Nuclear early-warning systems are the silent sentinels of the atomic age, designed to detect incoming missile attacks and give national leaders a chance to respond. The United States (and NATO allies) and Russia built extensive early-warning networks during the Cold War, and other nuclear-armed states like China, India, and Pakistan have since begun developing their own. These systems typically include space-based infrared satellites to spot missile launches, ground-based radars to track warheads, and command centres that fuse the data. They have undoubtedly helped prevent accidental nuclear war by debunking false alarms – yet they also suffer from hidden vulnerabilities that rarely receive public attention. Aging technology, cyber risks, false readings (both false positives and false negatives), and gaps in coverage all pose dangers. History is replete with near-disasters caused by early-warning failures, and current efforts to modernize these systems are racing against time to patch the weaknesses. This article examines the early-warning systems of the U.S./NATO and Russia (with a brief look at China and South Asia), detailing their current state, known flaws, and notable failures.

United States and NATO: Dual Phenomenology and Aging Infrastructure

Structure and Operation: The U.S. early-warning network (shared with NATO allies) rests on a combination of space-based sensors and terrestrial radars. During the Cold War the U.S. deployed the Defence Support Program (DSP) satellites in geosynchronous orbit to detect the infrared plumes of missile launches around the globe[1][2]. Today, the DSP has been largely succeeded by the Space-Based Infrared System (SBIRS) satellites, which provide improved launch detection and are being further upgraded in the 2020s. On the ground, the U.S. and NATO maintain large phased-array radar sites such as the Ballistic Missile Early Warning System (BMEWS) radars at Clear (Alaska), Thule (Greenland), and Fylingdales (UK), as well as PAVE PAWS radars on the U.S. coasts that watch for submarine-launched ballistic missiles. These radars and satellites form a dual layer: satellites catch the bright heat of a missile booster, and radars can then track incoming warheads once they rise over the horizon. This “dual phenomenology” – confirmation by both infrared and radar data – is a cornerstone of U.S. early warning, reducing the chance of a single sensor’s error causing a false alarm[3]. Indeed, in multiple Cold War incidents U.S. officers stood down from alert precisely because one of the two sensor types showed no sign of an actual attack[4].

Vulnerabilities: Despite its sophistication, the U.S./NATO system has shown its share of vulnerabilities. Some components are decades old – for example, early BMEWS radars date to the 1960s (though upgraded over time) and even the DSP satellites were first launched in the 1970s. Aging hardware and software can lead to glitches. Notoriously, in November 1979 a technician mistakenly loaded a training simulation tape into NORAD’s computers, causing screens to display a massive Soviet missile attack. U.S. bomber crews scrambled and the Pentagon’s “Doomsday Plane” was launched before it was discovered to be a false alarm[5][6]. A few months later, in June 1980, warning displays at U.S. command centres began showing fluctuating numbers of inbound missiles – first 2, then 200, then zero – due to a faulty 46¢ computer chip. Again a threat conference was convened and nuclear forces went on alert until the warning was debunked[7][8]. These incidents highlight how simple technical failures in aging computer systems nearly sparked disaster. While the U.S. early-warning architecture’s redundancy (satellites and radars) prevented catastrophe in those cases, the human factor was also critical – officers had to trust the second source of data and their instincts to avoid overreaction.

Another vulnerability is the risk of cyber intrusion or sabotage. U.S. nuclear command-and-control networks are heavily secured and often air-gapped, but experts warn that no system is entirely immune to cyberattack. A sophisticated breach could, in theory, spoof false warning data or cripple sensors and communications[9]. Thus, an adversary might attempt to hack or jam early-warning satellites and radars to blind the U.S. or, conversely, inject fake signals to provoke a false alarm. So far, there are no publicly confirmed cases of hackers compromising nuclear early-warning networks (and such systems are largely isolated from the internet), but Pentagon analysts consider the cyber-nuclear nexus a real threat in the future[10]. Even the Emergency Alert System for civilians has been hacked in the past (for example, a false missile alert in Hawaii in 2018, while not caused by hackers, showed how easily public warning mechanisms can sow panic). In short, the U.S. system’s complexity is a strength and a weakness: it provides multiple verification channels, but also a larger attack surface for malfunctions or malicious actors.

Modernization Efforts: The United States has been continually updating its early-warning assets to address these vulnerabilities. The SBIRS satellites (with staring infrared sensors) are more capable than the old scanning DSP satellites at detecting dim or low-signature launches, and a Next-Generation Overhead Persistent Infrared program is underway to improve tracking of emerging threats like hypersonic glide vehicles[11]. On the ground, many early-warning radars have been modernized with solid-state electronics and better resolution. For instance, the PAVE PAWS radar pictured below is part of a network now upgraded under the SSPARS (Solid State Phased Array Radar System) program, extending its service life. Additionally, the U.S. and NATO have integrated data from newer systems such as missile-defence sensors (e.g. Aegis ship radars and the Space Tracking and Surveillance System) to enhance early warning. That said, new types of weapons pose challenges – for example, hypersonic glide vehicles fly at lower altitudes than ICBMs and can manoeuvre, potentially evading U.S. early-warning satellites and radar coverage[12]. This has spurred research into novel sensors (like a dedicated Hypersonic and Ballistic Tracking Space Sensor constellation)[13][14]. In summary, the U.S. early-warning network is robust and redundant, but it must contend with aging components, cybersecurity concerns, and the need to detect new kinds of threats for which it was not originally designed.

Russia’s Early-Warning System: Gaps, False Alarms, and Upgrades

Structure and Operation: Russia’s strategic early-warning system (inherited from the Soviet Union) parallels the U.S. in broad outline but has notable differences. The Soviet Union deployed its own early-warning satellites (the Oko system) starting in the early 1980s, using Molniya-orbit satellites to peer toward U.S. missile fields at the horizon[15]. These satellites, however, had limitations – they viewed missiles against the dark space background by looking at the edge of the Earth (to reduce false alarms from sunlight)[16]. The Soviets also built a network of ground-based radars (such as the Dnepr, Daryal, and later Voronezh series) around the periphery of the USSR to detect missiles coming over the horizon. Today, Russia’s system includes several new Voronezh early-warning radars ringing its territory and a new space-based tier called the EKS or “Kupol” system. Since 2015, Russia has been launching Tundra satellites to rebuild its space-based warning capability after a period in the 2000s when the old Oko satellites failed. However, coverage is still not as continuous as the U.S. enjoys – as of 2023 only about six of the planned ten Tundra satellites had been put in orbit, with no geostationary backup yet, due in part to production delays and sanctions on high-tech components[17]. In practice, Russia relies heavily on its radar network for early warning, more so than the United States does[18]. Data from radar stations and satellites feeds into command centres (like the Serpukhov-15 bunker and the national defence center in Moscow) that would provide launch warnings to Russian leadership.

Vulnerabilities: Russia’s early-warning system has long been considered less reliable and more prone to gaps than its American counterpart[19]. One fundamental issue is the lack of assured “dual phenomenology” – if a Russian satellite reports a launch, there may be no radar in position to quickly double-check it (especially for missiles coming over polar or ocean horizons). Conversely, if radar detects something, Russia’s sparse satellite constellation might not corroborate it in time. This was starkly illustrated on September 26, 1983, when the Soviet Oko satellites reported what looked like multiple U.S. missiles inbound. Lieutenant Colonel Stanislav Petrov, on duty at the Serpukhov-15 warning centre, saw the electronic map light up with “missile launches” and had mere minutes to advise his superiors[20]. He noticed, however, that only five missiles were indicated (a real U.S. first strike would involve hundreds) and, critically, no radar stations picked up any sign of missiles[21]. Trusting his gut, Petrov declared it a false alarm – and indeed, it turned out that sunlight reflecting off high-altitude clouds had fooled the satellite’s sensors[22]. This famous near-miss (often credited with literally saving the world) exposed the vulnerability of Russia’s space-based warning: the old satellites could not reliably distinguish missiles from natural phenomena under certain conditions[23].

Russia has also experienced false alarms on the radar side. The most serious was the Norwegian rocket incident of January 25, 1995. A scientific rocket (a four-stage Black Brant XII) launched off the coast of Norway was detected by a Russian early-warning radar in Murmansk, which interpreted the high-flying projectile as a possible U.S. Trident missile[24]. For the first time in history, the Russian nuclear “Cheget” briefcase was activated for President Yeltsin – he and his top commanders had to consider within minutes whether this blip was the first salvo of a surprise U.S. attack (possibly an EMP weapon to blind Russian radars)[25]. Fortunately, as the trajectory became clear, Russian observers realized the rocket would fall into the sea and posed no threat, and the alert was cancelled[26]. Afterward it emerged that Norway had in fact notified Russia of the pending scientific launch weeks ahead, but the message never reached the right defence officials[27]. This incident underscored procedural and communication failures in Russia’s system – and how a single unrecognized radar track almost precipitated a nuclear crisis. It also highlighted the short warning times Russia faces. A submarine-launched missile from near Norway could hit Moscow in 10–15 minutes, leaving essentially no time for double-confirmation. Analysts note that Russian doctrine, as a result, tilts toward a “launch-on-warning” posture and even possible pre-delegation of launch authority to lower commanders in a fast-breaking scenario[28]. This raises the risk that overreaction to a false alarm could occur if leaders fear waiting for impact confirmation[29].

Beyond false alarms, Russia’s early-warning network has other little-discussed vulnerabilities. Many of its radars were aging or located in former Soviet republics that became inaccessible; during the 1990s, Russia actually lost coverage in some approaches (for example, a gap in the early 2000s when the last Oko satellite failed[30]). While new Voronezh radars have since been built to close coverage gaps[31][32], these large installations are themselves targets. In fact, during the Ukraine war, previously unthinkable attacks on Russia’s early-warning infrastructure occurred: in May 2024 a Ukrainian long-range drone struck the Voronezh radar station at Armavir (Krasnodar Krai), a site crucial for watching the south-western direction[33]. Another radar site was reportedly attacked the same month[34]. Although Russia did not lose its entire warning capability from these isolated strikes, they reveal a physical security weakness – the radars are fixed, big, and vulnerable to sabotage or bombing in a conflict[35][36]. Similarly, Russia’s satellites could be shot down or blinded in orbit by anti-satellite weapons or electronic jamming. Russian officials openly worry about “blinding” attacks: a nuclear blast in space could ionize the atmosphere and blind radar sensors, something early-warning crews might even mistake as the prelude to a nuclear strike[37]. Indeed, the Russian defence ministry has prioritized making new radars more jam-resistant and power-resilient, announcing plans to upgrade at least three Voronezh stations by 2028 with improved electronic protections[38]. All these factors – patchy space coverage, extremely short decision times, aging legacy systems, and susceptibility to attack – make Russia’s early-warning system arguably less robust than America’s. Analysts have called it “relatively outdated” and prone to technological failures that exacerbate escalation risks[39]. This precariousness is one reason Russia developed unique fail-safe measures like the semi-automated “Dead Hand” (Perimeter) system, which, in a doomsday scenario of leadership decapitation, could still trigger retaliation based on remote sensors detecting nuclear detonations[40]. (Notably, despite wild rumours, even Perimeter is believed to require a human decision in the final step[41].)

Modernization Efforts: Acknowledging these vulnerabilities, Russia has been actively modernizing its early-warning apparatus in the last decade. The ground-based Voronezh radar program is considered a success – nine new Voronezh radars are now operating, replacing most of the 1970s-era Dnepr/Daryal radars and restoring 360° coverage of missile launch corridors[42]. These modern radars are modular, easier to maintain, and can be brought online faster than previous generations. Russia has also deployed specialized high-frequency over-the-horizon radars (the “Rezonans-N” in the Arctic) to catch lofted missile trajectories and even stealth aircraft[43]. On the space side, the EKS “Kupol” satellite constellation is being deployed more slowly. As of 2021, the Russian military acknowledged having only four Tundra early-warning satellites in service[44], far short of global coverage. By 2023 this grew to six, with more launches planned[45], but Western analysts note that Russia has not met its original timeline (10 satellites by 2020) due to technical delays and sanctions impacting satellite components[46]. The new Tundra satellites are reportedly more capable than their Soviet predecessors – there are indications (though debated) that they might have a limited “look-down” infrared ability to view warheads against Earth’s background[47]. Even so, experts like Dr. Theodore Postol have argued that if Russia’s new satellites truly had modern sensors, they wouldn’t need so many in highly-elliptical orbits; the continued launch of multiple Molniya-orbit satellites suggests Russia is still compensating for coverage gaps[48]. In effect, Russia is catching up on early warning, but the system remains less automated and less trusted, even by its own operators. Russian writings emphasize maintaining human control and even advocate waiting for actual nuclear impact evidence before retaliating[49] – a risky proposition given the short flight times. In summary, Russia’s early-warning system today is a mix of state-of-the-art radars and patchwork space sensors, making it a focal point of concern. Observers worry that an accident or misinterpretation in this system – compounded by Russia’s launch-on-warning posture – could spark a nuclear exchange by mistake[50].

China’s Nascent Early-Warning and India/Pakistan’s Gaps

China: For decades, China lacked a dedicated nuclear early-warning system, relying instead on a minimal radar network and slower intelligence methods. This is now changing rapidly. In 2019, Russian President Putin revealed that Moscow was helping Beijing build a missile attack early-warning system[51] – a remarkable collaboration, since until then only the U.S. and Russia had such systems. China has since invested in both satellites and ground radars. According to U.S. Defence Department assessments, by 2022 China had at least three early-warning satellites in orbit[52], likely a result of cooperation with Russia’s Tundra/EKS technology. China has also fielded large phased-array radars in its northeast, northwest, and central regions to cover possible missile trajectories from India, Russia, or the Pacific[53]. One example is a new radar in Heilongjiang province reportedly looking northeast towards the Korean peninsula and Japan[54]. Chinese and Russian militaries have even conducted joint computer-simulated missile defence exercises (“Aerospace Security” drills) to integrate their warning data[55]. All these moves indicate that China is fast-tracking the creation of a modern early-warning network.

Despite this progress, China’s capabilities remain limited and largely untested. Chinese early-warning satellites are believed to be experimental or in early deployment – perhaps not yet providing full global coverage. Analysts still describe China’s missile defence and warning sensor layer as the “largest gap” in its strategic forces[56]. In other words, China until recently would have had little to no warning of a nuclear missile attack, raising the risk of being caught by surprise or misidentifying other events as attacks. This is a vulnerability seldom discussed openly by Chinese officials (due to secrecy), but Western experts note it could make China jittery in a crisis. For example, without reliable warning, China might be more inclined to an accidental launch-on-false-warning or to delegate nuclear forces to hair-trigger alert, just to avoid losing them on the ground. The introduction of early-warning sensors should, over time, ameliorate that risk by giving Beijing more confidence and decision time. However, Chinese early-warning systems will face familiar challenges: distinguishing real launches from false alarms, integrating data from satellites and radars (a new undertaking for the PLA), and protecting these systems from interference. Given China’s heavy censorship, any near-miss incidents or technical glitches in its warning system may not become public knowledge. It is worth noting that China tested hypersonic glide vehicles that can approach the U.S. from the south polar direction, exploiting gaps in U.S. early-warning coverage[57]. By the same token, China must realize that its own early-warning could be outflanked by unconventional trajectories unless it attains global sensor coverage. The ongoing China-Russia cooperation – including reported plans for Russia to deploy its Voronezh radars on Chinese soil – will significantly enhance China’s coverage[58]. In summary, China is moving from having virtually no early-warning toward a fairly modern system in the coming years, but until that is fully in place, its nuclear posture carries an underappreciated vulnerability.

India and Pakistan: The two South Asian nuclear rivals have only rudimentary early-warning arrangements, which is a serious concern given their short missile flight times (as little as 5–10 minutes across the border). India has developed a series of long-range tracking radars as part of its Ballistic Missile Defence program – for instance, the Swordfish radar (derived from Israel’s Green Pine) can detect incoming missiles from hundreds of kilometres away and is integrated with India’s missile launch detection network[59]. India’s satellites are not currently known to include dedicated missile-warning sensors, but there have been discussions about using indigenous satellite technology for that purpose[60]. In effect, India’s early warning is based on an overlapping ground radar network and perhaps some space-based experimental infrared assets. Pakistan, on the other hand, has even fewer capabilities; it lacks a known equivalent of Swordfish and does not operate early-warning satellites. Islamabad likely relies on simpler air defence radars and intelligence to sense an incoming Indian attack. The net result is that both countries face extremely compressed decision timelines, heightening the risk of miscalculation. A technical glitch or misread radar blip in a tense Indo-Pak crisis could lead to a launch-on-warning with almost no chance for verification. Thankfully, no major false alarm incidents have been publicly reported from South Asia’s nuclear history – possibly because neither country has continuous automated warning systems (they have to some degree kept nuclear warheads de-mated and rely on human command chains). Still, as both nations modernize (India, for example, is considering space-based warning as its missile programs grow), they will need to guard against the same pitfalls the U.S. and Soviet Union encountered. False positives or false negatives in a fog of war scenario (e.g. a radar picking up a routine missile test or satellite launch and mistaking it for an attack) are a latent danger. Moreover, any cyber-attack on command networks or blinding of radars (via electronic warfare or physical strikes) could be destabilizing since redundancies are few.

Hidden Vulnerabilities and Notable Failures

Early-warning systems, as seen above, are complex systems-of-systems with many potential failure points. Some vulnerabilities are technical; others are organizational or even stemming from human psychology. This section highlights key categories of vulnerabilities – many of which have only come to light through incidents and declassified stories – along with examples of when things went wrong.

• Aging Technology & Maintenance Issues: Many early-warning sensors and command systems were built in the 1960s–80s and, despite upgrades, are prone to breakdown. Small failures can have outsize effects. The 1980 NORAD false alarm (showing 2,200 phantom missiles) was traced to a single microchip error[61]. In 2010, a 70-cent computer part caused a backup communications network at NORAD to fail, though fortunately without false attack indications (anecdotes like this are often quietly fixed without public fanfare). Aging hardware is also a security risk – older computers might be more vulnerable to power surges or even malicious firmware. Recognizing this, Russia is overhauling power supply systems at its radar stations by 2028 to improve reliability[62]. The U.S. has similarly poured funds into modernizing missile warning centres’ IT infrastructure. Nonetheless, until legacy systems are fully replaced, the spectre of an old circuit glitch or sensor degradation causing a false alert remains. Each upgrade cycle is effectively a race between obsolescence and catastrophe prevention.
• False Positives (Phantom Attacks): History has shown that early-warning systems do sometimes cry wolf – and it is usually quick thinking (or sheer luck) that averts disaster. Besides the 1979 and 1980 U.S. false alarms mentioned, the 1983 Soviet Petrov incident is a textbook case: a satellite report of incoming missiles that was completely wrong[63]. In that case, the false positive was caused by a rare alignment of sunlight and cloud tops, a scenario designers hadn’t anticipated[64]. Another false positive was the 1995 Black Brant rocket incident in Russia: a single research rocket’s trajectory mimicked a possible nuclear strike on radar screens[65]. False alarms can also stem from human error (as simple as selecting the wrong input on a computer, as in 1979[66]) or software bugs. Every false alert is dangerous because it compresses decision time – military officers may have mere minutes of uncertainty to decide if it’s real or not. The common thread is that multilayered verification is vital: in 1983, the absence of radar confirmation saved the day[67]; in 1979/1980, the presence of satellite data that showed “no launches” overruled the computer displays[68]. Thus, one unspoken vulnerability is over-reliance on a single source – a lesson bought with several near-miss events.
• False Negatives (Missing a Real Attack): The flip side is an early-warning system failing to detect an actual missile launch – a terrifying scenario that could leave a country blind to an incoming strike. While no full-scale attack has occurred (thus no known false negatives in war), there have been close calls. In 1962 during the Cuban Missile Crisis, for instance, a U.S. BMEWS radar in Alaska temporarily lost track of a test missile, causing momentary confusion. More pertinently, Russia’s sparse satellite coverage means it might not see a missile coming in certain trajectories. Experts noted that Russian satellites until recently could not look down over U.S. missile fields – they viewed at an angle – so a missile that didn’t rise above the horizon of the satellite’s view cone could slip past detection[69]. Submarine-launched ballistic missiles (SLBMs) flying on a low loft could fall into this blind spot[70]. In fact, Russian officials conceded that only when a missile’s warheads show up on radar (much later in flight) can they be sure – which by then leaves perhaps 10 minutes or less to react for SLBMs. The danger of a false negative is panic on the other side: the attacked nation might absorb a first strike without warning, potentially losing its deterrent forces. This is why nations strive for early warning – lack of it is itself a destabilizing vulnerability. Until recently, China essentially lived with that vulnerability (no satellites watching globally)[71]. Another example: in 2010, a U.S. DSP satellite actually missed the launch of a Scud missile during a combat situation, failing to cue an alert – it was a minor theatre case, but it spurred the deployment of newer SBIRS sensors with better sensitivity. In summary, while false alarms get more attention, an undetected real attack is the ultimate nightmare the entire system is meant to prevent – and any gaps that make it possible (sensor coverage holes, inoperable satellites, etc.) are top priority to fix.
• Cyber and Electronic Warfare Threats: Modern early-warning networks depend on vast software code, digital communication links, and sensor electronics – all of which could be targeted by cyber or electronic attacks. A sophisticated adversary might attempt a cyber infiltration to disable warning systems or feed them disinformation. For example, malware might be inserted (via insider or supply chain compromise) that masks a real launch or generates a fake one. Even “air-gapped” systems (not connected to the internet) are not immune – the Stuxnet incident showed that determined attackers can reach isolated networks via infected removable media[72]. The U.S. Government Accountability Office has repeatedly warned of cyber vulnerabilities in defence systems, and early-warning is no exception[73]. In addition to cyber, electronic jamming of radar or satellite uplinks is a concern. During a conflict, an opponent might try to jam a key Russian radar, for instance, to create confusion or blind a segment of coverage. Russian developers explicitly cite improved jamming resistance as a goal for radar upgrades[74]. There is also the prospect of anti-satellite (ASAT) weapons – Russia, China, and the U.S. have all tested ASAT missiles that could destroy or disable early-warning satellites in orbit. In a severe crisis, one side might be tempted to knock out the other’s “eyes,” but doing so could be interpreted as preparation for a nuclear attack (since it removes the victim’s ability to see an incoming strike)[75]. Thus, cyber/ASAT attacks on early-warning systems are highly escalatory. They remain an under-discussed vulnerability because no such attack is known to have occurred yet – but military planners are certainly aware of the risk.
• Human Factors and Decision Pressures: Finally, a vulnerability that is impossible to engineer away is the human element. Early-warning systems inevitably rely on human judgment at critical junctures – to interpret ambiguous data, to decide if an alert is credible, to communicate effectively under stress. The Stanislav Petrov episode in 1983 exemplifies how an individual’s decision (in this case, to delay and double-check) can avert catastrophe[76]. Conversely, humans can also err by believing a false alarm or by miscommunicating information. In 1995, Russian authorities did not properly disseminate Norway’s prior notice of a scientific launch, contributing to their surprise[77]. Human psychology under extreme time pressure is a wild card: cognitive biases or worst-case assumptions may take hold. If, for instance, geopolitical tensions are already high, leaders might be more inclined to view any warning as the real thing. This is why both Washington and Moscow have, at times, explored confidence-building measures like sharing early-warning data. In 1998 they agreed in principle to establish a Joint Data Exchange Centre in Moscow, where U.S. and Russian officers would jointly monitor missile launch data – aiming to prevent misinterpretation on either side[78]. A pilot project ran around the Y2K rollover with Russian officers present at NORAD in Colorado Springs[79]. Unfortunately, the permanent centre never materialized (plans were dropped amid later political tensions)[80]. Such cooperation could mitigate human misperception by ensuring both sides see the same picture in real time. Its failure to launch is itself telling political distrust can undercut even the best technical systems. Therefore, the human and political dimension remains a vulnerability – one “no one talks about” until it’s too late. It is sobering that deterrence stability hinges not just on machines, but on fallible people interpreting those machines.

Conclusion: Reducing the Risk

Early-warning systems for nuclear attack are often hailed as vital safeguards – and indeed, they have prevented more disasters than we will ever know. Yet, as we have seen, they are far from infallible. Technical limitations (aging sensors, limited coverage, new hard-to-detect weapons) and the potential for errors (mechanical, software, or human) mean that the risk of an accidental nuclear exchange cannot be entirely eliminated. In fact, analysts argue that the risk of a nuclear war by mistake may be growing in an era of fast-paced warfare and cyber threats, unless these vulnerabilities are addressed[81]. Both the United States and Russia (and other nuclear states) are taking steps to modernize their early-warning networks – deploying more advanced satellites, hardening systems against jamming or hacking, and improving data processing with AI assistance. These efforts aim to buy precious minutes or seconds, make sensors more trustworthy, and give decision-makers greater clarity in a crisis.

Ultimately, however, no technology can provide absolute certainty. A sobering question remains: what happens when warning systems fail or send mixed signals? The answer must lie in doctrine and diplomacy as much as engineering. Restraint policies – like not rushing to “launch on warning” without confirmation – are one way to cope with uncertainty[82]. Direct communication links and data-sharing between rival powers are another, so that an innocent rocket launch or a radar outage is less likely to trigger panic. The fact that false alarms in 1979, 1983, and 1995 did not lead to nuclear war was due to level-headed individuals and a dose of luck. We may not always be so lucky. Transparency about early-warning vulnerabilities, plus robust fail-safes and international dialogue, are critical to ensure that the first mistake is not also the last. As one U.S. and Russian joint study concluded years ago, the only cure for false alarms is building mutual trust – a reminder that the ultimate reliability of any early-warning system lies not only in its radar range or code, but in the wisdom with which its warnings (or silences) are interpreted[83].