Against Police Heroism

There is no free active shooter lunch

Occasionally, I use this blog to opine on social conventions or systems without providing the reader any evidence or new data to back up my argument. This was going to be such a post, regarding police tactics in Uvalde — my main point simply being that any policy which risks officer lives will eventually lead to even poorer execution of police-work¹, and this includes policies on school shootings.

However, I ended up putting in a bit more reading and writing a more specific defense of the police response at Uvalde and criticism of current wishful-thinking policies on active shooter responses.

A caveat which applies to this entire post is that it is regarding policy and training; it is not an argument against ad hoc “heroic” actions that result in “happy outcomes.” Such “happy outcomes” will always occur on their own when 1) Police engage shooters in the assessment phase, without intending to 2) These engagements do not spontaneously produce disastrous outcomes. But such outcomes (which are good, and may be termed “heroic”) cannot be promoted or discouraged by any policy; they exist outside of policy.

The DOJ recommendation in the Uvalde report is stupid

Uvalde has interested me since it occurred — not in the sense of conspiracy theories about intentional police enablement, but simply because I have an interest in the problems of police work. However, I have refrained from actually paying any attention to the subject until today, when the report of the Justice Department investigation is available to provide a condensed and hopefully accurate summary of the event.

The report is here; I have read the event summary (p. 7), the sections on historical changes in police response practices (p. 86), tactical observations and recommendations (p. 109), and policies and procedures (p. 357). For a general summary of the DOJ’s report and recommendations, the reader may refer to the New York Times story here.

This is the overview of the event, with emphasis on the major points:

At 11:33 a.m. on the morning of May 24, 2022, the subject entered Robb Elementary School equipped with a high-powered AR-15 rifle. He immediately started shooting and within a minute entered classrooms 111 and 112, which were connected via an interior door.

Within 3 minutes of the subject’s entry into the school, 11 law enforcement officers from the Uvalde Consolidated Independent School District (UCISD) and Uvalde Police Departments (UPD), including supervisors, arrived inside the school. Hearing continued gunfire, five of the responding first on scene (FOS) law enforcement ran toward classrooms 111/112. The other six FOS did not advance down the hallway, including UPD Acting Chief Mariano Pargas, who was in the best position to start taking command and control, and to start coordinating with approaching personnel. One of the officers said to “line up to make entry” and within seconds shots were fired from inside one of the rooms. Two officers were hit with shrapnel, and all responders retreated to positions of cover.

After three attempts to approach the classrooms, the focus of the responders shifted from entering classrooms 111/112 and stopping the shooting to evacuating other classrooms, attempting to negotiate with the subject, and requesting additional responders and equipment. With this shift from an active shooter to a barricaded subject approach, some responders repeatedly described the subject over the radio as “barricaded” or “contained.” Yet within four minutes from FOS arrival, 911 dispatch confirmed that class was in session and reported that they had received calls from victims.

Chief Pete Arredondo of the UCISD Police Department (UCISD PD) directed officers at several points to delay making entry into classrooms 111/112 in favor of searching for keys and clearing other classrooms. Occupants of other classrooms were at risk of further injury as a result of the high-powered nature of the shooter’s AR-15 style rifle and from possible crossfire once classrooms 111 and 112 were entered. At several points, UCISD PD Chief Arredondo also attempted to negotiate with the subject. Others called out over the radio for additional resources and indicated that they were waiting for a tactical team to arrive, such as Uvalde special weapons and tactics (SWAT), the Texas Department of Public Safety (TXDPS), and the U.S. Border Patrol Tactical Unit (BORTAC). Chief Arredondo, who became the de facto on-scene commander, was without his radios, having discarded them during his arrival, and communicated to others either verbally or via cell phone throughout the response.

Over the course of the incident, overwhelming numbers of law enforcement personnel from different agencies self-deployed to the school. Leadership on scene, however, had not established command and control, to include an incident command post (ICP), staging area, or clear perimeter around the hallway or the school. Thus, arriving personnel did not receive accurate updates on the situation or direction for how to support the response efforts. Many arriving officers—based on inaccurate information on the scene and shared over the radio or from observing the lack of urgency toward entering classrooms 111/112—incorrectly believed that the subject had already been killed or that UCISD PD Chief Arredondo was in the room with the subject. As leaders from additional law enforcement agencies arrived, including Uvalde County Sheriff Ruben Nolasco, the lack of clear communication and command structure made coordination difficult. Emergency medical responders faced similar challenges as they deployed. They struggled to identify who was in charge, and ambulances encountered streets blocked by law enforcement vehicles.

Concerned families were also arriving at the school. They likewise had difficulty obtaining information about their loved ones’ status. Incorrect and conflicting information was also being shared on social media with the UCISD posting that all students and staff were safe in the building and later posting messages about reunification that conflicted with the UPD posts.

At 12:21 p.m., 48 minutes after the subject entered the school, the subject fired four additional shots inside classrooms 111/112. Officers moved forward into formation outside the classroom doors but did not make entry. Instead, presuming the classroom doors were locked, the officers tested a set of keys on the door of a janitor’s closet next to room 112. When the keys did not work, the responders began searching for additional keys and breaching tools. UCISD PD Chief Arredondo continued to attempt to communicate with the subject, while UPD Acting Chief Pargas continued to provide no direction, command, or control to personnel.

After another 15 minutes, officers found a second set of keys and used them to successfully open the janitor’s closet. With working keys in hand, the officers then waited to determine whether a sniper and a drone could obtain sight of and eliminate the subject through the window. Those efforts were unsuccessful.

At 12:48 p.m., 27 minutes after hearing multiple gunshots inside classrooms 111 and 112, and 75 minutes after first responders first entered Robb Elementary, officers opened the door to room 111. A team composed of BORTAC members, a member of the U.S. Border Patrol Search, Trauma, and Rescue Unit (BORSTAR), and deputies from two local sheriffs’ offices entered the rooms, and officers killed the subject when he emerged shooting from a closet. The subject was killed at approximately 12:50 p.m., 77 minutes after the first officers entered the school and after 45 rounds were fired by the shooter in the presence of officers.

This is what (per the NYT summary) the DOJ now terms a “Unimaginable Failure” (the phrase does not appear in the report itself). This of course is merely a bureaucratic lynching of the responsible parties to appease the media and public. “Something bad happened; somebody in government should tell us that this was because someone messed up, and they have been punished.” Heads roll, the gods are appeased, etc.

The police response at Uvalde was neither an unimaginable failure nor an obvious triumph. All things considered, it was a reasonable response, and may have been the best possible in terms of reducing deaths.

Before discussing “policy” (a misleading term in this application), let us consider the principle recommended response of the report:

Officers responding to an active shooter incident [as in, the 5 of 11 first arriving officers who initially approached the shooter] must continually seek to eliminate the threat and enable victim response. The shooter’s immediate past actions and likely future actions serve as “triggering points” that indicate the appropriate response should be in line with active shooter response protocols. An active shooter with access to victims should never be considered and treated as a barricaded subject.

This, taken to a logical end, implies both that the five officers should have pressed forward regardless of casualties and regardless of danger to civilians from crossfire. (Note: This is the actual “policy” recommended by supra-police organizations for active shooter incidents.) Who is to say that this would have resulted in fewer civilian deaths? Couldn’t it easily, obviously, have resulted in more? And here is the crux of the meaninglessness of the “unimaginable failure” proclamation of this investigation. The investigation actually does not show any reason to believe that this is a better response to this scenario in general.

Side comment on tactics

Most Americans have wildly inaccurate ideas about the feasibility of resolving a shootout with handguns, such as would be available to police on scene in advance of tactical reinforcements (i.e. SWAT etc.). The author, yours truly, gets his notions from research on firearms as well as frequent visits to the Police Activity youtube channel.

Consider this recent shootout which closed a freeway in Houston (youtube.com, embed is not available). The shooter, a 19 year-old male, is able to shoot a police officer, run toward a vehicle, enter it, exit it, and run toward another vehicle — all while surrounded by officers who are firing their semi-automatic handguns (9mm most likely) from cover (their own vehicles) — before he finally falls to the ground. At this point, he is still able to attempt to reload his weapon — he has not been truly incapacitated or neutralized.

Realistically, were the police to not react by firing at him even more, he would still be capable of firing a weapon for several minutes — the resolution of “firing at him until he is dead” in this case is only available because he was shot and fell without cover. Otherwise, he might with his present injuries have been able to continue firing at police for 5 minutes, if we take the 1986 Miami FBI shootout as an example (wikipedia).

Let the reader understand the Houston example as a proxy for “multi-officer threat neutralization capacity” in sub-optimal but not poor conditions. Now imagine how such a tremendous (sarcasm) threat neutralization capacity would play out while sieging a classroom full of children. It’s possible children’s lives would have been saved. It’s possible all the police would have been killed and no children’s lives saved. It’s possible some of the police would have been killed, and more children’s lives lost, many more, especially by police fire.

Keep in mind, as well, that although the Houston example was an ad hoc response to a stopped car which suddenly became an active shooter situation, all “plans” are merely plans until someone gets shot in the face. There is no “plan” that can assure any result.

So, wait, what is the plan?

If the above, albeit brief, discussion of tactics has convinced the reader that from the moment of being fired upon during approach, the further actions of waiting for reinforcements, containing the shooter, attempting to apprehend the disposition of the shooter and civilians with drones, and finally storming the shooter with overwhelming and deliberate force (reducing, in theory, the chaos of the final showdown) were actually a really great approach (notwithstanding organizational and leadership chaos and some number of wasted minutes in the interim, i.e. a lack of perfection in execution²) — then the question might arise, isn’t this what the “plan” is? And if not, why not?

The plan of the UPD/UCISD

Bearing in mind that Uvalde Police Department (UPD) and school police (UCISD) officers were first on the scene, the “policies” in question were as follows:

Page 385 and 386 of the DOJ report detail UCISD training on active shooters which had been the substance of an active shooter training course just months before the shooting. Critically, the emphasis is on:

• ISOLATE—Drive or segregate the attacker in an area where their capacity to harm students, staff, or visitors is minimized until more first responders arrive.

In other words, the policy and training of the first responders at Uvalde was not to charge into rifle fire or fire blindly into crowded classrooms from doorways until no more sounds are heard.

Whatever one thinks of this policy, it was the policy. The officers followed it.

Therefore, there was nothing wrong (lowercase-w, follow-the-rules “wrong”) with the response at Uvalde besides the facts that 1) the local emergency response and law enforcement agencies in question were not flawless (they had not pre-rehearsed their response over and over with billions of dollars of free money), 2) local policy disagreed with a consensus of supra-police organizations and federal agency vogue, 3) the outcome did not deliver a lower death count than it did (which is axiomatically impossible; just as knowing whether the death count that did occur is higher or lower than what was possible (again, it may have been much lower than with the “generally accepted” plan).

The plan of the DOJ and consortium of related organizations that form “generally accepted” standards

It contrasts with the “generally accepted” policy for responding to active shooters in schools or elsewhere. The generally accepted policy, echoed in the DOJ’s recommendations, is that the first-arriving police force (i.e. individual officer or group of officers) should direct itself toward “stopping the killing and stopping the dying.”

In a recent “ideal example” cited by the DOJ report, a single officer in an Allen, Texas mall stopped an AR-equipped shooter last May. First of all this outcome, like the outcome to any shootout, would have been delivered to the officer by the grace of the Cover Gods, not by an idiotic policy of “run toward shooters always, until either you or they are dead.” Second of all, the shooter still killed eight and injured 7 — not much “better” than the outcome at Uvalde.

Just for an example of how realistic this policy is on a tactical level, in a school, consider the DOJ report’s on comments on the use of shields in active shooter situations (emphasis added):

Shields

According to the FBI’s 20-Year Review of Active Shooter Incidents (2000–2019), approximately 43 percent of the 333 identified active shooter incidents included the presence of a long gun/rifle by the subject. A tactical ballistic shield, capable of absorbing impact from rifle rounds, serves as an invaluable tool for law enforcement officers, providing an added layer of safety in high-risk situations. Given active shooter response protocols require immediate action, oftentimes from officers in a patrol function, in many law enforcement agencies throughout the U.S., it is unlikely that a responding officer will be able to respond with a shield. If a shield is available, they must be used in accordance with tactical training in order to be effective. However, an officer should never wait for the arrival of a shield before moving toward the threat to stop the shooter. Proper training on the use of a shield is essential for an officer to master the maneuvering, tactical positioning, proper handling, firearm position and accurate fire, and transitioning off the shield. Although many officers may receive some exposure to the use of a shield at academy training, it is typically with a non-ballistic rated shield, typically used for crowd control. Furthermore, they are often not retrained on their use unless they are for specialized training, such as SWAT. Without SWAT experience, most patrol officers will not have training on use of a shield in a tactical situation such as an active shooter.

But the current “generally accepted” policy is unproven

This policy, importantly, is a reversal from historical trends — an aberration from more long-tested methodologies which was spurred by the Columbine shooting. The DOJ report provides a good summary of this development on page 86. Essentially, before Columbine, the Uvalde policy represented standard practice — contain and wait for SWAT.

Throughout most of history, the police response to an active shooter incident was to secure a perimeter and call out a special weapons and tactics (SWAT) team and, in some cases, negotiators. Most officers lacked specialized, advanced training and preparation to handle such situations.

Modifications to the new kamikaze pseudo-policy of “stop the killing, stop the dying” have followed observations (investigations) of later shootings. For a long time, the principal concept was that four or five officers would charge toward the active shooter with their handguns and imaginary anti-rifle shields. In Marjory Stoneman a lone officer, following this generic concept, did not try to stop the shooting. So now the concept is that if only one officer is present, one officer should go.

Post-Columbine, rapid deployment in a crisis is predicated on the first four or five responders forming an active shooter response team and making entry. In a situation where seconds count, any delay in response could be detrimental to potential victims. However, due to the different response times to an active shooter situation, the hard-and-fast recommendation of having a four-to-five-member rapid deployment response team has been reduced to where some agencies authorize a single officer to enter an active shooter location.385 Recent tactical training also emphasizes solo officer response when the situation demands. These tactics have arisen due to incidents where immediate action by a lone officer could have helped save lives.

Note that none of the language is concrete, because no one making these recommendations actually knows that they work.

There has not been any scientific study of active shooter interventions — no randomized control trial demonstrating that any benefit accrues from police rushing in not just without overwhelming tactical superiority, but in an obvious potential case where the shooter has that superiority. What is actually the point of a plan such as “throw meat shields at the problem and hope for the best?” (Again, as in my caveat at the head of this post, encounters during police assessment will still happen on their own and resolve “happily” at times, regardless of policies.)

The present “generally accepted” policy is all just wishful thinking structured around analysis of a handful of events which may have no relationship to any particular future event — they do not represent a “best practice,” taking into account all possibilities, but merely a Monday-Morning Quarterback fantasy overlaid over a few known past outcomes. Consider one of the past outcomes in question, and the resulting fantasy (emphasis added):

The establishment and proper use of an incident command structure has also been a common lesson learned across incidents, as was prominently the case in the active shooter response at the Washington Navy Yard Building in 2013. Though law enforcement entered the building within minutes, it took over an hour to locate and kill the shooter, due to the size of the building. Among the key lessons learned was the value of establishing an incident command structure early on to coordinate resources.

“Establishing incident command structure early on,” while it may or may not accelerate the locating of hostiles, is not an achievable goal — to be put in place everywhere in America, at all times, not just here and there, sometimes, would require billions of dollars of police and other agency training. This money does not exist and shouldn’t be spent this way anyways.

But even more to the point — the “ideal” response occurred, and didn’t accomplish anything. Police wandered aimlessly, at who knows what risk of attack from the shooter, or of friendly fire, or accidental civilian shooting (I am speaking entirely hypothetically), and the resolution occurred well after a SWAT team would have arrived had they been waited for to begin with.

Making a pseudo-policy based on a principle of “just one more unrealistic expectation and we’ll get it right this time, bro, just one more I promise” is nuts (technical term) — and it’s only by stupid luck that these previous instances of “stop the killing, stop the dying” haven’t yet produced an “unimaginable” disaster.

A better plan

Police should seek to contain shooters until overwhelming force arrives — essentially as took place in Uvalde.

Such a policy — in which “heroics” are not unrealistically and recklessly demanded of police — will not rule out heroics in practice. Containment still requires assessment and approach of the shooter; whenever it happens that police initially encounter the shooter where is not, yes, barricaded as in Uvalde, engagement will follow with a chance that the shooter is neutralized. No missed opportunities for reasonably risky early termination of the threat to civilians will follow a policy discouraging unreasonably risky confrontations.

Further, over the long term, the “generally accepted” policy will produce worse outcomes. Police should not be encouraged to confront shooters at a known disadvantage; this must lead to increased risk of police killing of civilians (e.g., police will be more trigger-happy because they are more at risk).

While the current pseudo-policy is not backed by actual “science,” the problems with it are obvious and do not require studies or proofs. It is patently unrealistic — just bureaucratic wishful thinking.

If you derived value from this post, please drop a few coins in your fact-barista’s tip jar.

click button or visit https://ko-fi.com/unglossed

1. Hence why police work is never “good” and can never be “good” when it comes to traffic stops, as there is a substantial risk to police life and extreme asymmetry in knowledge of the threat (only the threat knows if they are a threat). It is not therefore “good” to introduce extra risk to police officers in school shooting contexts; it will only (over time) make the police response to an active shooter like the police response to a traffic stop, more militant than it needs to be.

2. Here one could criticize the lack of adequate training on crisis agency coordination and delegation in general and active shooter response specifically. But is such a criticism actually “fair”? Police do not have infinite resources; and the more complicated, high-competence, and low-probability the task, the more likely it is that training would be wasteful in general and incomplete in particular locales (regardless of however seriously any local police leader might take the problem). Uvalde was deficient in these regards in spite of years of supra-police investigating and report-making and office-creating with the goal of promoting partnerships and standardization in active shooter response. What good is knowing the best training practices if they are unworkable (and what good are the practices)?