They Told The Pentagon The SAS Had Failed. The Pentagon Pulled Up The Satellite Feed And Went Quiet.
He set the red folder on the table. He did not sit down. He spoke for 4 minutes. The last sentence was the one that mattered. The SAS had failed. Three of the officers present nodded. Two wrote something down. Nobody asked a question. The briefing ended at 0904. The folder remained on the table as the room cleared.
A clerk moved it 3 ft to the left at 0912. Nobody opened it. The folder would still be there the next afternoon. By then, the Pentagon would have pulled up the satellite feed and the conversation in the screening room would have gone quiet. That is where this story ends. It is not where it begins. It begins in a fuel depot outside Baqubah 14 months earlier on a road that carried most of the traffic into western Diyala.
It begins with a man named Abu Khalid al-Janabi, who had survived four American-led operations, two Iraqi-led operations, and by the time Elkins first read his name on a target list in August 2005, had been personally responsible for the coordination of at least 280 deaths. He had survived for three reasons.
He kept his phone off. He never slept in the same building twice, and he had never, as far as the joint task force could determine, been seen by anyone whose description of him could be matched to a photograph. That last fact was the one that defeated everyone who went looking for him. It was also the one the SAS patrol did not find a way around.
They found a way under it. The SAS element running surveillance on Abu Khalid was a patrol of six men. They were attached to the Joint Special Operations Task Force, but reported separately. They held the file number 2217. They held no name in press materials because there were no press materials. They were, in the filing system of the coalition, something closer to a placeholder.
For 11 days in October 2006, the patrol did not transmit at all. That was the silence Elkins had read as failure. It was not failure. Abu Khalid al-Janabi was born in 1968 in a village south of Samarra. He joined the Iraqi army in 1986, served in an engineering unit during the war with Iran, and was released in 1992 with a technical specialty that would later define the rest of his life.
He knew how to build things that were meant to come apart. By late 2004, he was running a network of 40 to 60 active bomb makers across the Baghdad belt. The coalition had assigned him a file number. They had assigned him a photograph, though the photograph was 12 years old and showed him in a uniform. They had assigned him a ranking, third tier senior facilitator, and a budget for his capture, though the budget figure was never published.
What they had not assigned him was a current face. That, at the table at the Pentagon, was the persistent problem. The first serious attempt to close the file came in March 2005. A signals intercept placed Abu Khalid in a safe house in Yusufiyah. An airstrike went in at 0240 local time. The building came down inside 9 seconds.
When the bodies were identified 3 days later, two four of them were his cousins. He was not among them. He had left the building 11 hours earlier in the back of a fuel truck through a checkpoint that did not record the traffic. The second attempt came in July. A drone observed a convoy leaving a known associate’s compound near Samarra.
Two missiles, three vehicles destroyed, 11 men killed. Abu Khalid had been supposed to travel in the lead vehicle. The lead vehicle had been sent 20 minutes earlier empty as a test. In October, a night raid went into a compound in Fallujah that had been identified through three separate source streams. 140 personnel, six helicopters, two hours of search.
The compound was empty. It had been empty, the interrogations of local residents later established, for at least 11 days. The fourth attempt was the one that cost the most in political capital. A walk-in source at a coalition gate in Baqubah claimed to have been Abu Khalid’s driver for 6 months. He gave an address. He gave a schedule.
He gave a description of a specific white sedan and a specific Wednesday. In January 2006, three intelligence services coordinated a capture operation around that sedan. The sedan arrived on schedule. It contained two men, neither of whom was Abu Khalid. The source had fabricated the entire account in exchange for a cousin’s release from detention.
He did not know Abu Khalid. He had seen him once in 2003 from 40 m away. The fifth attempt leaned on metadata. In April, a burst of phone activity around a village north of Baqubah produced a pattern that three separate analysts called conclusive. The pattern was conclusive. The phone was not Abu Khalid’s.
It had been given 3 weeks earlier to a courier whose route the analysts mistook for the facilitator’s routine. When the compound was hit, the courier was inside. He was 19 years old. The sixth attempt came in June 2006. 800 personnel drawn from the Joint Task Force and two conventional American brigades conducted a coordinated sweep of 11 compounds across three districts over 4 days.
The sweep produced 23 detainees. 19 of them had no operational connection to Abu Khalid. Four of them had known him briefly years earlier. None of them knew where he was. By the end of June 2006, the target file had been opened for 26 months. The coalition had expended resources that, in any bureaucratic accounting, ran into the hundreds of millions of dollars.
Abu Khalid had been photographed once in that period, from a distance, in profile, in poor light. The photograph was useless for biometric matching. What the institution had not tried, and what it did not know how to value, was sitting a long way down the food chain. In August 2006, the SAS patrol submitted a proposal.
The proposal was two pages long. It described a 6-week surveillance window around a stretch of road between Baqubah and a village called al-Zaydan. It requested no additional personnel. It requested no air support. It requested a single concession, to be removed from the coordination grid for the duration. The last request was the one that killed it in review.
To be off the coordination grid meant that the patrol would not check in on schedule, would not appear in the daily briefing slides, and would not be visible to the officers managing the theater. That was, from the institutional point of view, unworkable. A patrol off the grid was a patrol that could not be directed.
A patrol that could not be directed was a patrol the commander could not be said to be commanding. The proposal went up the chain. It reached Elkins’s desk on the morning of August 23rd, 2006. He read it once. He wrote three words in the margin. He sent it back. The three words were not a chance. The briefing where Elkins spoke about the proposal out loud was 4 days later.
It was held in a secure room at Camp Victory, not at the Pentagon. 14 officers were present. The list included two British liaisons, one Australian, three American colonels, four majors, two CIA officers, and two from a coordinating body that has not been named in any public record of the meeting. Elkins ran the briefing.
The British liaison who had carried the SAS proposal up the chain was seated at the far end of the table. His name was Lieutenant Colonel Harold Pike. He had been in theater for 11 months. He was 44 years old and had commanded a Pathfinder platoon in Sierra Leone. He did not speak during the briefing. Elkins had the proposal in front of him.
He held it up. He set it down. He tapped it twice with the index finger of his right hand. “This,” he said, “is what happens when you ask Boy Scouts to solve an engineering problem.” Three of the 14 officers smiled. Two looked at the table. Pike did not look up. Elkins continued. “The SAS,” he said, “had good discipline and excellent individual skills.
He would concede that on the record. But the target they were proposing to surveil was protected by a network of 40 to 60 active participants spread across an area the size of Connecticut. A six-man patrol was not going to penetrate that network. It was going to sit in the desert, miss its windows, blow its own cover, and require an extraction that the task force would have to mount with conventional assets.
” He said the proposal was the kind of thing officers wrote when they wanted a story to tell at regimental dinners. He said it was not the kind of thing a task force commander approved. Then he said the line. “Send the amateurs in first,” he said. “Then we’ll clean up after them.” He laughed when he said it. Two of the American colonels laughed with him.
The Australian liaison, who had worked with SAS elements in East Timor 7 years earlier, did not laugh. Pike, at the end of the table, closed the file in front of him and did not open it again. The line entered the meeting log as request denied, not operationally viable. That was the end of the paper trail. It was not the end of the proposal.
What Pike did the next morning was not in the meeting log, either. He met privately with the SAS patrol commander in a prefab office at Camp Victory. The prefab had a window unit that did not work. The temperature inside the room that morning has been estimated later at 38° C. The meeting lasted 22 minutes. The patrol commander will be referred to from this point forward as Barnard.
That is not his full name, and it is also not his call sign. But it is what Pike called him in the private account Pike later gave to an Army historian in 2014. By that account, Barnard was 34 years old, had served in the regiment for 9 years, and had spent 14 months in Iraq across two separate rotations. He was, in Pike’s description, a quiet man who did not volunteer more than he needed to.
In the prefab, Pike told him the proposal had been denied. Barnard did not react visibly. Pike told him what Elkins had said in the room. Barnard nodded once. Pike told him the patrol would be folded back into the general tasking cycle within the week. Barnard said six words. “We have a cousin in Al Zaidan.
” Pike did not ask what that meant. The patrol was formally back on the grid on September 2nd. The daily slides at the task force briefing began including a line for patrol 2217. Routine sustainment, area familiarization, liaison with Iraqi army battalion in Baqubah. Nothing that required a response from the task force.
Nothing that required air or logistics. That was the official record for 7 weeks. What the patrol was actually doing in those 7 weeks would not appear on any slide. It would not be in any briefing. It would not, at any point, be visible to Elkins. It was, in the most literal sense, off the grid that counted.
The cousin in Al Zaidan was a man named Saad. He was 41 years old and worked as a fruit seller at a market near the intersection of two roads, one of which ran north toward Baqubah, and the other of which ran east toward a village Abu Khalid was known to have used as a staging point during the Anbar offensives of 2004.
Saad had been a childhood friend of one of Abu Khalid’s brothers. He had last seen Abu Khalid in person in 2002 at a family wedding. He did not like him. Saad was not an asset. He was not registered anywhere. He did not receive money for what he provided because he did not provide anything until he was asked, and he did not volunteer anything that was not asked for.
He had, for 3 years, been quietly watching the traffic on the road that ran past his fruit stand. He had noticed, over that time, that certain vehicles appeared on certain days and did not appear on others. He had noticed the pattern without cataloging it. He had never been asked about it. The SAS patrol began talking to Saad in the third week of September 2006.
They talked to him four times. Each conversation lasted less than 20 minutes. Each conversation took place at his stand during normal business hours under the cover of a vehicle purchase of fruit. The four conversations produced a document of 11 pages, which was held by Barnard in a leather-bound notebook whose spine had broken the previous year, and which he now kept closed with a single rubber band.
The 11 pages described a white Toyota Hilux, its license plate, its windows, its tire pattern, and three dates on which it had passed Saad’s stand in the previous 4 months. The pages also described two drivers, one of whom appeared to be the same man each time, and one of whom did not. The 11 pages did not contain Abu Khalid’s name.
That omission was deliberate. Worth noting that at this point in the operation, six different task force elements were still looking for Abu Khalid in six different districts, none of them Al Zaidan. The assessment on the main task force boards placed him, as of early October, somewhere in the Salah al-Din Governorate, which was 80 km north of where the SAS patrol was now sitting.
The assessment was based on two things, a phone intercept from September 24th, and an analysis by a CIA cell in Baghdad that treated the intercept as authoritative. The intercept was real. The phone was not Abu Khalid’s. It had last been his 7 months earlier. The CIA analyst who assessed it had joined the cell 5 weeks before.
No one at the task force was wrong to follow the intercept. The framework they were using required them to follow it. A phone intercept with a known voice match from 2004 was a lead. A fruit seller who had not been asked a question was not a lead because the fruit seller was not in the framework. There was no box on any of the task force’s slides for what Saad was providing, which was something closer to the texture of a single road on three specific Tuesdays.
That was the structural issue. It was also the reason the patrol had requested in August to be removed from the coordination grid. The grid could not carry the information they were gathering. The grid could carry leads. What Saad had were not leads. They were observations a neighbor had made about traffic on a road over 4 months.
On October 8th, 2006, the patrol stopped transmitting. That was the beginning of the 11 days of silence that Elkins would later describe as failure. What the patrol did in those 11 days is known in broad outline because it was reconstructed later from the patrol’s own post-operation report. The report was 14 pages long.
It was submitted on October 23rd, 2006, and was reviewed by three officers, two of whom, the record shows, read it within 24 hours of submission. The third officer read it on November 17th. By then, the question had been settled. The 11 days began with the patrol repositioning from the Iraqi army base they had been using as a cover posting to a set of three observation points around Al Zaidan.
The positions were selected on the basis of Saad’s 11-page document. None of them was in the village. Two of them were in agricultural outbuildings owned by a family that had, without knowing why, agreed to allow the patrol overnight access in exchange for a payment that was not recorded in any cash account.
The third position was in a drainage culvert approximately 400 m east of the road junction. The patrol stopped transmitting at 02:47 on October 8th because transmission from any of the three positions would have been detectable by anyone watching the local spectrum. Abu Khaled’s network did watch the local spectrum.
They did so with commercial equipment purchased in Kuwait 11 months earlier. The equipment was not sophisticated. It did not need to be. A 5-W coalition radio transmission from a drainage culvert in Al Zaidan was, against the local RF background, conspicuous. So, the patrol went dark. For the first 4 days, nothing happened on the road that matched the file.
On day five, a white Toyota Hilux passed Saad’s stand at 11:42. Saad did nothing. He had been told to do nothing. The patrol observed the vehicle from 700 m. The plate matched. The tire pattern matched. The lead driver matched. The vehicle did not stop. It continued north and disappeared into a compound 3 km past the village.
The patrol observed the compound for 72 hours. On day eight, the compound produced movement. A second Hilux, not the one observed earlier, left the compound at 05:17 on a Tuesday morning. Four men were in it. The vehicle drove 2 km south and parked behind a grain silo. 11 minutes later, a third vehicle, a black sedan, pulled in beside it.
A man got out of the sedan and stood by the Hilux for 4 minutes. He did not speak to the men inside the Hilux. He looked at a phone. He put the phone away. Then he got back in the sedan and left. The patrol had their target. The target was in the sedan. Barnard was 90% sure of the identification from Saad’s description of the build, the gait, and one specific detail about the way the man held his right hand.
A stiffness from an old injury that Saad remembered from the 2002 wedding. Barnard did not move. He did not transmit. He did not request authorization. He did not position the patrol for a take. He let the sedan leave. What he did instead is the reason the operation worked. The thing to understand, which the patrol understood and the grid could not, is that a six-man element attempting to take a target protected by a 40- to 60-man network at 05:28 on a Tuesday on an exposed road has a very specific failure mode.
The failure mode is not that they do not get the target. The failure mode is that they get the target and do not get out. And the second half of the operation ends in the loss of the patrol and the target’s predecessor network reconstituting around the vacancy. Barnard knew this. He had read two after-action reports on operations that had ended exactly that way.
He had been on one of them 2 years earlier as an NCO. He was not interested in the kind of success that ended with a named patrol being lost. So, he let the sedan leave and the patrol observed the compound for another 36 hours. And during those 36 hours, they mapped every pattern the compound produced. The vehicles in.
The vehicles out. The foot traffic. The radio transmissions from within. The light visible through the shutters after dark. The hour the generator switched on. The hour it switched off. On the evening of day 10, the patrol identified the 1 hour in the compound’s routine during which the target was reliably in the interior, alone or near alone, and when the compound’s network of watchers was attenuated by shift change.
That hour was 02:15 to 03:15. The patrol moved at 01:47 on day 11, which was October 19th, 2006. They moved in two teams. Four men on the compound. Two men on the road 300 m north to take the driver if the driver moved toward the compound when the first sound came from inside. They carried suppressed weapons.
They carried one breaching charge. They carried Saad’s description of the interior layout, which Saad had obtained from a second cousin whose wife had cleaned the compound 2 weeks earlier and who had agreed to describe three rooms in exchange for a favor the patrol had never specified. They entered the compound at 02:11.
The breach was one charge on a side door that had been identified as the least watched entry point. The charge detonated at 02:12. The four-man team was inside the target room in 23 seconds. Abu Khaled al-Janabi was in a chair facing a laptop with a cup of tea beside him. The laptop was displaying a map.
The tea was still warm. The patrol fired 11 rounds. Six of them were fired by one operator. The engagement lasted 17 seconds. Three men in the target room were killed. Abu Khaled was taken alive. He was bound and moved out of the compound in 4 minutes. The patrol extracted east across 2 km of open ground to a position where they had pre-staged vehicles at 2300 the night before.
They crossed back into coalition-controlled territory at 04:17. At no point during the operation did they transmit. The first transmission from the patrol was a single short burst at 04:19. It said, “Jackpot. Returning.” That was the first thing the task force had heard from patrol 2217 in 11 days.
It was also the first thing the Pentagon had not been told. Because the cable Elkins had read in Washington had gone out at 04:30 Baghdad time on the 17th of October. It had been sent by a colonel at the task force who had, 36 hours earlier, concluded that patrol 2217’s 11-day silence meant the patrol had been compromised, killed, or both.
The colonel’s assessment had gone up the chain. The chain had passed it to Elkins. Elkins had converted it into a single word in the third line of a cable, “Failed.” The cable was sent 2 hours before the patrol made contact. At 16:14 Washington time on October 18th, a staff officer walked into Elkins’s office with a request from the British liaison in Baghdad.
The request was brief. It asked that the Pentagon pull up the live satellite feed over a set of coordinates in western Diyala. The coordinates were for a compound 3 km north of a village called Al Zaidan. Elkins was in a meeting when the request arrived. He read it when the meeting ended at 16:22. He did not understand what he was looking at.
He asked the staff officer who had made the request. The staff officer said it had come from Lieutenant Colonel Pike in Baghdad and that Pike had asked specifically for Elkins to see the feed before any further reporting went up the chain. Elkins walked to the screening room on the fourth floor. Three officers were already there. The CIA liaison from the briefing the previous morning was there.
The feed was being brought up. The feed showed the compound. What the feed showed at 16:24 Washington time on October 18th, 2006, was a compound in western Diyala at what was then 00:24 the following morning local time. The compound was quiet. There were no vehicles moving. There was no smoke. There was no active perimeter.
There was a marker on the feed. The marker was a small white cross placed by the analyst who had pulled the feed at the point in the compound where, approximately 30 hours earlier, a six-man patrol had removed a man from a chair beside a laptop and a cup of tea. Beside the marker in the margin of the display was a six-line summary.
The summary listed time of entry, number of rounds fired, number of personnel in the target element, the identity of the individual taken alive, and two figures at the bottom. The bottom two figures were the ones that made the room go quiet. The first was the elapsed time from breach to extraction, 6 minutes, 11 seconds.
The second was the personnel count on the coalition side. Six. No one in the screening room spoke for 11 seconds. Then the CIA liaison said, quietly, one word. Six. Elkins looked at the feed. He looked at the summary. He did not speak. He had been in the building 36 hours earlier, in the briefing room on the fourth floor, telling 11 other officers that a six-man SAS patrol had failed.
The patrol had not failed. The patrol had taken Abu Khalid al-Janabi alive at 0217 on October 19th, 2006, from a chair next to a cup of tea in a compound that had not appeared on any task force board because no one at the task force had been looking at the right road. The meeting at the Pentagon the following day was not a formal briefing.
It was held in the same room where Elkins had closed the SAS proposal 36 hours earlier. Seven officers were present. The red folder with the black stripe and the secret no foreign stamp was still on the table. A clerk had moved it during the night cleaning 3 ft to the left. No one had opened it. Elkins walked into the room at 0910.
He had the summary from the screening room in his hand. He sat down. He set the summary on top of the red folder. He did not open the folder. He did not open it that morning. He did not open it that afternoon. He did not, the record appears to show, open it at all. The folder sat on the table for 3 days. On the fourth day, it was filed by a junior officer who had not been present at the original briefing, and who processed it in accordance with the standing procedure as a closed operational file.
The junior officer did not read it. He filed it under the operation name that had been assigned to the original proposal in August 2006. The operation name was Thresher. The file under that name contained the cable Elkins had signed, the proposal he had denied, the meeting log entry from August 27th, and nothing else.
The record that the operation had succeeded, and that it had succeeded by the method Elkins had laughed at, was in a separate file in Baghdad held by Pike. It was never merged. The numbers are worth sitting with. The hunt for Abu Khalid al-Janabi had run in some form for 26 months. Six coalition operations had been conducted against him directly.
The operations had involved, in total, over 2,400 personnel at various points. The cumulative cost, on the most conservative published estimate, was $180 million. The operations had killed 37 people, four of whom had been directly in his network. The operations had produced no progress on the file. The operation that closed the file had involved six men, one fruit seller, 11 pages of notes in a broken-spined leather notebook, and 7 weeks of quiet.
It had cost, on the patrol’s own accounting, $11,000. Most of that was fuel. The patrol fired 11 rounds. The engagement lasted 17 seconds. The official record, as it currently stands in the Pentagon archives, does not reflect any of this. Elkins retired from the Army in July 2009, 33 months after the events at al-Zadan.
His retirement remarks, delivered at Fort Leavenworth, made no reference to the operation. They referred in general terms to lessons learned from the coalition campaign in Iraq. He spoke for 24 minutes. He did not take questions. Pike retired 4 years after that, in 2013. He gave one interview to an Army historian, in which the events of October 2006 were discussed across two sessions totaling 4 hours.
The transcript of that interview remains sealed. Barnard is not publicly identifiable. He served for another 6 years. What he did after leaving the regiment has not been confirmed by any open source. Saad continued to sell fruit at the market near the intersection of the two roads. He was asked once, 3 years later, whether he had understood what he was doing when he talked to the patrol at his stand.
He said he had understood only that the men asking him questions were polite, and that they had not pressed him for things he did not know. He was asked if he had ever told anyone in his family what he had done. He said he had not. There was no point. Nothing had changed. The white Toyota Hilux had not passed his stand since.
The folder on the table at the Pentagon, by the time the story of Abu Khalid’s capture appeared in the third sentence of a four-paragraph press release from the coalition office in Baghdad on 2nd of November, had been filed for 10 days. It remained in the filing system, a closed operational record. The stamp still read secret no foreign.