TTAG Charlie Hebdo Simulation: Preliminary Results

I don’t have all the data from the event yet — all the videos are still en route, for example — but Robert wanted me to post something about the event immediately, so I obliged. The office shooting simulation we ran in Plano, Texas involved two scenarios, and was repeated for a total of 9 repetitions. There were some problems with the event, but we also had some interesting findings and noted some areas to investigate further with a larger test. Here’s a quick look at the methodology, issues, and initial results…


Keep in mind that this isn’t the final word yet. Robert’s desire was to re-create the Charlie Hebdo environment exactly, but with an added armed defender. Given the extremely short time frame in which we had to throw this thing together (24 hours), Patriot Protection in Plano, Texas did not have sufficient time to re-create the offices exactly and instead re-used their existing setup which was adequate for the test. Their configuration in the training area consisted of a hallway with six small rooms on either side and a larger room at the end of the hallway.

Volunteers were collected from the local area. A total of 26 volunteers and some paid staff participated in the event, although not all of them were able to go through the scenarios.

The rooms in the training area were populated with volunteers, one or two people to a room down the sides and four people in the room at the end. For scenario #1, the armed defender was positioned in the room at the end of the hall. For scenario #2, the defender was positioned in one of the rooms on the side of the hall.

Each defender was told that there were people in the adjacent rooms, and to act as they would normally. They were given no specific instructions about how to react. The defenders were armed with a GLOCK 17 converted for UTM man marker rounds holstered in a non-retention holster and a single 18 round magazine.

The two attackers were armed with AR-15 rifles converted to fire UTM man marker rounds and a few 30 round magazines. The attackers in this instance were highly trained individuals who were employed as instructors at Patriot Protection, and were similarly instructed to attack the office area as they normally would given their training.

Each scenario started with placing the defenders and volunteers in the office area. Once they were set up, we had them chat among themselves to focus their attention on something other than the scenario at hand. After a few minutes, the attackers entered the area and began their assault. In every situation we modeled the defender had advanced warning that attackers were coming as there were shots fired before they got to the room where the defender was located.


Every live fire exercise has issues, and this one was no different.

The biggest problem reported by the volunteers was that their masks were constantly fogging up. Given the cold nights in north Texas and the lack of heating in the training area, the required safety equipment became more of a burden than expected. Many volunteers complained that they were unable to see the attackers at all when they finally entered the room, and were forced to simply shoot in their general direction. Obviously in a real world situation fogged up face masks would not be an issue, so this is a problem that we experience trying to re-create the scenarios only and detracts from the applicability of the testing to real world scenarios.

Another problem was the use of a flashlight by one of the attackers. A flashlight was not included in the testing protocol, but according to reports from the volunteers one of the attackers used a flashlight to blind the defender in one or two scenarios. These tests deviated from the methodology and will be excluded from the final results.

The use of UTM’s man marker rounds was another problem. While they give a good indication of whether someone is hit or not, they do not accurately model the ability to incapacitate or kill someone. People who were hit multiple times continued shooting, despite being instructed to simply stop shooting and get down once hit. This posed a huge problem when trying to analyze the results.

The final problem was time. Robert insisted that we perform this test within 24 hours of having the idea in order to capture media attention, and that unfortunately meant that the methodology was rushed, the volunteer pool was small, and we did not have sufficient time on the day of the event to even run all of the volunteers who did manage to make it to the location. Due to the small sample size, it will unfortunately be impossible to draw any hard and fast conclusions from the testing. However, we can identify some areas for further investigation in a larger scale test with more time allotted.

Preliminary Results

I do not have all of the data yet. Patriot Protection recorded the whole event on their camera system for me, and unfortunately they have been unable to provide me with the full tapes as of yet. However, the one thing I do have is footage from a camera mounted to one of the attacker’s rifles. Using this footage and descriptions of what happened from the volunteers and attackers, combined with what we have learned from previous testing in similar situations, I am able to make some preliminary statements about the results of the testing.

1. Attacker’s experience matters

In the previous simulation, we used volunteers as attackers to simulate a relatively poorly trained individual attempting to attack a school. In that scenario, nearly ever one of the armed defenders were able to either kill the attacker or successfully prevent them from entering the classroom and killing more students. In other words, guns would indeed save lives in that scenario. The difference in this scenario is that we used two extremely well trained individuals as attackers, and as a result the defenders were placed in what might be described as a “no win scenario.” Two heavily armed individuals against one person with 18 rounds is definitely not a fair fight, especially when the attackers do this as a job every day and are highly trained.

2. One terrorist = victory, two terrorists = defeat

Despite that “no win scenario” description, there were some notable minor victories for the defenders. From the footage available to me at this time, it appears that in the majority of the scenarios the defender is able to successfully kill one of the terrorists before they themselves are killed (the definition of “killed” is a hit on the attacker in a vital location before the attacker can fire back). While it might seem like another “no win” scenario, the fact remains that a single armed defender — even one with very little training — is able to successfully kill and stop at least one terrorist. The problem in this scenario is that there are two terrorists as attackers, and a single armed defender does not appear to be able to stop both of them.

Even when confronted with heavily armed and well trained attackers, a single armed defender is capable of stopping at least one of the attackers.

The question this raises is whether there is any deterrent effect that an armed defender provides to the remaining attacker. In our scenario our attackers were instructed to keep shooting, but in the real world it is entirely possible that the remaining attacker(s) would reconsider their course of action in the face of armed resistance. That outcome is something seen in American shootings time and again, that once confronted with armed resistance (usually the police) the attacker either flees or takes their own life.

3. Armed defenders give others time to escape

In one of the early scenarios, a relatively new shooter decided that instead of trying to confront the armed terrorists she would use her gun to cover her retreat and give her co-workers time to escape. This plan worked perfectly, and she was able to escape from the room while returning fire towards the attackers, allowing nearly everyone in the room to escape before she too turned tail and ran.

In the face of overwhelming numbers and firepower, it appears that this tactic of using the firearm as a means to give everyone else time to escape is extremely effective. There was only one person who used this tactic, but they used it to great effect. Due to the extremely small sample size it is unknown whether this tactic would work for others as well, but it is something that we need to investigate further and try to model in more detailed studies down the road.

Even when the defenders were unable to escape, the presence of an armed defender definitely slowed down the progress of the attackers. The comparison of how long it takes the shooters to kill people and move on between a room with no defender and a room with a defender is clear, and indicates that an armed defender at the very least slows down the attackers. In the case of Charly Hebdo, it would seem to indicate that an armed defender would have been able to give their other coworkers a better chance at running and escaping.

4. People barricade themselves rather than confronting enemies

In every scenario, the defender was instructed that there were other people in the offices around them. For seven of the nine scenarios, the defenders chose to hide in a corner and wait for the attackers rather than confronting them in the hallway and attempting to save their coworkers. For two of the scenarios, however, the defenders edged their way into the hallway and attempted to confront the attackers before they reached the room. Exactly what percentage of people would react in this manner under similar circumstances is unknown due to the small sample size, but in our tests it was approximately 22% of people. This is something that definitely merits further investigation on a larger scale.

Moving Forward

Within the next couple days I will get the footage from Patriot Protection, analyze the tapes, and come to a final conclusion about the results of the experiment. However, even with those final results there’s no way to say that they are applicable in any manner to real life. The sample size for this test was far too small and the methodology far too rushed to draw any definite conclusions one way or another, but it does give us some interesting indications of where to look in a deeper and more complete test. For example, comparing performance with a single armed attacker versus two attackers given identical situations? Or what if the defender was trained simply to delay the attackers and give others time to escape, is that an effective strategy? Could armed guards be effective? None of these questions were specifically tested, and they definitely merit further investigation.

I know the gun control activists will want to point to this as definitive proof that armed defenders are useless, but that’s simply not the case. Not only is there insufficient data to draw any conclusions whatsoever, but we’ve previously proven that in some cases armed defenders are nearly 100% effective at stopping similarly armed attackers. What we are investigating now is the probability of success in different scenarios, and by playing with the variables we can get a better understanding of how best to protect ourselves and our loved ones.

Stay tuned.