One week, three countries, three very different security pictures

What moved

Myanmar recorded 144 events and 81 fatalities during 18 to 24 July. Across the preceding 12 complete weeks, covering 25 April to 17 July, it had averaged roughly 260 events and 220 fatalities per week. The latest total was therefore about 45% lower for events and 63% lower for fatalities. Both were the lowest figures recorded anywhere in the full 13-week review period.

Part of the difference came from the absence of another catastrophic week like the one seen in late May, when more than 600 fatalities were recorded. That extreme week pulled Myanmar's recent average upwards, but it does not explain the entire fall. Even compared with the more typical middle of the preceding 12-week range, 81 fatalities was still markedly lower. Myanmar genuinely experienced a less deadly week.

The important question is what that decline meant. It did not result from a ceasefire, negotiated settlement or nationwide military withdrawal. The week simply contained fewer exceptionally lethal incidents while violence continued across many fronts.

Armed clashes remained the deadliest category, causing 43 recorded fatalities. These clashes included ground fighting between the military and the different resistance or ethnic armed forces operating across the country. Direct attacks against civilians caused another 15 deaths. Only three shelling or artillery events were recorded, but they caused 14 fatalities. A small number of attacks therefore produced a disproportionately large human cost.

Airstrikes and drone attacks also remained frequent. Forty such events were recorded across ten parts of the country, although they caused fewer reported fatalities than armed clashes during the week.

These categories are not separate pieces of random violence. They reflect how Myanmar's civil war is being fought. The military has struggled to regain territory from resistance and ethnic armed forces using ground troops alone. Aircraft, drones and artillery allow it to strike contested or opposition-controlled areas without first occupying them.

A village may see no conventional battle and still be bombed. A resistance group may hold a town but be unable to protect it from the air. A road may appear open on a national map while artillery fire or drone surveillance makes civilian movement dangerous. This is why fewer ground battles do not automatically mean less exposure for civilians.

The national decline also hid very different regional pictures. Mandalay Region recorded the week's highest fatality count, with 21 deaths. Kayin State recorded 15. Magway Region, Tanintharyi Region and Kayah State recorded eight each, while Rakhine State and Chin State recorded six each. Sagaing Region followed a different pattern. It recorded 37 events, the highest number anywhere in the country, but only two fatalities.

Sagaing was therefore the busiest region in the data, but not the deadliest. Other regions experienced fewer incidents that caused much greater loss of life.

This is partly because Myanmar's battlefields are not all alike. Sagaing, Magway and Mandalay form part of the country's central dry zone, where many locally organised People's Defence Forces fight the military. Rakhine State has a different conflict. There, the powerful Arakan Army controls most townships and has established administration across much of the state, while the military retains key positions and continues attacking Arakan Army-held territory from the air. Kayin and Kayah states have their own ethnic armed organisations, local resistance forces, contested roads and long histories of conflict with the central military.

A national total compresses all of these fronts into one number. The number tells us that violence fell. It does not tell us that every region became safer for the same reason.

Public reporting helps explain what some of the weekly figures looked like on the ground. During 18 and 19 July, airstrikes and drone attacks reportedly killed at least six civilians and injured more than 26 people across Rakhine State, Mandalay Region and Sagaing Region. Those injured reportedly included children and Buddhist monks. More than 25 homes and a church were damaged.

In Rakhine, strikes hit areas controlled by the Arakan Army. The group had captured 14 of the state's 17 townships, as well as neighbouring Paletwa in Chin State, according to the reporting reviewed. The military still held important positions around Sittwe and Kyaukphyu and continued relying heavily on air power against Arakan Army-controlled areas.

In Mandalay, a 70-year-old man was reportedly killed during drone strikes after telecommunications and internet access were cut. In Sagaing, four fighter jets reportedly dropped more than 60 bombs on two villages. No deaths were reported from that particular bombardment, but the scale of the attack illustrates why event totals and fatality totals need to be read separately. A major bombing can destroy homes, terrify communities and force people to flee without immediately producing a large recorded death toll. The absence of deaths does not make sixty bombs a quiet incident.

A broader ACLED assessment published after the source week found that the military had intensified aerial bombing and repression against civilians during the first half of 2026. ACLED described the more regular use of specialised formations of two to five jets conducting sustained attacks against individual targets. That matters when interpreting a lower weekly event count. One recorded airstrike event may involve several aircraft and repeated bombing runs. Fewer event records do not necessarily mean a proportionate reduction in firepower, destruction or fear.

Ground fighting also continued along the road and railway corridor linking Mandalay with Myitkyina, the capital of Kachin State. Military forces had occupied Maezar and Nantthae villages in Sagaing's Indaw Township in early July. After roughly two weeks, they reportedly withdrew as fighting intensified around Nabar, a road and rail junction connecting Sagaing with Kachin State.

Resistance sources believed the troops were being moved towards more important positions along the corridor. Fighting continued around Indaw, Katha, Mawlu and Nansiaung, with further clashes reported across the boundary with Kachin State. To an outside reader, this may sound like another sequence of unfamiliar village names. The route matters because it is one of the main land connections between central Myanmar and the north. It carries civilian travellers and commercial supplies. It can also move troops, weapons and military logistics towards Kachin State.

Control of one junction can affect a much larger area. If the military secures the route, it can reinforce northern positions and support operations elsewhere. If resistance forces disrupt it, they may isolate military units, slow resupply and make it harder for the state to retake nearby towns. Civilians experience the same contest differently: a battle over a railway station may mean transport stops, a road closure may interrupt food deliveries, and a military column may trigger village evacuation.

The humanitarian crisis did not fall with the weekly numbers. By 13 July, the United Nations estimated that more than 16 million people across Myanmar needed humanitarian assistance and nearly 3.8 million were displaced. Many had been forced to move more than once. Recent fighting in Magway Region alone had displaced more than 100,000 people. Armed conflict, movement restrictions and administrative requirements continued to prevent aid organisations from reaching some communities.

These figures were already in place before the latest week began. A lower weekly event total does not return people to their homes. It does not reopen a road, rebuild a clinic or restore a burned harvest. The humanitarian burden therefore continues after the battle count falls.

Why it matters

Myanmar's figures fell sharply, but the wider civil war did not meaningfully retreat. The latest week appears to have avoided another large concentration of deaths like the late-May spike. Instead, violence remained spread across many regions and methods: ground clashes, civilian attacks, artillery, airstrikes and drones.

Put plainly, there were fewer exceptionally deadly incidents. There was not an end to the war.

The way the conflict is being fought matters as much as the total number of events. Air and drone attacks allow the military to strike areas where it has limited ground control. A reduction in conventional clashes may therefore coincide with continued bombing of villages, towns and resistance-held territory.

The conflict can change shape without becoming less severe. A battle may end in one region while aircraft strike another. Troops may withdraw from a village only to reinforce a road junction nearby. Fatalities may fall nationally while people living beside one corridor remain in immediate danger.

That is why the latest decline should be treated as real but limited. It tells us Myanmar had a less deadly week. It does not tell us the forces sustaining the civil war have weakened, reconciled or stopped attacking civilians.

The humanitarian situation makes this distinction even more important. Displaced families do not return because a chart moves down. People need to believe that fighting will not resume, roads will remain open and homes will not be shelled or bombed after they return. Without those conditions, a lower week may offer temporary relief without producing lasting recovery.

What to watch next

The clearest warning sign would be another sudden concentration of deaths. Myanmar does not need hundreds of additional events for the picture to worsen quickly. One major airstrike, artillery attack or battle could push fatalities sharply upwards, especially if it hits a village, displacement site, school, monastery or crowded public building.

The central dry zone deserves close attention, particularly Mandalay, Magway and Sagaing. These regions combine extensive resistance activity, military operations, contested transport routes and large civilian populations. If ground operations intensify around roads and railways, the effects could spread beyond the immediate fighting through displacement, disrupted trade and blocked humanitarian access.

The Mandalay-Myitkyina corridor is particularly important. Further military reinforcements towards Indaw, Katha, Mawlu or Nansiaung would suggest the route remains a priority. Resistance attempts to block or retake key junctions could lead to more concentrated fighting and greater use of artillery, drones or aircraft.

Air activity will be another essential clue. If ground clashes remain below recent levels but airstrikes stay frequent, Myanmar may look calmer in the national totals while selected communities remain highly exposed. That would be a change in how the war is being fought, not a meaningful improvement in safety.

Rakhine State also bears watching. The Arakan Army controls most of the state, while the military retains strategic positions and air power. Repeated strikes against Arakan Army-administered towns could deepen displacement and increase pressure around Sittwe, Kyaukphyu, transport routes and major infrastructure.

A further question is whether air attacks are increasingly aimed at civilian life around opposition-controlled territory rather than only armed positions. Damage to churches, homes, schools or markets can make an area unliveable even if military control does not immediately change.

A more convincing improvement would require several developments moving together: civilian deaths continue falling; airstrikes and shelling become less frequent; fewer people are newly displaced; roads and transport corridors remain open; aid organisations gain safer and more reliable access; and the decline persists across several weeks rather than one.

Until those signs appear together, the clearest reading is:

What moved

Of Indonesia's 102 recorded events, 72 were peaceful demonstrations. None caused a recorded death. Another 21 events were classified as riots, mostly involving forms of mob or communal violence, and accounted for four deaths. Three direct attacks against civilians caused another four deaths, while three armed clashes caused two.

The distinction matters. Someone looking only at the national event count might imagine Indonesia was experiencing widespread violent disorder. Someone looking only at the 10 fatalities might assume demonstrators were being killed during confrontations with police. Neither picture fits the evidence. Most of the week's visible activity involved people gathering, marching or demonstrating without fatal violence. The deaths came from a much smaller number of incidents involving armed groups, attacks on local residents and communal conflict.

The demonstration total also did not represent 72 branches of one centrally organised national campaign. Public reporting from inside the week shows several separate protest gatherings even within Jakarta. On 20 July, police prepared for a demonstration by student executives from Indonesia's state Islamic universities and other groups in the Gambir area. A second protest was planned outside the Financial Services Authority by a different youth activist network. More than 700 police personnel were deployed across the area.

On 24 July, police prepared for four further demonstrations in central Jakarta. Different groups planned to gather near government, legal, education and financial institutions. Their organisers included legal-reform activists, student and youth groups, and a coalition of university organisations.

These examples do not independently account for all 72 demonstrations recorded by ACLED. They do show why the national protest layer should not be described as one movement with one demand. Some demonstrations were directed at government policy. Others focused on education, law, regulation, local administration or a specific institution. Several may have involved only a modest number of people. They entered the same national total because they shared a method, public demonstration, not because they shared one political command centre.

The protests were fragmented, but they were not occurring in a political vacuum. On 17 July, one day before the ACLED reporting window began, students and civil-society groups demonstrated in central Jakarta under a broad "motion of no confidence" theme.

The protesters called for greater civilian sovereignty, changes to the state system and protection of education. They also criticised major government programmes, including the free-meals scheme and the Red and White Village Cooperatives, arguing that implementation had not met public expectations. Police redirected the gathering away from its intended location near the Horse Statue and towards another part of the National Monument area.

The timing matters. That demonstration cannot be counted as one of the 72 peaceful protests between 18 and 24 July because it happened the day before the source window opened. It is immediate political context, not direct evidence for every protest inside the week. What it reveals is the language through which some student groups were beginning to connect different frustrations.

Complaints about education, government spending, civilian authority, public accountability and the role of the state were being placed inside a wider argument about the direction of President Prabowo Subianto's administration. Similar grievances had appeared in larger student demonstrations in June, when students criticised fuel-price increases, government spending priorities and the growing role of the military in civilian affairs. Some argued that expensive flagship schemes were being prioritised over education, healthcare and immediate cost-of-living pressures.

That does not prove Indonesia was heading towards one unified national protest movement. It does suggest that several ingredients for broader mobilisation were already present: economic pressure, student networks, distrust of major spending programmes, concern over civilian institutions and frustration with how demonstrations were managed. For now, those ingredients remained distributed across different groups and cities rather than fused into one national campaign.

The week's geography makes the split even clearer. Jakarta recorded 19 events, East Java 11, West Java 10 and Central Java seven. None of these four major population centres recorded a fatality in the weekly aggregate. Highland Papua recorded only six events, but those incidents caused six deaths. East Nusa Tenggara recorded three fatalities, while Banten recorded one.

Across the full 13-week review period, Highland Papua, Central Papua and Papua together accounted for almost three-quarters of Indonesia's fatalities, despite contributing a much smaller share of the country's total event volume. Jakarta and Java therefore produced much of the visible activity. Papua carried much of the recurring fatal burden.

Papua has faced a decades-long conflict between Indonesian security forces and armed groups seeking independence. The dispute traces back to the former Dutch colony's contested incorporation into Indonesia, formalised through a heavily disputed 1969 process. Since then, cycles of separatist resistance, military operations, political exclusion and allegations of serious abuses have kept the conflict alive.

The current armed movement is commonly associated with the West Papua National Liberation Army, often referred to by the Indonesian initials TPNPB, and the wider Free Papua Movement. It does not operate like a conventional army holding one uninterrupted frontline. Violence is spread across remote mountain districts, roads, airstrips, mining areas and isolated communities. Armed groups may control or influence particular routes without administering large continuous areas. Security forces may hold towns and facilities while struggling to control the surrounding terrain.

During the source window, the West Papua National Liberation Army claimed responsibility for killing three Indigenous Papuans near the Trans-Papua Highway in Yahukimo. The armed group said the victims were intelligence agents assisting Indonesian forces and released material depicting the killings. Indonesian authorities rejected that allegation and described the victims as civilians.

Red Cloud Watch cannot independently determine whether the victims had any connection to the security forces. What can be said is that three local residents were deliberately killed, the armed group accepted responsibility, and accusations of collaboration were used to justify their deaths.

That creates a particularly dangerous environment for civilians. People living in remote districts may use government-built roads, travel on aircraft licensed by the Indonesian state, visit public clinics or interact with security personnel because those are the only services available. An armed group may interpret that contact as proof of cooperation. At the same time, Indonesian security forces may view people living in rebel-influenced districts with suspicion. Local residents can therefore find themselves trapped between two demands for loyalty, neither of which leaves much room for ordinary survival.

The killings took place near the Trans-Papua Highway, a road system intended to connect isolated communities across difficult terrain. In Papua, roads and airstrips are not merely infrastructure projects. They determine whether food, fuel, medicine, teachers, health workers and passengers can reach remote communities. They also carry security forces and government personnel. That makes civilian transport vulnerable to being interpreted as military logistics.

Earlier in July, armed separatists killed an American pilot shortly after his aircraft landed at a remote airstrip in Yahukimo and then burned the plane. The armed group claimed the aircraft had violated a flight restriction and alleged that civilian flights were being used to transport Indonesian troops. Indonesian authorities denied that allegation and said the aircraft served isolated communities. Seven Indigenous Papuan passengers were spared.

That incident occurred outside the 18 to 24 July source window, but it helps explain the week's later killings. When armed groups begin treating pilots, roads, local passengers or transport providers as possible military collaborators, civilian infrastructure becomes part of the conflict. The most important consequence may not appear immediately in a fatality total. It may appear when pilots refuse particular routes, operators reduce services, residents avoid highways, food becomes more expensive or medical access deteriorates.

The three deaths in East Nusa Tenggara came from an entirely separate story. On 18 July, groups from Narasaosina and Waiburak villages clashed in East Adonara. Police said the violence grew from a fight between groups of young men and was connected to a continuing dispute over customary land.

Three people were killed, seven were injured and 20 homes were burned. Additional mobile-brigade police were deployed, including 34 officers who patrolled the area and staffed a position between the two communities. Authorities said normal activity and religious services had resumed, but police remained in place to prevent another confrontation.

This was not part of Papua's separatist conflict. It was not an extension of the student protests in Jakarta. It was a local dispute that became communal violence. Calling it local does not make it minor. Three deaths, injured residents and burned homes leave behind grief, displacement and an argument over who suffered what. Police can separate the communities and reopen roads. They cannot erase the memories that may drive retaliation.

Available public reporting helps explain six of Indonesia's 10 recorded fatalities: three in East Adonara and three in Highland Papua. It does not independently account for every fatality recorded in Highland Papua or the death recorded in Banten. Those totals should remain grounded in the ACLED aggregate rather than decorated with stories the available reporting does not support.

Why it matters

Indonesia shows how a national number can be accurate and still give the wrong impression. The loudest part of the week was peaceful protest. The deadliest part occurred in smaller, more remote settings and arose from completely different causes.

That difference matters for anyone trying to understand political or operational risk. A national institution focused only on event volume might concentrate on Jakarta and Java. An organisation responsible for aviation, mining, construction, healthcare or staff movement may find Papua much more consequential despite its lower number of recorded events.

The required responses are different too. Student protests call for political listening, proportionate policing and credible engagement with demands. Papua requires civilian protection, access to isolated communities, restraints on military escalation and some form of political process capable of addressing a conflict that policing alone has not resolved. East Adonara requires mediation, investigation and measures to prevent retaliatory violence between two neighbouring communities.

Treating all three as one category of "instability" would obscure the different mechanisms driving each.

The Papua killings raise the deepest concern. If accusations of collaboration become a recurring justification for killing local residents, armed groups can exercise control through fear even where they do not hold territory. People may avoid roads, aircraft, clinics or government services not because those facilities are physically closed, but because using them might attract suspicion. Fear can therefore isolate communities more effectively than a roadblock.

There is also a possible connection between Papua and the wider protest landscape. At present, they remain largely separate. Papua's conflict is geographically distant from most of the peaceful demonstrations that dominated the weekly event count. But a major civilian casualty incident, widely shared video, serious allegation of military abuse or further attack on civilian aviation could bring Papua more directly into student and civil-society mobilisation in Jakarta and other cities.

That would not mean peaceful protesters had joined the separatist movement. It would mean a remote conflict had become a national political grievance.

What to watch next

The first question is whether Indonesia's separate layers remain separate. For the protest layer, watch whether different groups begin returning under the same demands rather than organising around dozens of local issues. The numbers become more politically significant if student organisations, workers, professional groups and civil-society coalitions repeatedly mobilise around a shared argument about government spending, education, civilian authority or the role of the military.

The sharper trigger would be a serious policing incident. Blocking a march or moving protesters away from a symbolic location may produce complaints. A widely recorded injury, death or apparently excessive use of force can give fragmented groups a common symbol and a reason to organise together.

That is a plausible pathway, not a forecast. Indonesia has a large and active civil society. It also has strong historical memories of student mobilisation and state repression. Officials do not need to turn every demonstration into a national crisis. A badly handled confrontation can perform that service for them.

In Papua, watch the places where conflict meets daily survival: civilian airstrips and small aircraft, the Trans-Papua Highway, mining and construction sites, clinics and schools, local government workers, and Indigenous residents accused of assisting either side.

The most worrying pattern would be a repeating sequence: an armed-group attack; a larger security operation; new allegations of civilian harm or displacement; retaliation against residents or transport providers; and further military deployment and armed clashes. Papua could become more dangerous without producing a dramatic increase in national event volume. A small number of attacks against aircraft, roads or local residents may be enough to isolate entire communities.

Also watch whether service providers begin changing behaviour. If pilots refuse routes, contractors suspend work, transport operators withdraw or clinics lose regular supplies, the conflict's indirect effects may spread much farther than the locations of the attacks.

East Adonara should remain a separate watch item. Police restored immediate order, but three deaths and 20 burned homes create the ingredients for revenge. Further threats, organised weapon gathering, continuing displacement or another confrontation between the same communities would suggest that the dispute was not contained.

A less concerning picture would require several developments to hold together: demonstrations remain peaceful, police avoid a catalytic confrontation, fatalities in Papua decline, civilian transport continues to operate, and there is no retaliation in East Adonara.

Until then, the clearest reading is:

What moved

Pakistan recorded 267 events and 146 fatalities during the week of 18 to 24 July. Across the preceding 12 weeks, it had averaged roughly 265 events and 153 fatalities per week. The latest week was also close to the previous four-week pattern.

Statistically, very little changed. That was the troubling part. Across the full 13-week period, Pakistan recorded between 76 and 292 fatalities per week. The country was not moving through a quiet period occasionally interrupted by violence. It was sustaining a high and volatile level of conflict in which even an average week carried a severe death toll.

Of Pakistan's 267 recorded events, 171 were protests, almost all peaceful. None caused a recorded death. Sindh accounted for 147 events, more than half the national total, but only one fatality. Punjab recorded another 21 events and no deaths.

This matters because a national chart would make Sindh look like the centre of Pakistan's security story. By event volume, it was. By human cost, it was not. The protests represented a broad public-order layer: visible, politically relevant and potentially disruptive, but largely separate from the armed conflicts driving the week's deaths.

Peaceful demonstrations can still reveal frustration over governance, policing, livelihoods and public services. But they should not be blended with militant violence as though a street rally in Sindh and a suicide assault in Khyber Pakhtunkhwa were expressions of the same national escalation. They were not.

Armed clashes accounted for only 29 recorded events, but caused 112 fatalities, more than three-quarters of Pakistan's national death toll. Direct attacks against civilians caused another 16 deaths. Roadside bombs and other remotely detonated explosives caused nine.

The geographic concentration was stark. Balochistan recorded 61 events and 79 fatalities. Khyber Pakhtunkhwa recorded 37 events and 66 fatalities. The rest of Pakistan combined recorded one fatality. In plain English, Sindh made Pakistan look busy. Balochistan and Khyber Pakhtunkhwa made it deadly.

Khyber Pakhtunkhwa has faced militant violence for years, especially in districts near Afghanistan and the former tribal areas. The TTP emerged in 2007 from a collection of Islamist militant factions opposing the Pakistani state. It seeks stricter Islamic rule and has attacked police, soldiers, government facilities and civilians. Pakistani military campaigns later pushed it from much of its territorial base, but the movement rebuilt parts of its network after the Afghan Taliban returned to power in 2021.

As the source window began, Pakistani forces were already carrying out operations across Bannu, Lower South Waziristan and Dera Ismail Khan. On 19 July, authorities reported killing three suspected militants during a raid on a hideout in Bannu. Police also detained 28 suspects as operations expanded across several nearby areas. The campaign followed an attempted vehicle bombing against a police station on 15 July.

In Lower South Waziristan, security sources said they destroyed an explosives-laden vehicle and motorcycle allegedly being prepared for a suicide attack. In Dera Ismail Khan, counter-terrorism police reported killing a suspected TTP member accused of involvement in an earlier officer's death.

These were claims made by Pakistani security agencies. Red Cloud Watch cannot independently confirm every identity, weapon recovered or account of how each person died. What the reports do show is that authorities believed they were dealing with connected militant infrastructure, not one isolated attacker.

Then came the week's largest publicly reported attack. During the night of 23 to 24 July, a suicide bomber and gunmen attacked a joint security checkpoint in Tank district. According to the Pakistani military, the attackers first tried to storm the position. When they were pushed back, an explosives-laden vehicle was driven into the perimeter wall, causing severe damage.

The assault killed 15 people: 12 soldiers, two police officers and a former forestry official. The military said 12 attackers were killed and blamed the TTP, which did not immediately comment.

This was not a quick ambush against an exposed patrol. It was a coordinated attempt to overwhelm a defended position using gunmen and a vehicle bomb. Such an operation requires explosives, personnel, transport, knowledge of the target and enough organisation to bring those parts together.

The attackers did not capture the checkpoint. But describing the attack only as "repelled" would hide the cost. Fifteen people were killed. The position was badly damaged. Further clearance operations followed. Other checkpoints across the region now had to consider whether they could face the same combination of gunfire and explosives.

Balochistan's conflict has a different history. It is Pakistan's largest province by land area, rich in gas and minerals but comparatively poor and sparsely populated. It is also home to the port of Gwadar and major projects linked to the China-Pakistan Economic Corridor.

Baloch separatists argue that the province's resources have been extracted without fair political power, local consent or an equal share of the benefits. The state says it is bringing development while fighting armed groups that attack security forces, civilians and infrastructure.

The conflict also exists alongside peaceful Baloch activism over disappearances, alleged extrajudicial killings and political marginalisation. These civic movements are not the same as the armed insurgency. Balochistan is also home to civilians, workers, tribal communities, public servants and families trying to pursue grievances without being swallowed by the conflict.

The source window began under the shadow of major attacks earlier in July. Pakistan's military said 42 police and military personnel had been killed in three attacks in Balochistan since 6 July, while reporting 54 militants killed in response. Those official figures sat outside the latest week, but they explain why large operations were already underway as the 18 to 24 July period began.

During the week itself, Mastung became the centre of several different incidents. On 22 July, Pakistani authorities said security forces had killed 10 militants during operations in Mastung and Surab over the previous 48 hours. They also reported arresting two alleged would-be suicide bombers and a suspected facilitator. Those were official claims and should remain labelled as such. The identities and circumstances of every reported militant death were not independently verified.

During the same period, gunmen positioned near a road in Mastung opened fire on a passenger coach travelling from Quetta to Karachi. Three passengers were killed and five were injured.

A separate explosion badly damaged a bridge on the Quetta-Karachi highway. Traffic was suspended for several hours, leaving queues of passenger buses and goods vehicles on both sides. Authorities later allowed smaller vehicles to use alternative routes and imposed an indefinite curfew in the surrounding area.

The following day, armed men attacked vehicles carrying judges from Quetta towards Mastung. District and Sessions Judge Abdul Hakeem Kakar and his guard were killed. Another judge and one other person were injured. Responsibility remained unclear. Later reports said Islamic State Khorasan may have claimed the attack, but Dawn said it could not verify that claim.

These incidents were not necessarily part of one coordinated operation. But together, they struck three foundations of ordinary life: civilians travelling by bus, the main highway linking Quetta with Karachi, and judges representing the state's civilian justice system.

A bus shooting tells passengers that routine travel may be dangerous. A bridge explosion tells transport companies and traders that a major route can be cut. The killing of a judge tells public officials that their work may make them targets even outside the courthouse.

None of these attacks captured a city or defeated Pakistan's security forces. They did something subtler. They made everyday movement and administration feel less secure.

Why it matters

Pakistan shows why "unusual" and "serious" are not the same thing. The latest national totals were not exceptional compared with the preceding 12 weeks. The baseline itself was exceptional.

A monitoring system focused only on sudden spikes might overlook Pakistan precisely because severe weeks had become familiar. Some countries demand attention because conditions suddenly worsen. Pakistan demanded attention because an extraordinary burden was beginning to resemble business as usual.

The national total also hid several different Pakistans inside one chart. Sindh produced most of the recorded events through protest and public mobilisation. Balochistan and Khyber Pakhtunkhwa produced almost every death through insurgent attacks and security operations.

Those layers require different responses. Peaceful protests require political engagement, functioning public institutions and restraint by police. The TTP insurgency involves suicide attacks, militant networks, fortified checkpoints and a worsening dispute with Afghanistan. Balochistan combines separatist violence, attacks on infrastructure, resource grievances and a long-running struggle over whether peaceful dissent is being treated as a security threat.

Calling all of that "instability" would be technically correct and practically useless.

The 146 fatalities capture the week's most immediate harm. They do not capture everything that follows. A damaged bridge delays food, fuel, passengers and freight. An attack on a bus can make operators reduce services or passengers avoid a route. The killing of a judge can discourage other officials from travelling, accepting postings or handling sensitive cases. A major assault on a checkpoint can force authorities to reinforce dozens of other positions.

These effects accumulate. One disrupted road is temporary. Repeated attacks can increase transport costs, raise security spending and leave communities feeling cut off. One official requesting protection is manageable. Hundreds doing so can stretch personnel and budgets.

Militant groups do not always need to hold territory to weaken the state's presence. They can make administration more expensive, movement more uncertain and ordinary life more constrained.

Pakistani operations may disrupt armed networks, seize weapons and prevent attacks. They can also bring raids, curfews, arrests, road closures and civilian-harm allegations. Where communities already distrust the state, heavy security measures may strengthen the argument made by militants: that local people are treated as a hostile population rather than citizens entitled to protection and justice.

That does not mean every harsh operation creates new militants. Human behaviour is mercifully less mechanical than that. But a state can win individual gun battles while losing public trust.

The dangerous cycle is not difficult to imagine: militants attack; the state launches wider operations; communities face disruption or alleged abuse; anger and fear deepen; armed groups gain new narratives or support; and further attacks follow. One week cannot prove that cycle is strengthening. The persistence of attacks and counter-operations makes it a pathway worth watching.

What to watch next

The first warning sign is a widening of targets. Much of the lethal violence still centres on security forces, armed groups and state-linked facilities. The danger to the wider public grows if attacks increasingly reach passenger buses, highways, bridges, railways, markets, religious sites, judges, teachers, mines, energy projects or foreign-linked infrastructure.

One attack does not establish a strategy. Repetition does. Another attack along the Quetta-Karachi corridor would raise the possibility that transport disruption is becoming deliberate rather than incidental. Further killings of judges or administrators could suggest an effort to intimidate the civilian state, not only its military and police.

The second warning sign is the complexity of militant operations. The Tank attack combined gunmen with an explosives-laden vehicle against a defended target. Similar attacks would suggest continued access to explosives, reconnaissance and coordinated personnel. Even failed assaults can reveal growing capability if militants repeatedly reach hardened positions.

The third issue is whether domestic violence spills further across the Afghan border. Pakistan says the TTP operates from Afghan territory. Kabul denies it. Relations between the two countries have already deteriorated into cross-border strikes and fighting, raising the risk that another major TTP attack triggers retaliation against Afghanistan, border closures or a wider military confrontation.

That would turn an internal militant problem into something larger, with border communities, refugees and trade routes caught between the two governments.

Balochistan presents a different political warning. If peaceful campaigners, families of missing people, lawyers and political organisers are increasingly treated as extensions of the insurgency, the state may weaken the distinction it needs to preserve between civic grievance and armed violence. The less room there is for peaceful dissent, the easier it becomes for separatist groups to claim that armed resistance is the only remaining route.

A more convincing improvement would require more than one lower weekly fatality total. Attacks and counter-operations would need to decline in both western provinces. Major roads would need to remain reliably open. Public officials and civilian transport would need to operate without extraordinary protection. Suicide and vehicle-bomb attacks would need to become less frequent. Tensions with Afghanistan would need to recede rather than produce another round of strikes.

Until those changes begin moving together, the clearest reading is: