What Do FBI Statistics Really Say About “Gunfights”?

Image from FBI.gov

Image from FBI.gov

Not as much as many people would lead us to believe. We frequently hear claims like: “Check out the stats, man, it’s always at 20 feet or less, and nearly always at 10 feet or less.” Or “that zero to three feet distance that most gunfights take place at in the real world.” Various figures are cited as the source for these contentions, most commonly “the FBI.”

Well, neither of those statements can be proven true based on the information supplied by “the FBI.” The available facts are actually a lot less conclusive than many people portray them as. Consequently, there is a great deal of unsupportable extrapolation of the facts that do exist.

The data that is being referred to is compiled and published as part of the FBI’s annual Law Enforcement Officers Killed & Assaulted report, which is available to download from the FBI’s website. It is published around the end of October each year for incidents occurring the previous year.

Since LEOKA 2012 has just been released, let’s examine the assertions about them. Unfortunately, the FBI does not gather information about distance of gunfights in routine submissions to the Uniform Crime Reporting (UCR) Program. The only information on distances is gathered as supplemental data in cases where police officers are actually killed. These data are compiled in Table 36 — Distance Between Victim Officer and Offender, 2003–2012, in the current report. However, this table reports only a very small subset (44) of “gunfights” police engaged in. The data that is collected about the total number of police gunfights (2,259) is compiled in Table 70 — Type of Weapon and Percent Injured, 2003–2012. No distances are provided in this table.

Image from FBI.gov

Image from FBI.gov

Even the data in Table 36 must be interpreted carefully, because the distance given is for where the officer was actually killed. For example, in a previous year, a State Trooper returning home was ambushed from across the street at a distance of 35 yards with a rifle and severely wounded. His assailant then walked over to him and executed the Trooper at close range with a handgun. Since the Trooper was killed at almost touching distance, he was reported as a zero to five feet casualty. But the “fight” actually took place at 35 yards.

Table 70 does list an interesting piece of information that provides a reality check on the “zero to three feet distance” assertion. Of the 2,259 Assaults with Firearms, only 9.8% resulted in injury. However, 13.2% of the officers who were assaulted with knives or other cutting instruments were injured. Statistically, it seems unlikely that within the same zone of influence, knives would result in a 35% higher casualty rate than firearms.

Image from FBI.gov

Image from FBI.gov

Over the past ten years, injury rates for firearms have been the lowest of all weapon types:

  • Personal weapons – 28.5%
  • Other dangerous weapons – 23.8%
  • Knives or other cutting instruments – 12.7%
  • Firearms – 9.3%

If most police gunfights took place within three feet, it would seem reasonable to expect that the injury rate would most likely approach 100% rather than 9.3%. Using these data to partially vet the distance of police gunfights would seem to indicate that police gunfights do not take place at near-contact distances. More likely, gunfights actually take place at considerably longer distances, where the popular spray-and-pray point-shooting techniques employed by criminals are rather ineffective.

Image from FBI.gov

Image from FBI.gov

Looking at the overview of officers killed in 2012 also provides some interesting information. Forty-four officers were killed with firearms, 32 with handguns. However, only seven officers fired their weapons, and two additional officers attempted to use their weapons. Twenty-four of the officers killed with firearms were within five feet from the offenders. Combining the data of officers who fired their weapons with the data of distance the officers were killed at would seem to indicate that officers tend to be slain in “shootings” rather than “gunfights.”

A more justifiable assertion based on the data in Tables 36 and 70 would be that if a police officer is shot (at) by a criminal at a distance of under five feet, the possibility of death is high. The data gathering and analytical capabilities of the FBI and UCR are probably not necessary to arrive at that conclusion. The only study that actually looked at distances of gunfights in a detailed and methodical way was done for Police Marksman magazine by Richard Fairburn in the 1990s. The average distance reported in that study was 14 yards.

Another frequent claim goes something like, “There’s a commonly held and cited statistic that three out of four times, when officers are killed during shootings, it happens in low light.” That “commonly cited statistic” is a fallacy.

Image from FBI.gov

Image from FBI.gov

The Summaries of Officers Feloniously Killed provides much more detailed information. The summaries from LEOKA 2011 yield a very specific time break-out:

  • Mean 12:44pm
  • Mode 11:00am
  • Median 1:20pm

Of the 64 incidents reported in the 2011 Summaries, only 36% took place between the hours of 2100 and 0700. Of those, in only one incident would a flashlight have had any relevance whatsoever, and even that incident is questionable. For instance, if an offender uses his car to run an officer down at 12:20am, a flashlight is unlikely to help the situation. This isn’t to say that flashlights aren’t useful. However, when using statistics as the underlying premise for an assertion, they should be researched for veracity, rather than be “commonly cited.”

become a member

The purpose of this article is not to say that gunfights and shootings don’t take place at close range but rather that stating “the FBI” says they do is incorrect. Statistics can be useful, but their data collection methods and interpretations must be well understood before the final product becomes useful intelligence. And the actual source for the statistic must be determined and vetted to be real and appropriate. If we’re going to use numbers as a basis for our training and decisions, let’s get them correct.

Reply to Allen Wilkins
  • (will not be published)

29 Responses to “What Do FBI Statistics Really Say About “Gunfights”?”

  1. Jo

    Data are good to have and lots of work to not compile it but deciphering can be challenging. But, overall the bottom line is random gunfights are hard to talk about as to best practice and all that under that “gunfight” umbrella. We are not operators on a specific task, the private Joe’s react to situations that presents itself.

  2. john estes

    Excellent article. Those myths have been going around for at least 30 years as I can remember.

  3. Patrick

    1. We carry firearms in order to kill people who attack us, protected by the very specific and limited laws of self defense.
    2. We want to hit what we shoot and to kill what we hit, and to be completely exonerated in a court of law.
    3. We understand that, depending on state, prosecutor, local politics, re-election and pressure, we are always in danger in these courts.
    4. We understand that even a charge of manslaughter carries a potential penalty of 10 years.
    5. We understand that, beyond a very close distance- say 20 feet- we begin to loose our ability to claim unavoidable danger.
    6. We accept that for every 5 feet after 20-25 feet, we increasingly risk our actions being labeled as criminal.
    7. We accept that after 40-50 feet, we may no longer be able to win a court case and could end up in prison regardless of the legitimacy of our action.
    8. We understand that training and skill are far more important than after-market custom builds and endless debates over caliber and manufacturer.
    9. We master the tools available to us, that we can afford, and become experts in as many firearms as possible, exposing ourselves to as much and as varied training as we are able, to be able to respond in all conditions.
    10. We learn to shoot, hit, and kill, any target within 50 feet, in any position, with either hand, light or dark, without hesitation or remorse.
    11. We pick weapons we can handle, that available to us, and we train rigorously with them as we are able, in order to perform within the bounds of self-defense.
    12. We do actual research to confirm the actual performance of the rounds we have chosen, and know exactly how they will perform with the firearms WE personally carry, within the physical boundaries of self-defense.

    • dustydog60

      We carry firearms in order to STOP people who attack us…

      Never say your goal was to unalive someone.

  4. John McSherry

    The pistol/revolver is the weapon of last resort (military). But in the civilian world, the pistol is the first choice. We all know that civilians are the wild-card in any community task force, where discipline is non-existent. IFF is a measure of shoot first, without regard to friendlies. I would rather see identification transponders issues to para-military civilian militias, than shooting wildly in all directions. Assassinations are becoming more common, which brings engagements down to Zero. If we quarantine the larger cities, anarchy will take care of itself. Here in Appalachia Ohio, DNR has earned us about ambush possibilities. Instead of a gun, some good camo.

  5. Robert

    Garbage article.
    Look at actual, documented gunfights, like the 1986 Miami FBI shootout, or any actual bank robbery or convienent store holdup. Pull up liquor store robbery videos on Youtube.
    <10 yards, almost without fail.
    7 yards has been the average for longer than the author has been alive, and it still is.

    • Claude

      No, that’s not what Dick said at all. “While the average puke has a problem landing handgun hits at 20 feet” is rather different from saying that’s where police gunfights take place.

  6. Ed Harris

    All of the above has no meaning to “civilians”. The average civilian gun fight will take place at 3 to 5 yards.Lets look at the reason. Most civilians will be in a gun fight with a robber or burglar. Very few robbers walk up to their victim gun in hand. And most residence’s hallways are 5 to 7 yards long. A room-3 to 4.

  7. Reggie

    It was once said that FBI statistics on shooting were 777, meaning 7rounds, 7seconds, 7yards. Now rumors has it that it’s changed to 333. Just trying to get some clarification. Thanks.

  8. Don Russell

    notice that ZERO mention is made of the scores of cops saved each year by the mandated wearing of concealed body armor? that’s one of the reasons that the % of deaths is so low with guns. Knives pierce armor, and slashes with a sharp knife, to the arms/legs can easily result in death.

  9. Don Russell

    most attacks are not made with guns, according to the Annual Crime Survey. You absolutely are justified to fire on a club or knife-wielder, or someone much larger/stronger than you, out to 10ft or so of distance. This fact tips the training to be more like 1 second and 10 ft, 2 shots or less. If your hand is not on your gun (say, in a pocket holster) or out in sight, the odds for you are not good, not at all, even if the attacker has no gun. Don’t give gun attacks too much priority, faster is better!

  10. Aaron

    FBI and LEO stats on gunfight distances are irrelevant to lawful self defense.

    Law enforcement has the mission to apprehend and stop criminals; therefore they will necessarily engage at distances further than required for self defense.

  11. Brian Jackson

    One large problem with any ‘pure statistics’ related report is that gun fights are dynamic incidents, as the example of the attack that started with a rifle, street shootings take place a different distances and the report that is filed to the FBI does not have variable distance options for data entry.

  12. Chris Upchurch

    Table 70 does list an interesting piece of information that provides a reality check on the “zero to three feet distance” assertion. Of the 2,259 Assaults with Firearms, only 9.8% resulted in injury. However, 13.2% of the officers who were assaulted with knives or other cutting instruments were injured. Statistically, it seems unlikely that within the same zone of influence, knives would result in a 35% higher casualty rate than firearms.

    It doesn’t seem at all unlikely to me. I’ve done quite a bit of force on force and a fair amount of empty hands versus knife attacks (not Tueller drills at 7 yards, stuff up close where real knife attacks happen). It’s a lot harder to prevail against the knife without getting touched.

    • victoria hawke

      most of my practice is done at 20 ft or less mostly at point blank to abt 15ft and lots of hip shooting and weak hand. i have had 4 encounters and they were 10 ft and less. practice what you want and remember the DA is not usually your friend

  13. dan

    Using police shootings as the only basis for practice is imho not valid for many of us civilians. many law enforcement shootings are in relation to traffic stops. frequently the officer is standing off usually at hi/her vehicle and calling the driver to exit when the shooting starts or many times retreating when he observes a gun.

    Civilians may be much closer to the assailant do to the assailant’s intent, ie mugging, assault, sexual attack etc.

    Although I do shoot at greater distances once in a while most of my practice is at 21 feet or less. The Tueler drill is set at 21 feet and shooting an assailant with a weapon other than a firearm beyond this distance may prove difficult to defend in court.

    An assailant with a firearm is a whole different situation, of course.

  14. Stan

    This is “Law enforcement data”, not Personal defense data and should not be interpreted as such. Law enforcement are usually called to the scene expecting danger. It seems to me that the most dnagerous situation to train for is still “Counter ambush” because that is where the fatalities lie.

  15. Larry Conrad

    Another short coming in using LE shootings is there is no data showing how LE fatal shooting stack up against civilian shootings, do they take place at simular distances? We do not know, then when you factor in drive by shootings the distances will be altered again. Anyone serious about self defense with a hand gun should practice out to at least 15 yards if not all the way out to 25.

    • Allen Wilkins

      the man who does our class’s here works off the 10 feet or less theory. his training is very rigid

  16. Ernie White

    If I read this right there is no mind set as to how for apart a gun fight is…
    It could be from in your face to 35 yards away!
    I shoot at any where from 7 ft. out to about 25 ft.
    Sometime even more. with my carry pistol.

    • EstevanCruz

      Solid idea to train for distance. I also train with that mentality, and I also try to train with alternating hands to bring my weak side up in effectiveness.

  17. W. Hock Hochheim

    The FBI stats are always very broad and well…shaky. They were much worse decades ago. Some states have done a much better job recording shootings details, but POLICE shootings only! These police things are abstract to citizen shootings. The situations are so very different. I recall the decades-old, police “Rules of 3.” 3 YARDS not feet -which does open it up from 0 to 9 feet. (I don’t know when the gossip/street-lore changed it from yards to feet?) More than 3 shots are fired. Okay. Usually lasts “about” 3 seconds – about, meaning many happen real fast. These are kind of somewhat harmless assumptions, and certainly NEVER to be taken as total gospel by us. But I could go on and on about the coding and collection for the FBI reports. Interesting to read, but not so good for building training doctrine.

    • Don M

      Police are civilians. They do, occasionally, know the law better than non-police, but that is also not universal.