A WEEK after the carnage in Paris, terrorists struck again, this time in an attack on a hotel in Bamako, the capital of Mali. There were no immediate indications that attacks in Mali and France were directly linked, but both incidents underscore the spread of a deadly style of attack that came to prominence in a jihadist assault on Mumbai seven years ago.

The first characteristic of this form of terrorism is that the main weapon used is guns, rather than bombs. This is partly because assault weapons are relatively easy to obtain. East Africa is awash with guns that have come down from Somalia. Large parts of the Sahel and West Africa have been flooded by weapons brought south across desert smuggling routes from Libya. And in parts of Western Europe attackers have had relatively little difficulty buying guns smuggled in from the Balkans.

Bomb-making, on the other hand, requires either access to explosives (whether commercial or military, both of which are tightly controlled) or the ability to make explosives from commonly available materials—a dangerous and difficult job. Moreover once explosives have been acquired or made, a skilled bomb-maker is needed to assemble them into a device that will reliably detonate. Both tasks are easier said than done. For instance, a series of four planned suicide bombings in London on July 21st 2005 failed because the bomb-maker who prepared the explosives got his mixture wrong.

Guns can be used after minimal training, and are frighteningly efficient at killing large numbers of unarmed civilians, particularly if they are in a confined place. When four jihadists from the Shabab, an affiliate of al-Qaeda in east Africa, attacked the Garissa University College in northeastern Kenya on April 2nd 2015 they killed 148 people. An attack by jihadists almost two years earlier on the Westgate shopping mall resulted in 67 deaths. Moreover they do not need large teams. A lone attacker shot and killed 38 people in Sousse, Tunisia, in June this year.

A second characteristic of this style of attack is that the perpetrators have no expectation of coming out alive. Hostage-takers in the 1970s generally expected to be released as part of any deal to free their prisoners. In the Munich Olympics crisis of 1972, for instance, Palestinian terrorists demanded the release of more than 200 prisoners (as well as safe passage out of Germany) before a botched rescue attempt resulted in the deaths of all of their Israeli hostages. The advent of suicidal attackers, on the other hand, means there is less scope for negotiation. Often the main reason such attackers have for taking hostages is to complicate efforts by security forces to regain control of the site, since hostages may be killed in the crossfire.

The advent of suicidal attackers armed with guns is forcing hostage-rescue teams to throw away their old playbooks. These were premised on the idea that rescue forces could bide their time while building up an intelligence picture of the attackers. In the Iranian Embassy siege in London in 1980, for example, the security forces negotiated over the course of six days before launching a 17-minute assault to rescue hostages.

Yet one key lesson of the attacks in Garissa and Mumbai, in which 166 people died, was that the longer such hostage-takers are left in control, the higher the resulting death toll. In Paris, French security forces could immediately call on elite hostage rescue teams who train intensively for just the sorts of rescue operations that they conducted. But in remoter parts of the world such as Mali, Kenya or Somalia, even the best units will not have the training, equipment or expertise of those in the rich world, and rescue efforts may be riskier and more chaotic.

In the Westgate attack a police counter-terrorist unit entered the mall from the roof while soldiers from the Kenyan army entered on the ground floor. Neither team was communicating with the other, according to a blow-by-blow account in Foreign Policy. When the two units bumped into one another, a fire-fight broke out in which the commander of the police team was killed. The policemen promptly withdrew, as did the soldiers.

In the most recent attack in Bamako, Malian commandos were assisted by French special forces when they stormed the hotel just a few hours after the hostages were taken. The presence of the French soldiers was fortuitous: they were there as part of an approximately 1,000 strong contingent that has been conducting counter-terrorism operations in Mali.

Although governments can change the way in which they respond to these sorts of attacks, they cannot prevent them entirely. For the past few decades much counter-terrorism work has focused on preventing bomb attacks by, for instance, changing the formulation of commercial fertilisers to make them more difficult to turn into improvised explosives, or by hardening targets. Most embassies, hotels and international companies in risky parts of the world already have barriers to prevent car bombs from coming too close and security films covering their windows, to reduce the number of casualties from flying glass if a blast does occur. But guarding against armed attackers requires additional defences such as armed guards, high walls and gates that can be locked quickly.

Policemen can’t be everywhere, all the time, but they can at least guard the most obvious magnets for terrorists. Kenya increasingly posts policemen armed with assault rifles at the entrances to shopping malls and international hotels. In the most recent attack on Bamako, the target, a Radisson Blu hotel, was a juicy one: it was frequented by diplomats, foreign businessmen and was used by international airlines including Air France and Turkish Airlines. Questions need to be asked as to why it was so poorly guarded. Even so, guards with guns outside hotels may do little more than displace attackers to softer targets.

The recent attacks suggest two responses. The first is the need to do more to tackle gun-smuggling. In Europe it should be possible, for instance, to prevent the legal sale of decommissioned assault rifles that can be converted back to lethal use. It would be naive to assume the supply could be totally dried up in conflict-ridden regions such as Mali or Somali, but tackling the trade in small arms might at least provide new lines of intelligence on the terrorists. Western spies find it difficult to penetrate jihadist groups such as al-Qaeda and Islamic State, but they ought to find it easier to recruit informers among the criminal networks that smuggle guns.

The second is that the rich world needs to do more to build up special forces and hostage-rescue capabilities across Africa and the Middle East. Much work is already happening quietly, but the growing prevalence of raid-style attacks by gunmen demands a proportionate increase in the forces needed to counter them. In the past Western governments whose citizens were being held may have had the luxury of time to fly out their own forces. For instance Germany freed 86 hostages on a hijacked plane in Somalia in 1977 and Israel freed 102 hostages a year earlier after covertly flying a rescue force into Entebbe, Uganda. Now governments may have to accept that an immediate rescue effort by local forces may save more lives than a counter-assault by better trained Western soldiers hours, or days, later. If so, it is worth making the effort to ensure that local forces in the most exposed places become as competent as possible.