Australia's Gun-Control Success Story Has a Statistical Problem
Wendover Productions tells a compelling story: Australia suffered a massacre, passed sweeping gun restrictions, confiscated hundreds of thousands of firearms, and then experienced a dramatic decline in mass shootings. The historical record is powerful. The problem is what the video does with the mathematics. It turns evidence that a change occurred into something resembling a 200,000-to-one probability that gun reform caused it. Those are two very different claims.
The distinction matters because the video builds its strongest argument around that number. And the original researchers themselves were considerably more careful about what their findings could establish.
The one-in-200,000 figure is not a probability of causation
The 2018 study by Simon Chapman, Michael Stewart, Philip Alpers and Michael Jones used a Poisson model to compare shooting rates before and after the reforms. It recorded 13 qualifying mass shootings between January 1979 and June 1996, and zero between July 1996 and February 2018. Its calculated p-value was approximately 0.0000047, with a bootstrap result of 0.0000069. The researchers also tested for temporal clustering rather than simply assuming that shootings were evenly distributed.
hoplofobia.info
That is meaningful statistical evidence that the rate changed. It is not a probability that the NFA caused the change. The American Statistical Association explicitly warns that a p-value does not measure the probability that a hypothesis is true or that random chance alone produced the observed result. It measures how unusual the data would be under a specified statistical model.
doi.org+1
Wendover nevertheless tells viewers that the probability of the absence being caused by natural fluctuations is one in 200,000, then announces a 199,999-in-200,000 probability that something other than chance explains the result, "something like the implementation" of the NFA.
That is an invalid conversion. The complement of a p-value is not the probability that the alternative explanation is correct. The model does not compare the NFA against every other possible explanation, assign probabilities to competing causes, or establish that the legislation deserves the entire credit.
There is an important qualification: Wendover did not invent this interpretation. Chapman and Alpers used similar language in their own 2018 response to criticism, describing the result as a one-in-200,000 chance occurrence and arguing that the reforms likely averted approximately 16 shootings. But their technical paper in Annals of Internal Medicine explicitly states that establishing a definitive causal connection without a randomized controlled trial is not possible. Wendover adopts the more confident interpretation while leaving out that crucial distinction.
opencart.addletonacademicpublishers.com
opencart.addletonacademicpublishers.com+1
The insurance analogy does not repair the problem. Insurers use probability models to price risks; Chapman's research tested whether the historical shooting rate changed. The existence of a profitable insurance industry cannot establish why an Australian shooting rate changed after 1996.
Nor were 16.3 shootings actually counted as prevented. That figure is the expected number under the hypothetical assumption that the pre-reform rate would have continued unchanged. It is a counterfactual projection, not an independently verified tally of attacks that the NFA stopped.
The video changes its definition of a mass shooting
Wendover initially describes 12 shootings with at least four victims killed before Port Arthur, compared with three shootings involving 12 victims over the following 26 years. Later, when presenting the Chapman model, it says that two shootings actually occurred between 1996 and 2018.
These figures are not presented using one consistent definition.
The Chapman model defines a mass shooting as an incident involving at least five victims killed, excluding the perpetrator. Under that definition, the study recorded zero incidents after the reforms through February 2018. The earlier JAMA study separately identified three post-reform fatal shooting incidents involving three or four victims. Those incidents did not qualify under its five-victim definition.
hoplofobia.info+1
The video never explains why its three-event count becomes two when discussing the Chapman model, or identifies the different threshold that would reconcile the figures.
There is no universally accepted definition of a mass shooting. The JAMA researchers themselves acknowledged this. That makes consistent definitions particularly important when comparing rare events. Changing the threshold without explaining it changes the evidence being presented, not merely the terminology.
Australia did not have the national firearms registry Wendover describes
One of the video's clearest factual problems concerns registration. Wendover says that after 1996, every firearm and firearm sale had to be recorded in a national registry.
The NFA did introduce nationwide registration requirements. It did not create one central national firearms database. Individual states and territories continued operating their own registries.
This was not a minor administrative detail. A 2023 Parliamentary Library report explained that a national register had been recommended in the 1996 agreement but had never been implemented. The Australian Government's November 2024 factsheet explicitly states that the planned National Firearms Register is not itself a national registry: each jurisdiction will continue maintaining its own records.
Parliament of Australia+1
The distinction became painfully practical during the 2022 Wieambilla shootings.
Nathaniel Train had a suspended firearms licence but was still able to purchase ammunition in Queensland using his physical licence. Dealers could not verify the validity of interstate licences through an effective national system. The Australian Government subsequently identified national licence verification as a vulnerability the new register was intended to address.
Document 14H+1
The national information-sharing project only commenced on 1 July 2024, following National Cabinet's agreement in December 2023. Its operational target is mid-2028.
In other words, something Wendover presents as an achievement of the 1996 reforms remained an outstanding national reform nearly three decades later.
Attorney-General's Department+1
Falling firearm deaths do not automatically mean the same number of lives were saved
Wendover moves from the decline in firearm mortality to the conclusion that the reforms saved lives. That is a reasonable question to investigate, but the video omits an important qualification from the very JAMA research it relies upon.
The 2016 study found that the annual decline in firearm deaths accelerated from approximately 3% before the reforms to 4.9% afterward. However, non-firearm suicide and homicide rates also changed substantially. Their combined annual trend shifted from a 2.1% increase before 1996 to a 1.4% decline afterward.
The researchers concluded that the decline in non-firearm deaths was greater in magnitude and that they could not determine whether the change in firearm deaths was attributable to the gun reforms. They also discussed improvements in emergency medical care and changes in the lethality of non-firearm methods as possible explanations for broader mortality trends. Wendover presents the firearm figures without giving this conclusion comparable attention.
JAMA Network+1
Later research provides a more complicated picture.
A 2020 synthetic-control study by Bradley Bartos and colleagues found that the buyback was associated with a significant reduction in Australia's overall homicide rate during 1997–2007. That supports the possibility of a wider public-safety benefit.
PubMed
A more detailed 2024 study by Patrick Duenow and Luke Connelly separated firearm deaths from deaths involving other methods. It estimated reductions equivalent to approximately 35 fewer firearm homicides and 77 fewer firearm suicides annually at Australia's current population.
But it also identified possible substitution. Other-method homicides increased, potentially offsetting the firearm homicide reduction; its aggregate homicide analysis found no net reduction during the study period. Evidence of substitution in suicide was less conclusive, particularly over the longer term.
Wiley Online Library+1
These findings do not establish that the NFA failed. They strengthen the case that it reduced firearm-specific deaths while demonstrating why those reductions cannot automatically be translated into an equivalent number of lives saved overall.
The distinction matters. A policy can make firearms less available and reduce deaths involving firearms without necessarily producing an identical reduction in total deaths. Wendover does not adequately investigate that difference.
The research affiliations deserved disclosure
The JAMA study disclosed that Simon Chapman had been a member of the Australian Coalition for Gun Control between 1993 and 1996, while Philip Alpers was director of GunPolicy.org.
Neither affiliation invalidates their research, and the disclosures were made by the authors themselves. But these are relevant relationships when a video relies heavily on their interpretation to establish the success of a politically contested policy. Wendover identifies their expertise without providing this context.
JAMA Network+1
Storyblocks is also disclosed as the video's commercial sponsor. Nothing in the supplied material establishes that the sponsorship influenced its gun-policy conclusions, so it would be irresponsible to suggest otherwise.
The word "stopped" needs a date stamp
There is also an important development since the video was published. On 14 December 2025, two gunmen attacked a Hanukkah gathering at Sydney's Bondi Beach, killing 15 victims and injuring 42 people. This does not erase Australia's decades-long reduction in mass shootings, but it demonstrates that the reforms reduced the danger rather than eliminating it.
ABC News+1
On 26 August 2026, National Cabinet agreed to another update to the National Firearms Agreement, including improved criminal-intelligence sharing, citizenship requirements for firearm licences and restrictions on 3D-printed firearms. Australia's firearm policy remains an evolving system, not a problem permanently solved in 1996.
Home Affairs
The verdict
Australia's 1996 reforms deserve serious credit. The disappearance of mass fatal shootings for more than two decades and the sustained decline in firearm-specific deaths are substantial outcomes.
But Wendover turns a statistical test into a probability of causation, muddles the definitions behind its shooting counts, and describes a national registry as an achievement when it remained unfinished for decades.
The strongest independent evidence supports the conclusion that Australia's reforms reduced firearm-related deaths. It does not establish that the NFA alone caused the entire change, that precisely 16 mass shootings were prevented, or that every reduction in firearm deaths represents a life saved.
The historical achievement is real. The mathematical certainty is manufactured by interpretation.
References
1. Wendover Productions — Australia Had a Mass-Shooting Problem. Here’s How it Stopped. Original video and primary source for the claims examined in this article.
2. Chapman et al. — Fatal Firearm Incidents Before and After Australia's 1996 National Firearms Agreement Banning Semiautomatic Rifles. Annals of Internal Medicine, 2018. The original Poisson model, its event counts, statistical results and explicit qualification about causal proof.
3. Wasserstein and Lazar — The ASA Statement on P-Values: Context, Process, and Purpose. The American Statistician, 2016. Establishes why a p-value cannot be interpreted as the probability that a hypothesis is true.
4. Chapman, Alpers and Jones — Association Between Gun Law Reforms and Intentional Firearm Deaths in Australia, 1979–2013. JAMA, 2016. Supplies the mortality trends, shooting definitions, alternative explanations and the researchers' own causal limitations.
5. Chapman and Alpers — Australia’s 1996 Gun Law Reforms Halted Mass Shootings for 22 Years: A Response to Criticism from Gary Kleck. Contemporary Readings in Law and Social Justice, 2018. Documents the researchers' public interpretation of the one-in-200,000 result and their disclosed affiliations.
6. Bartos et al. — Controlling Gun Violence: Assessing the Impact of Australia's Gun Buyback Program Using a Synthetic Control Group Experiment. Prevention Science, 2020. Provides an alternative synthetic-control assessment finding lower overall homicide rates after the buyback.
7. Duenow and Connelly — The Effect of Gun Buy-Back Law Reform on Homicides and Suicides in Australia. Health Economics, 2024. Separates firearm deaths from deaths by other methods and investigates substitution effects.
8. Parliamentary Library — Status of the National Firearms Register. Parliament of Australia, 2023. Establishes that the national register recommended in 1996 had never been implemented and documents the Wieambilla licensing gap.
9. Australian Department of Home Affairs — National Firearms Register Factsheet. November 2024. Explains that the new system connects existing registries rather than replacing them with a single national registry.
10. Australian Department of Home Affairs — National Firearms Register. Provides the implementation history and the mid-2028 operational target.
11. Department of Finance — FOI 24-25/091, Document 14H. Documents the suspended-licence ammunition purchase by Nathaniel Train and the proposed national verification solution.
12. ABC News — Nathaniel Train purchased ammunition despite suspended gun licence. 18 February 2023. Reports the practical consequences of Australia's interstate firearms-licensing gaps.
13. ABC News — Bondi Beach shooting death toll rises as police scour crime scene. 15 December 2025. Establishes the subsequent mass shooting's death toll and injuries.
14. Australian Department of Home Affairs — Firearms. Updated 26 August 2026. Documents the latest National Firearms Agreement and its additional national measures.