Vangjel's own data shows that US negligence caused Calipari's death, and that the driver (Carpani) acted properly. Vangjel's distance data shows that Carpani stopped the vehicle as quickly as he could have been expected to, given the very late warning our troops provided. The US numbers don't add up in the sense that they don't support the US conclusion, that we acted properly and the Italians didn't.
Vangjel's report also clearly shows multiple factors of US negligence in training, equipment and communications. (Much or most of the information supporting this was ineptly redacted. Redacted and unredacted versions of the Vangjel report can be found here.)
Elapsed time from first warning to deadly fire: less than two seconds
The Vangjel report includes no maps or diagrams, except perhaps in annexes which are mentioned, but which I haven't been able to locate. However, he specifies distance measurements which indicate the layout of the scene. In the following analysis I take Vangjel's distance specifications at face value, even though there's ample reason to question them.
The first relevant position is the point where the gunner saw Sgrena's vehicle (what I'll call "sight point"). 21m beyond this point is the "alert line." 36m beyond this point is the "warning line." 45m beyond this point is where the vehicle came to rest (what I'll call the "stop point"). 38m beyond this point is the position of the gunner (these numbers aren't explicitly indicated by Vangjel, but they're derived directly from the data he reports).
Vangjel reports that when the gunner saw the car, he pointed his spotlight at it. (I think the spotlight was a fairly large handheld device that looked something like this.) This presumably required some reaction time on the part of the gunner, but ostensibly he had the spotlight shining on the car by the time it arrived at the alert line.
With regard to the reaction time required to position the spotlight, note that the gunner was holding it with only one hand (his left). It's presumably a fairly large, unwieldly device, and he had to both operate a switch (probably a trigger) to turn it on, and point it in the right direction (and then using only one hand he had to keep the beam focused on a "speeding" vehicle).
I assume a speed of 50 mph, because this is more-or-less what Vangjel claims. Given Vangjel's claim that the spotlight was shining on the car by the time at arrived at the alert line, this means that the gunner managed to perceive the car, raise the spotlight, turn it on and place the beam on the car, using only his left hand, all within less than a second (a car traveling 50 mph would traverse the 21m distance between the sight point and the alert line in 0.94 seconds).
Vangjel makes a similar claim regarding another soldier who was operating a laser pointer. The spotlight and the laser pointer were the only warning devices used.
What happened next? Shots were fired. An astonishing admission by Vangjel is that "warning shots" were followed virtually immediately by shots that were attempting to "disable" the vehicle: the gunner "fired warning shots as [the car] reached the Warning Line, and fired on the vehicle in an attempt to disable it immediately after it crossed the Warning Line."
More confirmation that warning shots weren't fired until the car reached the warning line: "When the car got to the Warning Line, [the gunner], while still holding the spotlight in his left hand, used his right hand to quickly fire a two to four round burst into a grassy area to the on-coming vehicle’s right (the pre-set aiming point) as a warning shot.
If warning shots were fired as the car "reached" the Warning Line, and then shots were fired at the car "immediately after it crossed the Warning Liine," that's tantamount to admitting that there was virtually no pause between the "warning" shots and the shots directed at the vehicle.
So how long had the spotlight been on, at this point? Vangjel claims the spotlight hit the car before the car reached the alert line. Given reaction time, that seems doubtful. In any case, it couldn't have been much before. Vangjel acknowledges shots were fired directly at the car "immediately after it crossed the Warning Line." According to Vangjel, the distance from the alert line to the warning line is 36m. A car traveling 50 mph would traverse this distance in 1.6 seconds.
In other words, Vangjel admits (indirectly) that deadly fire was directed at the vehicle less than 2 seconds after the first warning was given.
Note also that the gunner "dropped the spotlight" as soon as he began firing directly at the vehicle (so he could control his weapon with both hands). In other words, the total amount of time the spotlight pointed at the car was apparently less than two seconds. Note that aside from the laser pointer, this spotlight was the only warning given.
Note that Vangjel realizes it's poor policy to require the gunner to also handle the spotlight: "Recommend an assessment of the current technique of requiring the gunner to operate both the spotlight and the weapon in the turret of the vehicle. This will allow more reaction time and simplify duties and responsibilities of the gunner."
Driver stopped as quickly as he could
Vangjel's own data seems to show that the driver hit the brakes and stopped the car as quickly as humanly possible.
Let's assume the driver stopped the car in reaction to seeing the spotlight. How far did the car travel from that position (where the spotlight hit it, i.e., the alert line), before it finally came to a stop? According to Vangjel, the car came to rest 81 meters beyond the alert line.
This is consistent with the driver attempting to stop as soon as he could. A car traveling 50 mph in the rain needs about 250 feet (76m) to stop, once perception time and reaction time are taken into account (link; many thanks to the poster of this comment). The small car had a relatively heavy load (3 adults), which would increase stopping distance.
In other words, the car came to rest within about one car length of where one would expect, assuming the driver reacted as quickly and decisively as humanly possible to the very first warning signal.
By the way, Vangjel indicates many other cars (15-30) were able to stop safely (maybe the gunner who had just been rotated out had done a better job of wielding the spotlight, or maybe the weather had just gotten worse). However, he acknowledges "many of the vehicles, though, screeched their tires when stopping." In other words, our troops did this for well over an hour, and noticed that civilians were having trouble stopping safely. Yet they lacked the training and/or equipment to deploy any of a number of safety devices to provide oncoming drivers with earlier warning.
Notice that we have cameras in space that can miraculously see through clouds and accurately measure the speed of a small car. But we can't afford to equip our troops with some stuff like this.
Speed was only slightly above normal
Vangjel indicates a radar operator was sent to the site to measure typical traffic speed (at the alert line). 57 vehicles were clocked on another Friday night (although conditions were dry). Result: about 45 mph. In other words, Carpani was going only slightly faster than average traffic does, at this location. And this is assuming that we accept Vangjel's speed estimate, and reject the Italian claim that they were going much slower.
Bullet holes are in the wrong place
Vangjel says "11 rounds struck the vehicle with only five of those impacting the front of the car ... A forensic examination of the car was performed after its removal from the scene. This analysis disclosed 11 entrance bullet holes. They are consistent with 7.62 mm bullets. Three bullets perforated the front section of the car at the bumper, right head light, and right fender. Two bullets perforated the windshield. Six bullets perforated the right side, right door, right front and rear passenger windows."
(By the way, elsewhere Vangjel acknowledges that a bullet flattened the left-front tire, but that bullet seems to be erroneously omitted from the above count.)
Given that the car was traveling directly toward the gunner, and ostensibly came to rest 38m from the gunner (i.e., the car ostensibly did not reach the gunner or pass him), it's hard to understand why a relatively large proportion of the bullet holes are on the side of the car, rather than the front. It's also hard to understand why most of the other windows were shattered (including the rear window), but not the windshield.
Vangjel says the gunner fired "towards the car’s engine block in an attempt to disable it ... perceiving the on-coming vehicle as a threat, he fired to disable it and did not intend to harm anyone in the vehicle." In that case, it's hard to understand why there's apparently no damage whatsoever to portions of the car surrounding the engine (such as the grill and the hood) and a great deal of damage to portions of the car surrounding the passengers (such as the side and rear glass).
"Checkpoint" vs. "roadblock"
Vangjel focuses quite a bit of attention on this important distinction. A checkpoint is designed to examine vehicles and then let them pass. A roadblock is designed to stop all vehicles and get them to turn back, no matter what. A roadblock is often setup in a hurry and taken down in a hurry. In Vangjel's milspeak, a checkpoint is called TCP, or "Traffic Control Point." A roadblock is called BP, or "blocking position."
Vangjel refers to three categories of checkpoints: "deliberate," "hasty" and "flying." Flying means "conducted immediately and often with little or no planning." Vangjel says "although not a TCP [Traffic Control Point, or checkpoint] mission, the mission given to 1-69 IN to block Route Irish on 4 March 2005 fell into this [flying] category."
In other words, this incident occured at a roadblock, not a checkpoint ("it is clear that these BPs were not established as TCPs"), and it was a roadblock created with "little or no planning."
Poor training
Vangjel gives many indications the troops were poorly prepared for the mission they were given. He says "there is no written SOP [standard operating procedure, i.e., documented instructions] or TTP [tactics, techniques and procedures] in 3ID, 2/10 MTN, or 1-69 IN [the various units involved] for the execution of the blocking mission and establishing a blocking position."
He also says "there is no evidence to indicate that the Soldiers were trained to execute blocking positions before arriving in theater ... the only training received by 1-69 IN Soldiers on blocking positions was that employed along Route Irish during after-curfew Rhino Bus Runs, and occurred during the Left Seat Right Seat Ride process with 4-5 ADA." In other words, these troops had received only a limited amount of on-the-job training, with regard to how to operate a roadblock.
They had also been working in this area barely two weeks: "The Commander, 1st Cavalry Division assigned the unit the mission of patrolling and securing Route Irish as of 15 February 2005"
Vangjel also says "there is no clear guidance in these units on what equipment is required for establishing a blocking position (e.g., different road signs)."
This poor training probably accounts for why the roadblock was placed on a curve, in a limited-visibility position where both the gunner and the driver would be surprised to see each other: "[officer] was under a time constraint to establish the BP quickly and expected to be in position for a very limited time, i.e., no more than 15-20 minutes. Further, the position was on a tight curve that caused [officer] to make less than optimal choices in positioning his vehicles."
Poor equipment
Vangjel refers to a variety of techniques and devices that are often used at checkpoints, such as "barrels filled with sand, water, or heavy concrete blocks (emplaced to slow and canalize vehicles), concertina wire (emplaced to control movement around the checkpoint), and signs stating the speed limit into and out of the checkpoint (in both English and the local language.) ... The equipment for a TCP includes warning signs, triangles, sawhorses, traffic cones, and/or tire poppers."
How many of these techniques were used at this roadblock? None.
By the way, it seems that 25 signs were in storage waiting to be modified: "the last part of that phrase 'or you will be shot' was to be covered with tape ... the signs had not been modified, and, therefore, not reissued as of 4 March 2005."
Poor communication
These troops were in place much longer than they were supposed to be: "normally, these blocking positions, which were hasty in nature, would be held for 10-15 minutes before the TOC would order the road opened ... [The officer] expected to maintain the blocking position no more than 15 minutes." These troops were kept in their vulnerable position for well over an hour. Why? Poor communication.
Poor communication (by our military, not by Italians) was a factor in the incident. Vangjel admits this: "ineffective battle tracking procedures (communications, log posting, and information sharing) at the 1-76 FA TOC caused A Company, 1-69 IN to be left in their blocking positions longer than expected ... communications problems involving a unit new to the AOR that caused the Soldiers to be left in position longer than expected."
Vangjel explains: "The 1-76 TOC had two means of communicating with 4th Brigade, its higher headquarters: Voice Over Internet Protocol (VOIP)2 and FM. The 1-76 FA Battle Captain was using only VOIP to communicate with 1-69 IN, but experienced problems with VOIP, therefore losing its only communication link with 1-69 IN, other than going through 4th Brigade ... As a result, the Battle Captain was unable to pass updated information about the blocking mission either directly to 1-69 IN, or to 4th Brigade. He did not attempt to contact 4th Brigade via FM communications."
In other words, even though Negroponte arrived at "the Camp Victory gate at 2010 hours," at 2050 these troops were still in place to protect his convoy, which had long since passed. "There is no evidence to indicate that 1-76 FA passed on the information about the VIP departure and arrival times to any unit ... As a result, A Company, 1-69 IN’s Soldiers were directed to remain in their blocking positions."
To add to the confusion, incorrect information was conveyed to the troops: "At 2030 hours, Captain Drew asked again about collapsing the blocking positions. He was told that the word from 3ID was not to move off the blocking positions, that the convoy would be coming down Route Irish in approximately 20 minutes." In other words, 20 minutes after Negroponte had already arrived at Camp Victory (near the airport), our troops were being told to expect his convoy 20 minutes later.
In other words, due to poor communication a hastily-planned roadblock designed to stay in place for 10-15 minutes was instead kept in place for well over an hour, in a position that created hazard to our troops and to civilians. Vangjel says our troops were not happy about this: "Staff Sergeant Brown, the acting Platoon Sergeant, was seeking to determine how much longer they were to remain in position. As such, he was standing with Second Lieutenant Acosta near the overwatch vehicle, their backs to the on-ramp ... Captain Drew, Second Lieutenant Acosta, and Staff Sergeant Brown were all concerned about the length of time that the Soldiers had been manning their blocking positions ... Captain Drew was concerned that leaving his Soldiers in a static position for more than 15 minutes left them open to attack ... Captain Drew checked with the 1-69 IN TOC at least two times seeking to collapse the blocking positions and return his Soldiers to their patrolling mission."
This surprising failure in communication sort of makes moot the whole question of whether or not the Italians properly notified US authorities (and there's ample indication such notification took place). If their commanders had wanted to tell these troops to be careful not to shoot Sgrena, we now understand that failures in communication gear would have blocked that message.
What our government said, then and now
Here's what our government said at the time:
"About 9:00 pm, a patrol in western Baghdad observed the vehicle speeding"
We've established that even the Vangjel claim of 50 mph is only slightly above normal for this location. So the assertion of "speeding" is highly questionable, at best.
"towards their checkpoint"
It was a hastily-placed roadblock, not a checkpoint (the latter term implies a variety of safety techniques that were clearly not present).
"and attempted to warn the driver to stop by hand and arm signals"
Vangjel mentions no such thing. The gunner literally had his hands full. He had to drop the spotlight in order to use both hands on his weapon. Therefore this claim appears to be an outright fabrication.
"flashing white lights"
At the last moment (very shortly before fatal fire was used), a spotlight was turned on for about two seconds. This is hardly what's implied by the term "flashing white lights."
"and firing warning shots in front of the car"
We've established that the shots into the car were virtually simultaneous with the "warning" shots. This is tantamount to understanding that there were no "warning" shots at all.
"When the driver didn't stop"
This implies that there was a period where the gunner waited, giving the driver a chance to stop. We've shown that deadly fire was used less than two seconds after the first warning, and we've shown that the driver stopped as quickly as humanly possible.
"the soldiers shot into the engine block"
Vangjel claims the gunner aimed at the engine, but there is apparently no sign that bullets reached the engine.
"which stopped the vehicle"
Vangjel doesn't claim that the vehicle stopped as a result of bullets striking the engine. In any case, that would not be a particularly quick way of stopping a vehicle. A "speeding" car can coast for quite a distance with the engine turned off. The only plausible explanation for how the car stopped as quickly as it did is the driver slammed on the brakes.
In other words, virtually everything our government initially told us about the incident has now been proven to be a lie or a distortion.
Speaking of propaganda, let's not forget the "senior US military official" who said "the investigation ... will ultimately prove the officer's car was traveling in excess of 100 mph." That's about as laughable as the idea that we have satellites constantly monitoring the speed of every car in Baghdad.
Other interesting statements by Vangjel
Vangjel acknowledges we don't really know the exact position of the vehicles involved, because they were all moved soon after the incident, before careful measurements could be taken. The area should have been treated as a crime scene, but it wasn't.
Vangjel also acknowledges that no thorough investigation of the scene ever took place subsequently, because we were apparently unable or unwilling to provide adequate security at this dangerous location (but we had no trouble requiring our troops to stay in this exact position, at night, much longer than they needed to be there, creating hazard to themselves and others).
Our troops have been well-aware of these problems
One wrote this:
"Without signs and roadblock kits of any form, Marines resorted to firing a warning burst at the approaching traffic. When the approaching traffic didn’t read the mind of the Marine and come to a sudden stop after the warning burst, they were engaged by multiple weapons systems and destroyed. Some of these vehicles, if not most, were civilians attempting to leave the battlefield too late, and did not obey the warnings of the Marines blocking the road. Natural human reaction to being fired upon is to flee the scene of the fire, yet we somehow expected these individuals to know that we wanted them to stop, without communicating our desire to them in any other form short of gunfire. The fault was not that of the Marines, forced into these blocking positions in the ever-fluid latter stages of combat operations. The fault was in not having a plan for this in advance and facilitating the requirements for this action for the Marines. In the future, I recommend pre-planning to the point of having 'stop' signs in the language of the country we’re attacking, along with collapsible tetrahedrons, concertina wire, roll out tire spike strips, and other traffic control devices. U.K. forces, due to their experiences in Northern Ireland, displayed a proficiency at blocking and directing civilian traffic that made our efforts look amateurish[emphasis added]. In order to minimize the collateral damage that endangers mission accomplishment, we need to effectively convey our intent to civilians at traffic control points."
Like this soldier, Vangjel realizes we need to learn from our allies: "Recommend that permanent Coalition participation be included in the Force Protection Working Group to solicit lessons learned from other nations’ experiences in operating ECPs, TCPs, and BPs in an insurgency environment."
Vangjel's very first recommendation is a clear indication of the root causes of the incident: "Recommend the Force Protection Working Group consider the use of additional non-lethal measures (e.g., spike strips, temporary speed bumps, and wire) be emplaced to slow down or stop vehicles before the use of disabling shots. The intent is to provide as many non-lethal options as possible before asking a Soldier to focus on firing the weapon."
This passage was redacted. Likewise for other passages like this suggestion: "Highly visible and quickly deployable checkpoint and roadblock warning signs for Soldiers on patrol."
It's hard to understand the purpose of these redactions, unless there was a desire to avoid admitting how we failed.
Update, 5/2, 12:55 pm
A while back I wrote a diary speculating about the exact location of the incident. I included an aerial photo. I was wrong. At the risk of making the same mistake twice, I think this photo shows the scene (pdf). Look for the large interchange at the bottom-right of the image. Zoom in for a close view. A variety of details in the Vangjel report seem to match up with this location.
I think this map of Baghdad is also particularly helpful. Look for the first large interchange east of the airport.