Isegoria

Before the war, on a tactical level the Armenian army was superior to the Azeri army:

It had better officers, more motivated soldiers, and a more agile leadership. In all previous wars with Azerbaijan, this proved to be decisive. But Azerbaijan found a way to work around it. This is where the drones came in: they allowed the Azeris to reconnoitre first the Armenian position and then the placement of reserves. Armenian positions then could be extensively shelled with conventional artillery, weakening their defences. Drones then guided the onslaught towards the Armenian reserves, bringing in artillery, multiple-rocket systems with cluster munitions, their own missiles, or using Israeli-made LORA ballistic missiles to destroy bridges or roads linking the reserves with the front. Once the Armenian side was incapable of sending reserves into battle, the Azeri army could move in any number it wished to overwhelm the isolated Armenian positions. This procedure was repeated day after day, chipping one Armenian position away each day and resupplying artillery during the night.

This tactic also worked well in mountainous territory the Armenians thought would be easy to defend. In the mountains, there is only one road connecting the front to the rear, which made it even easier for drones to spot targets. When the battle over Shusha demonstrated that the Armenians would not stand a chance even in this territory, the Armenian army started to disintegrate and Yerevan had no choice than to agree a ceasefire on adverse terms.

One of the military lessons from Nagorno-Karabakh is that computers and networks matter:

Like in Syria and Libya, Russian air-defence systems proved to be ineffective against small and slow drones. This has inspired a debate in the West about whether Russian air-defence systems are generally overrated. But this verdict would be premature.

Armenia’s most ‘modern’ air-defence systems, the S-300PT and PS series and the 9K37M Buk-M1, were both developed in the 1980s. While the missiles are still potent, their sensors are designed to detect, identifiy and track fast-moving fighters, and their moving-target indicators disregard small, slow drones. Like many 1980s systems, a lot of computing is predetermined by hardware layout, and reprogramming requires an extensive refit of the entire system, which the Armenians had not done. These systems are also incapable of plot-fusion: accumulating and combining raw radar echoes from different radars into one aggregated situation report. Plot-fusion is essential to detecting small and low-observable targets such as advanced drones or stealth aircraft. None of the export versions of Russia’s air-defence systems that it has sold to Syria, Turkey, North Korea, and Iran are capable of plot-fusion. (In the latter two cases, these are disguised as ‘indigenous’ systems like the Raad or Bavar 373.) There is therefore a huge difference in performance between Russian air-defence systems protecting Russian bases in Armenia and Syria and those Russian air-defence systems exported to Armenia and Syria.

Azerbaijan’s drones roamed free because Armenia had no jammer able to interrupt the signals linking the drones to their guidance stations. Only in the last days of the war did Russia use the Krasukha electronic warfare system based at the Armenian city of Gyumri to interdict Azeri deep reconnaissance in Armenia proper. Still, the Azeris also used the Israeli Harop loitering munition, which was able to work under adverse conditions (although at reduced effectiveness) as it does not, unlike drones. require a guidance link. Hence among armies that are likely to prepare to fight wars in the future – not only the US, China, Russia but regional powers such as Turkey, Israel, and South Africa – this experience will certainly prompt further research into artificial intelligence and autonomous lethal weapons systems. Rather than banning this class of ammunition by a prohibitive arms control treaty, as envisioned by Europe, they will experiment with how to make use of the new technologies and best integrate autonomous lethal weapons systems into their combined-arms manoeuvre forces, thereby increasing their operational tempo and effectiveness.

The information war has been just as fierce as the actual war in Nagorno-Karabakh, with both sides posting daily combat footage to proclaim victories:

Disinformation and propaganda, spread through official and unofficial accounts, have made it difficult to objectively assess the course of combat thus far. Furthermore, the relative accessibility of combat footage — whether from drones, cellphones, or cameras — paints a stylized picture of the battlefield for any analyst. They are official propaganda, and it is worth noting that on the modern battlefield, some systems have cameras or live video feeds, while many do not, distorting perceptions on combat effectiveness. A social media feed composed largely of drone video footage could lead one to believe in the dominance of such systems, even in a conflict where many casualties are still inflicted by armor, artillery, and multiple launch rocket systems. This tactical footage has led to familiar debates on the utility of tanks, the prowess of drones on the battlefield, or the proliferation of sensors.

There is a thirst for drawing lessons from contemporary conflicts that feature modern weapon systems. However, the result is often generalizing from a few cases, and at times, learning things that are not true. What can be discerned from this war is hardly revelatory. Remotely operated systems offer the utility of tactical aviation, close air support, and precision guided weapons to small nations, and to even relatively poor countries, for a cheap price. They saturate the battlefield with disposable sensors, shooters, and sensor-shooter packages in the form of loitering munitions. Notably, they enable precision artillery and strike systems to engage fixed positions, as has been seen across modern conflicts from Ukraine to Syria. Furthermore, tanks are vulnerable to counters, as they always have been, but it is unclear what other vehicles offer a better combination of firepower, protection, and maneuverability on the battlefield.

The war illustrates that in an offensive, or counter-offensive, the only thing worse than being in a heavily armored vehicle is being outside of one. If anything, the tank appears to be the most survivable vehicle, given the small warheads on drone carried munitions. These munitions often disable or mission kill the vehicle, but the crew can still survive anything other than a direct hit. Much of the hand-wringing in Western circles that comes from watching these conflicts stems from the epiphany that there is no way to avoid casualties on the modern battlefield, especially among an expensive force, replete with boutique capabilities that cannot be lost in large quantities. Furthermore, the ratios of support to maneuver units are important. Compared to forces like the Russian military, Western ground units feature poor availability of air defense and electronic warfare, and the expectations that existing air defenses or tactical aviation may be easily adapted to counter unmanned systems are probably unfounded. Armenia’s performance illustrates this problem. Drones are relatively cheap, and this military technology is diffusing much faster than cost-effective air defense or electronic warfare suitable to countering them.

That said, Azerbaijan’s unmanned air force has been operating against an opponent with incredibly dated short-range air defenses which are neither suitable nor effectively employed to defend against drones. Armenia does not have layered air defense, effective electronic warfare, or a large amount of tactical aviation. It has situated its air defense systems in relatively exposed fixed positions, in a mountainous region where air defense is even more difficult by virtue of the terrain. In truth, both sides are demonstrating tactical deficiency in their offensive and defensive tactics. While attaining some kills using optical sights, Armenia’s modernized Soviet systems (essentially technology that dates back to the early 1970s) were never meant to engage combinations of small drones, loitering munitions, precision artillery, or unmanned combat aerial vehicle systems. More advanced air defense capabilities like Tor-M2s are few, and have been intentionally held in reserve, although Azerbaijan has been reticent to use its fixed wing or rotary aviation. Armenia’s older S-300PS systems appear to have had no role in the conflict, and some launchers may have been destroyed early on, having never even been deployed.

The lessons from this conflict are consistent with those of other wars in the latter 20th century: It is much better to have a smaller ground force that is well defended from the air, than a vast armored force that is completely exposed to sensors and airpower from above. Well prepared defenses, if insufficiently protected or camouflaged from the air — which is increasingly difficult — are naturally vulnerable. The diffusion of remotely operated systems will outpace that of air defenses or specialized counter-drone systems, rendering older generations of air defense largely obsolete. Drones and loitering munitions will be, for some time, cheaper to acquire than the requisite defenses. And one can distribute forces, but they should be concentrated for assaults. There is no way getting around canalizing terrain, at least not until the battlefield features hover tanks. That tanks are vulnerable to anti-tank weapons should come as no surprise, but other vehicles, which trade survivability for maneuverability, seem to fare no better against anti-tank guided missiles. Vulnerable or not, it is unclear what other vehicle can achieve the tank’s mission on the battlefield.

The Army wants the first casualty of the next war to be a robot, not a human being:

Army studies of recent conflicts — Russia vs. Ukraine, Armenia vs. Azerbaijan — show you can have a dramatic impact by adding a small infusion of 21st century tech to a largely Cold War force, [Maj. Gen. Patrick] Donahoe said. How? One approach the Russians have employed to devastating effect is to use drones to spot targets for rocket launchers. Likewise, while the US Army is developing a host of new missiles, armored vehicles, and aircraft, most units will be using Reagan-era hardware for years to come. In essence, Donahoe wants to organize these existing weapons in new formations and add drones and ground robots to scout ahead.

[...]

Historical data on direct-fire engagements “shows that our enemies generally shoot first 80 percent of the time,” Sando said. “We don’t like those odds, [so] we want to avoid the close fight if we can. If we can’t avoid it, we want to enter it under conditions that are favorable to us.”

But how? Current Army doctrine prescribes “making contact with the smallest element.” In layman’s terms, if you must stumble upon the enemy and get shot at (the formal term for this is a, “meeting engagement”), then do it with the smallest vanguard possible, giving the main body time to prepare and maneuver without being pinned down. In the future, Donahoe said, the goal will be to make first contact with an unmanned element.

Cold War doctrine envisioned engaging the enemy along what’s called the Forward Line Of Troops, or FLOT. In the new concept, according to a briefing at the conference, a Forward Line Of Unmanned Aerial Systems (FLUA) will fly ahead through no-man’s-land into enemy-held territory, followed by a Forward Line Of Robots (FLOR) on the ground, followed in turn by the Forward Line Of (Human) Troops. The unmanned systems will flush out the enemy, stumble into meeting engagements and ambushes, take and receive the first hits, and map the enemy position for the human troops coming along behind them.

Of course, the Army can’t do this today. To execute the concept in reality, they need a lot more unmanned systems, so they’re going to build them.