the many ways in which Sadr was an unreliable ally, serving interests, including Iranian interests, that were hostile to Iraq’s long-term security.

Maliki backed off, at least for the moment.

•  •  •

Petraeus’s biggest gamble was his literal application of Galula’s dictum to live among the people. Very soon after taking command, he ordered joint security stations to be constructed all over Baghdad—seventy-seven of them, in all—starting in Ghazaliya, the once-prosperous neighborhood that Meghan O’Sullivan, his White House contact, had personally witnessed plunging into chaos.

Some of these stations were modifications of outposts that Casey had built, although back then, the troops had used them as day stations and would return to their large bases on the city’s outskirts at dusk. Most of the new ones had to be constructed from scratch, in one night; otherwise they would come under fire the next morning. Crews would roll in after sundown with a flatbed and crane, sometimes two cranes. By sunrise, the station would be fully erected, manned, and guarded, as if it had been there for months.

Petraeus also ordered the construction of concrete walls, winding all through Baghdad, as a brusque way of separating Sunni from Shia neighborhoods, thus keeping the peace between them, at least for a while.

The idea was that the constant presence of American troops would reduce the violence and that this, in turn, might make the local people feel sufficiently secure to help the troops find insurgents in their midst; this would reduce violence further, and on the cycle would spiral, resulting eventually in support for the war effort—and ultimately for the Iraqi government.

Several officers got the concept right away. They’d understood it for some time, having taken the course at the COIN Academy and read several articles, the online forums, and the recently issued COIN field manual. Now, with top cover from Petraeus, they pushed further, taking still greater risks at making contact and striking deals on their own initiative with local citizens, fence-sitters, and, in a few cases, militants.

But some officers resisted; the whole notion rubbed against everything they’d been taught about respecting a host-government’s sovereignty and the Army’s traditional primary mission of self-protection. Others were eager to follow the new commander’s orders but didn’t know what specifically to do.

So Petraeus asked Dave Kilcullen to write another essay, similar to his “Twenty-eight Articles,” which had circulated widely a year earlier. This one, Petraeus told him, should translate the principles of Odierno’s campaign plan into point-by-point instructions.

Kilcullen swiftly wrote a three-page memo called “Counterinsurgency Guidance.” It began with a précis of what Petraeus had said repeatedly in public: that the Iraqis will have to work out their own political solutions, but they can’t be expected to do so in a climate of “chaos, violence, and fear.” They needed help to change this climate, to create some breathing space—a zone of security—in which they might resolve their disputes calmly. “This,” Kilcullen wrote, addressing the troops, “is where you come in . . . Your actions at the tactical level resonate throughout the communities you protect and are amplified beyond.”

He listed ten things they needed to do, chief among them: “Secure the people where they sleep” (because securing them only in the daytime will make them more vulnerable to revenge by insurgents at night). Another point: “Get out and walk.” Armored vehicles were necessary to get from one place to another, but once you’re there, mingle among the people; the vehicles offer self-protection but “at the cost of a great deal of effectiveness”; patrolling by foot is the only way to build trust, and building trust is essential to gathering reliable intelligence.

With Kilcullen in the room, Petraeus read his memo, picked up a pen, and seemed about to sign it. Then he stopped, thought for a moment, and said, “Take this over to Emma and tell her that Ray should sign it.”

It was another instance of Petraeus seeking buy-in from those he needed to carry out a policy. Kilcullen took the document to Odierno’s office. Emma Sky went over it, making a few small changes. Odierno later signed it, sent it to all combat units, and also came up with the idea of having it translated into Arabic and circulated through the local media, so that the Iraqi people would have a sense of American intentions.

Petraeus and Odierno understood that the guidance was risky, that it would expose their troops to greater danger. Petraeus publicly acknowledged that, as a result of his new policy, more American soldiers would probably be killed, at least in the short run. They would be spending more time out in the open, near the enemy, in the fight. And because of the surge, there would be more of those troops taking risks. If the strategy worked, though, the payoff would be a steep reduction in military and civilian casualties over the long haul—and, further out into the future, a more secure and stable Iraq.

But this was a gamble. It was sure to work in theory; no one knew what would happen in practice.

•  •  •

For the first half of 2007, the violence didn’t let up in the slightest. Roadside bomb explosions hovered between 400 and 500 a month; suicide bombers set off another 200 or so blasts a month. And because the troops were now out in the open almost constantly, American combat fatalities soared by 50 percent, reaching a peak of 120 in the month of May.

But civilian casualties were dropping: from a high of 3,700 in January to 2,500 in March and 2,000 in June. Still, the death toll was too high, for Iraqis and for Americans. Back home, the surge was widely declared a failure. Even the commanders were getting a bit nervous.

Then on June 15, all five surge brigades were finally in place. (They had been mobilized at the rate of roughly one per month.) The next day, Petraeus and Odierno launched Operation Phantom Thunder, the long-planned, multipronged assault on the Baghdad Belts. Five days after that, a joint US-Iraqi Special Operations team mounted a raid on a key Mahdi Army stronghold in Sadr City.

The effects were almost immediate. The jihadists’ bomb-making operations were disrupted; their supply routes—both to Baghdad and to one another—were cut off; their veneer of invincibility was shattered. Meanwhile, the attacks on Sadr City persuaded many Sunnis that maybe the Americans weren’t taking the side of Shia militants after all. In August, Muqtada al-Sadr himself declared a cease-fire and took refuge in Iran. The same month, Odierno mopped up the al Qaeda cells that had eluded the first round of attacks and began putting into motion the hold-and-build (or, as he now called it, secure-and-sustain) phase of the COIN campaign.

The next few months saw a steep decline in all measures of violence. By August, explosions of roadside bombs dropped to three hundred a month, car bombings to fewer than one hundred; by the end of the year, the figures plummeted, respectively, to under one hundred and barely twenty. Sectarian deaths fell to fewer than five hundred in August and, by the end of the year, to nearly zero. American combat casualties began to drop as well, reaching numbers—ten to twenty a month—that hadn’t been seen since the start of the occupation.

As the violence subsided and the most militant insurgents lost their strength and appeal, the Sons of Iraq saw their ranks swell, which reduced the violence further. The recruitment lines were so long, the desire to sign up so feverish, that even some of the program’s supporters were a bit concerned—it wasn’t clear who some of these new comrades were or how enduring their allegiances might be.

Still, American soldiers and marines knew something new was in the air when they noticed Shiite and Sunni militiamen consorting openly in the streets of the capital. In the short run, this all worked to the Americans’ advantage.

The long run would be up to the Iraqis.

And that was the problem.

In the euphoria over the undeniable improvements in Iraq, the emergence of near-normal everyday life, a tendency arose both in the field and back in Washington to give all credit to the American military, to the surge, or to General Petraeus and his counterinsurgency strategy.

But there were other factors having little to do with anything that the Americans had done—apart from starting the war and thus triggering the social disruptions in the first place.

First, Petraeus and the surge had come late to the civil war. Many areas of violence had already been “cleared” through ethnic cleansing and exile; hundreds of thousands of Sunnis had abandoned their homes under threat of death for points west, either in Anbar Province or across the border into Lebanon or Syria.

Second, the Sunnis would not have been so eager to split from al Qaeda, much less join the Sons of Iraq, had they not begun to realize that they were losing the civil war with the Shiites. Only then did they see an alliance with the Americans as their best hope for survival.

Similarly, Maliki consented to the American assault on Shia militias in part because he had no choice, in part because he was slowly realizing that the Mahdi Army was at least as much a threat as an ally, that Sadr had joined his political coalition in order to take it over. And so Maliki allied with the Americans for the same reason the Sunni tribesmen did: to boost the prospects of his survival.

The surge facilitated this cooperation. It conveyed to everyone concerned that the Americans weren’t leaving after all. They were staying, in strengthened numbers; they had to be contended with, and perhaps they could be trusted.

True, the Anbar Awakening had preceded the surge, preceded the arrival of Petraeus, was well under way before publication of the COIN field manual, and was initiated by Sunnis, not by Americans. But it took a commander like Sean MacFarland to grasp the Awakening’s potential and respond sagely to the Sunnis’ initiative. That is, it took someone who’d been trained in COIN and influenced by the same events and ideas that had sired the field manual, not least Petraeus’s own experiences in Mosul. More conventional commanders would have responded to the violence by killing as many Sunnis as they could—with disastrous results on all sides.

Even with MacFarland in place, the Awakening would have remained a mere local phenomenon, had it not been for the changes that followed. The surge enabled Petraeus to spread variations of the Awakening across the rest of Iraq; it gave him enough troops to clear and hold more than one area at a time.

Finally, it took a commander with Petraeus’s background, instincts, and brazen assertiveness to know what to do with the surge—how to make twenty-one thousand extra combat troops (not exactly a massive escalation) the hinge of a strategic shift.

Even so, in the broader scheme, this shift, however dramatic, had only tactical consequences. The surge and the COIN campaign that went with it were—explicitly—mere means to an end. Petraeus said many times, as did Kilcullen in his Odierno-signed counterinsurgency guidance, that the point of it all was to give Iraq’s factions some “breathing space” so that they could focus in peace on hammering out their differences and forming a cohesive government with widespread legitimacy.

By the summer of 2007, it was clear that the American troops were fulfilling their part of the bargain. The question was whether the Iraqi factions would take advantage of the breathing space. Would they strike an agreement on sharing oil wealth? Would the Kurds and Sunni Arabs settle their property disputes in Kirkuk? Perhaps most important for the country’s future, would Maliki incorporate the Sons of Iraq into the national Iraqi army?

And if these things didn’t happen, would the surge and the new strategy turn out to have only prolonged the fighting and compounded the war’s tragic waste?