Kant's categorical imperatve and defence policy

A simple (not flawless) way of checking whether one's policy is ethical is to apply the categorical imperative; would the outcome be a good one if everyone behaved like this?

I have repeatedly argued on this blog (and elsewhere) that small powers should not create or maintain well-rounded miniature military forces, as such armed forces are near-useless for deterrence and defence. Or did you hear any good of the deterrence and defence values of the armed forces of Denmark, Norway, Estonia, Latvia, Lithuania during the Second World War? The only small power that resisted a great power successfully was Finland, and it did only so for a short time under uniquely favourable circumstances.

Likewise, I argued repeatedly that even some of the bigger NATO members such as Germany should orient their armed forces towards a specific threat scenario, complementing their alliance rather with contributions tailored to their situation rather than well-rounded armed forces in ignorance of any threat scenario. Germany could disband its navy with absolutely no harm to its or its allies' security, for example.

Moreover, military budgets should be set in the context of the military overspending of allies (that is, one should spend less due to others overspending rather than to emulate their silliness).

This doesn't seem to be an ethical course of action at first look and certain not very deep thinkers mis-use the word "freeriding" a lot even in the current real world situation.

To think so requires a static look at the world. My advice would change if the overspending by allies would end, for example. To devise a budget in context of allied budgeting may just as well lead to an increase of own spending when allied overspending ends.

A deterioration of air power or sea power or the primary threat scenario becoming unrealistic would lead to different conclusions, a different optimum, a different advice.

So what really drives my thinking about force structures is not a simplistic preference for less, but an appreciation of real world circumstances; the size of threats, the capabilities of allies, the own economic and fiscal capabilities and the own geographical position.


I do assert that much of the advice regarding more military spending, more warships, more combat aircraft, more tanks, more brigades is not driven by an appreciation of threat capabilities, allied capabilities and geography. Most calls for "more" rather seem to be driven by money interests, officer self-interests or very much subjective power fantasies. This ranges from millionaire CEOs and lobbyists to flag ranks who put their service before their country to simple fanbois.



Navies' obsession with peacetime hull quantities


Navies tend to seek to have many (and powerful = prestigious) warship hulls in peacetime. The more ships there are, the more jobs for junior officers, the more jobs for senior officers. The bigger the navy and the more central it is to the country's policies and self-image, the more prestige do the naval officers enjoy.

The principal-agent problem clearly leads to navy officers wishing for more naval expenditures than optimal for the nation's overall best interests.

There is a particular obsession of navies (and their fanbois*) with peacetime warship hull quantities. The U.S. navy outright obsesses over made-up hull quantity requirements that are supposedly necessary. Those quantities haven't been met in decades (if ever), but neither the U.S. nor its allies were blockaded, bombed or invaded by another country. There's also an anecdote about a German inspector of the navy (highest-ranking naval officer) who publicly regretted that he couldn't commission a single new ship during his time leading the German navy. The fact that Germany wasn't blockaded, bombed or invaded and commissioning an additional ship was thus proved to be nothing but a waste of resources didn't seem to have crossed his mind.

The obsession with peacetime warship quantities goes on despite the obvious fact that air power rules the surface of the seas. The U.S. Navy itself spent more money on buying aircraft than on buying warships a few years back, but the obsession was still about ship hulls, not about aircraft fuselages or naval aircraft fuselages. Land-based aviation can wipe out surface fleets at much lesser costs than any warship or submarine fleet could, and naval aviation is clearly technically disadvantaged to already dispersed land-based aviation (land-based aviation has more capable support aircraft, potential access to OTH radar data, isn't burdened by tailhooks or strengthening for stressful landings).

The overemphasis on warships leads to a neglect of other items that could be of great importance, similar to how Western air forces neglect ballistic missiles of the Iskander-ish class and air defences. 

Another analogy is modern armies neglecting preparations for dealing with prisoners of war (too few infantrymen or other temporarily detachable troops to handle them).

I'm writing about ships that can meet the traditional cruiser role of enforcing a distant naval blockade. Submarines have marginal target inspection abilities. They can try to enforce a naval blockade close to hostile ports, but a American-Sino war might see ships slipping through Indonesian and Philippine waters. What asset would inspect, identify and possibly board or (more risky in case of error) sink them? Traditionally, this would have been the job of cruisers. The USN has no high endurance frigate fleet for the job, and cannot spare its destroyers.

A simple answer could be auxiliary cruisers equipped with a relatively simple medium helicopter for boarding actions and another medium helicopter for sinking of ships. Those helicopters would also need to have decent thermal cameras, but that's not much of a cost driver nowadays if you are willing to improvise instead of gold-plating.

It would be very easy to modify a small container ship for auxiliary cruiser tasks, including armament with 106 mm recoilless guns and ManPADS. You would need containers for electricity generators, kerosene supply, helicopter hangar, helipad, extra galley, extra mess, bunks, a control centre, radio, washing rooms/showers, munitions storage, food (cold) storage, a medical container, containers for POWs/detainees, tools+maintenance machinery+spare parts, extra firefighting capability. 

In short; you'd want to have the containers developed, prototyped, tested and improved before a hot conflict.

About the same kind of auxiliary cruiser might also be effective in use by an underdog navy, but I lack the means and knowledge to determine whether modern satellite ocean surveillance including satellite AIS snooping might suffice to make the tasks of a merchant raider too hazardous if opposing forces have land-based air power or other assets in range to strike at it.

I have not seen any navy building up such a capability even though the East Africa piracy thing would have been a perfect excuse to get it funded. This is but one of many reasons why I doubt that Western navies are serious about deterring major war or being able to win major wars. The interventionist Western countries developed their navies into bullying/land attack forces with parallel and separate nuclear strike submarines. Even the U.S: navy, which had reason to pay attention much attention to PR China for well over two decades, is not all that serious about these parts of its job.


*: Many of which appear to believe that parroting calls for more warships mixed with calls for minor changes especially regarding long-obvious failure programs turns them into impressive thinkers.



Link dump May 2021






A large navy should occasionally have experimental ships, but the article's emphasis on the XQ-58 is ill-advised. Such drones have little ability to penetrate defended airspace on their own, little ability to find and identify semi-stationary or mobile land targets, and its payload is really small. Furthermore, outranging those land-based anti-ship missiles by a bit only puts targets close to some strips of coastline in range. An air war against China would either be about attacks on its sea and air power or in a more ambitious scenario a campaign against its electrical power network (particularly powerplant turbines) might also be worthwhile. This requires much deeper strikes. The vastly better alternative would be to focus on a distant naval blockade and defence of Japan/Taiwan, of course.

- - - - -


This is very similar to Germany in the 1880-1914 and 1920-1932 period and also very similar to modern sub-Saharan African politics.




No security policy consequences from burps (serious text)





The degeneration and reality disconnect of the right wing particularly in the U.S. is tragic-comedic, but the issue is an interesting one.

I suppose that cultured meat is going to be superior in price AND quality (different flavours) to meat from complete live animals within 10...20 years and the pork industry will collapse, while the cattle industry will likely prevail with a some shrinking due to less incomes from meat (higher income share from milk than today).

This means that indeed the methane burps (burps mostly, hardly any of the methane emissions are by farts) of cattle are an issue. Methane is a greenhouse gas (more potent, shorter-lived than carbon dioxide) and thus of concern. The extreme methane emissions of first and foremost cattle (not nearly as much other animals) stem from fodder that they have no evolutionary optimization for. The answer is not necessarily a reduction of meat consumption; it's to add few per cent (1...3%) food additives (such as seaweeds) that multiple studies found to be effective at suppressing 80% to almost all methane emissions (example) by changing the digestion processes. The challenge is to make this transition; the food additives production capacities first need to be created, and this requires incentive mechanisms by government becuase the benefits to society do not get internalized to the business decisionmakers by the market so far.

(Compare the obscene kindergarten-level of discourse chosen by right wing propagandists in the U.S. with this science-based summary! That's a fear and hate propaganda vs. reality contrast. I've known all of this stuff for about two years without being a industry insider, there's no excuse for media professionals to relay any of the nonsense on the topic.)

Is this about defence and freedom? It's peripheral. This topic and the prospect of water-saving greenhouses and hydrocultural production of food signal a likely coming revolution in our food supply. The challenges seem manageable and food production can very likely happen domestically. I don't see any need for security policy efforts aimed at securing food supply from other countries any more than so far.



Ten options when facing a threat of specific geographic origin (II)

.continuing from part I

(warning; long wall of text)

Let's remember the ten options* first:
  1. Economic attack
  2. Base strike
  3. Cordoning
  4. Mobile warfare
  5. Hunting patrols
  6. Convoying
  7. Secured zones
  8. Shadowing
  9. Infrastructure attack 
  10. Enduring the problem
Military history knows many examples, but they were not always used in the same mix, or with the same success. Inferior powers had to use or emphasize different options than superior ones. Terrain was very influential, as was technology and industrial support.
- - - - -
I'm concerned that we in the West are in part wasting resources by following obsolete paths, using an obsolete mix and structure of options to counter threats.

The expected shortness of WW3 led to an underappreciation of #1, economic attack. The one big exception is the campaign to keep small powers from acquiring or adding to a nuclear arsenal. The Western or American economic sanctions likely do just about nothing against Russia or PRC that would be important in a future hot conflict. Economic warfare against the opposing forces did not play much of a role in the years-long and even decades-long wars of occupation beyond mere interception of arms and munitions deliveries. Much more could have been done by targeting the economics of the warlords who had to pay at least the mobile part of their forces and for some arms buys.

Base strike is widely appreciated, but it won't help against threats without a clear geographic origin. An army of a million quadcopter kamikaze drones launched from shipping containers is not going to be countered by strikes on airbases or barracks. NATO paid much attention to cratering airbase runways until precision-guided bombs finally made direct hits on hardened aircraft shelters more promising. The transition wasn't rapid, though. Munitions and aircraft meant for runway cratering kept soldiering on for many years after the approach became the inferior one.

Cordoning is a very interesting approach. I wrote a lot and already long ago about the fact that there aren't enough troops to form a proper front-line that makes infiltration very dangerous. Modern inter-state land warfare can be very mobile or take the shape of a hedgehog defence as seen in Eastern Ukraine. A defended front-line fulfilled many functions, and this ten options model offers a partial guidance to alternatives to a fairly static front-line. #4 to #9 all have something to offer in this regard, and I found great depth in this specific topic. The Cold War's cordoning approach to land-based air defences in Central Europe was given up during the 90's in favour of superficially more promising clustering (which was still pointless with the available hardware given technological developments, but NATO air forces preferred to ignore that their air defences had become useless and kept pretending**). Cordoning at sea appears to have two revivals; the "island chains" talk regarding the West Pacific Ocean and a couple years ago there was talk about a revival of the GIUK gap defence line scheme against Russian submarines (which are mostly ex-Soviet submarines). Cordoning requires great strength, for the defender needs to make with a part of the cordon while the attacker can focus his main effort on this section, hence the success of breakthrough battles in late WW1 and in WW2. Mobile reserves to back this cordon up can help (such as ASW aircraft rushing to the area where submarine activity was detected), but this is prone to deception (such as feint offensives against land war front-lines).

Mobile warfare is relatively disordered and exhausting. It was normal in air war (where fighter patrols clashed with fighter patrols as long as neither side was clearly inferior and had to become more focused) and can be considered normal in land war and war at sea whenever the warring parties have few forces relative to the relevant area. This is the status quo of today for many land warfare scenarios. Hindukush, Greece vs. Turkey and Malaysia vs. Singapore are some of the exceptions (albeit the Singaporeans have an attacking strategic defence doctrine). We have still much military personnel in Europe (European NATO approx. 1.5 million), but the share of non-combat support troops has grown so large that the few combat and reconnaissance troops could not possibly meet the demands of a stiff cordon defence. The most pressing problem with mobile defence is that it's very much exhausting to move around a lot. Fuel and munitions resupply is an issue, but the exhaustion of the troops is the bigger one. It's doubtful whether we would feed our troops go pills as ruthlessly as done by the Wehrmacht in WW2 (Pervitin), and even that only extends the endurance from days to a few more days till complete exhaustion in a high threat and high mobility scenario. NATO currently seems to believe that it could approach this problem by simply adding fresh troops into the meat grinder. We would either have overwhelming strength if we "counter-concentrated" before the war started or we would feed one brigade (or battalion battlegroup) after another piecemeal to the theatre of war over the course of many weeks. The Russians with their huge geographic depth appear to be forced to do the same. I'm not seeing anyone having a good answer to the exhaustion problem. Navies have another problem with the mobile warfare approach; it's much easier to find (relatively) stealthy targets such as submarines when you focus your sensor assets on straits or a cordon than by hoping to find them in the vast oceans.

This leads to #5, hunting patrols. This seems largely pointless for today's naval warfare except maybe with ASW aircraft or against auxiliary cruiser commerce raiders. Surface task forces would be sunk by superior air power easily nowadays, so there would be only strike, not some hunting. Air forces don't have so much potential in this option either. Land forces on the other hand may be grossly neglecting this approach. The defensive reconnaissance post linked in part I already pointed that out:  A combination of #5 and #8 (shadowing) may be extremely efficient and promising at dealing with forces that infiltrated through a non-existing defensible front-line. The difference to mobile warfare may seem a bit fuzzy, thus a clarification: Hunting patrols are for the 'rear' area where the opposing force has no home advantage, whereas mobile warfare transitions rapidly between favourable terrain, unfavourable terrain and terrain that favours neither side. Shadowing is always done by inferior forces that merely need to possess the mobility and sensory ability to stay in contact and the means to communicate. This lends itself very well to "economy of force", doing the job in an area with as few resources as possible so the saved resources are available for advantageous use elsewhere. Now a bit more specific: Imagine a land war in which the opposing forces have little fortitude with mechanised battlegroups, but keep infiltrating with platoon- to company sized elements. Those (small) units do not get stopped by a defended front-line, and our reconnaissance is so undermanned that it cannot even meet its offensive-supporting tasks, much less provide a theatre-wide surveillance. Now let's add a cheap governmental militia force that reports such intruders, maybe shadows (tracks) them and some hunting patrols engage them once the conditions are favourable. What else could be done? Recalling multiple brigades to deal with minor intrusions? That's about what was done after Iraqi Fedayeen busted a single support small unit in 2003.

#6 Convoying; it's not  much of a thing any more. Sure, the USN convoys for its aircraft carriers and there are protected road convoys in occupation warfare. By definition all army vehicle movements with more than three vehicles are a "convoy". But we don't really protect civilian vehicles with convoys. The USN even stopped pretending that it could or would protect its own nation's strategic airlift ships. The notion that Western navies protect our maritime trade lanes is not about convoying at all (and is quite some nonsense in general). The Allies needed many hundreds of escorts in WW2, and we'd need just as many today if we were to protect maritime trade by convoying. Convoying doesn't really work for civilian aviation, either (except some occasional fighter flight escorting a VIP aircraft). It is highly questionable whether convoying is a sensible approach to secure the supply transportation between land forces 'in the field' and some forward depot. Convoys may actually make it easier to disrupt such transportation, as few convoys are more easily tracked and stalled than a multitude of smaller movements. Maybe - and this is something that requires experimentation - the dispersed seemingly chaotic movement of supply vehicles moving alone or in very, very small groups makes more sense.  Anyway, we don't have the assets to protect even only some main supply route movements with convoying during a major war anyway. All the assets developed and established during occupation wars are exceedingly useless against the very different major war threats. A MRAP stops neither a Su-25 nor does it stop kamikaze drones and it's a mere target to an armoured recce AFV with an autocannon as well.

#7 secured zones. This is in part about using favourable terrain to protect much with few resources, but the more pressing part is to protect high value targets that must not be struck. Let's call it out; Warsaw (and in Korea: Seoul). A major NATO capital that would not be allowed to see Russian armoured vehicles in its centre. How do we protect it? A relatively static defence (forces allocated specifically to the vicinity of Warsaw) would fix multiple NATO/EU brigades, which could tilt the balance where the real action is against them. So far I have not seen much regarding this issue, what I saw about Baltic defence scenarios was not paying much attention to the possibility of deep opposing forces raids. These are still a thing. Shall we bet on #4, #5 or #7 in this regard? We might inadvertently end up using #10.
I mentioned the integrated (area) air defence clustering before. One of my pet peeves is the concept of a 'bastion' where land forces could safely recover, where supply dumps would be safe. This would be a 'secured zones' concept, and it would require much area air defence, but camouflage, deception, concealment and dispersion would still be used. The ground raids threat could be reduced by pickets and defended river crossings. Such an approach was unthinkable during the Cold War when you had to keep in mind that a 100 kt thermonuclear warhead would render all personnel and material in 10 km radius militarily useless. We can think about such a fairly small secured zone nowadays, as the realistic defence scenarios aren't about global thermonuclear war, but rather limited disputes with limited objectives.

#8 shadowing (or 'tracking', but I prefer 'shadowing' because I really mean tracking from nearby, not from afar). Much was already mentioned about this under #5, but it should be noted that civilians may be removable from the theatre of war at sea and in the air, but not on land. Agents may track and report troops movements in conventional war similar to how Taleban motorcyclists shadow Western forces in occupation warfare. Furthermore, think about the man-marking in football, where one player stays with an offensive player of the other team to diminish his offensive dangerousness. The old-fashioned man-marking was largely replaced by a zone defence, where not one man attempts to mark a specific player. Instead, defence players are responsible for a zone and man-mark whatever dangerous offensive player enters this zone. It's one of my operations pet peeves that reconnaissance/scouting as well as counter-recce should be about assets allocated to defined areas with a particular level of ambition. To shadow and report intruders is already part of the lower ambition levels (the highest ambition level would require immediate elimination of all hostile intruders). Shadowing of intruders including reporting their location, strength, state and movement could (should) be a major building block of a land warfare doctrine. I just don't think that the run of the mill talk about information superiority yaddayadda comes close to what I envision. A defending nation (or alliance) should have a near-complete situation picture not because of fancy sensors, but because the 'blue' terrain is covered with militia (if need be "stay-behind) forces capable of moderate combat (mostly against support troops) and of course giving situational awareness reports. This recce should be area-bound, no manoeuvre forces commander would send a recce party ahead of a battlegroup movement. He should simply know what's there by the militia reports, enjoying a distinct advantage over the invaders.

#9 NATO was a bit desperate back in the 1970's. Its intel told it about huge hordes of Soviet motorized rifle divisions deep in Eastern Europe. The simple-minded conclusion was that this was spelling the doom of unstoppably many waves of divisions attacking the West. It did apparently not cross their mind (nearly enough) that this was a non-aggressive stance, suitable for a defence-in-depth and likely a sign of Soviets preferring quantity on paper over quality.***

(The widely-spaced dot landscape continued to the Urals.)
Its answer was mostly technological: Air-Land Battle, in which air power was supposed to bust bridges and massacre marching convoys deep behind the battlefield. Air defences and defending fighters be damned. Very low flying was meant to counter ground-based radars, but pulse-doppler fighter radars with "look down" capability had appeared in MiGs years before the specialised Tornado bombers entered service. It was an awfully technicized and expensive hope against the superior numbers. It was likely also utterly unnecessary and of doubtful deterrence value. NATO could and should have paid more attention to address the shortcomings of its land warfare concept. 

All this "deep strike" that already proved terribly expensive and just as much indecisive in Korea is still lingering strong in NATO air war force designer minds. Decades of bombing brown people and handily defeating their obsolete air defences added to their optimism, but hardly anything was ever gained by the West through those bombings. The biggest success was handing Kuwait back to the hands of some other dictator. The Serbs embarrassed NATO in 1999 when they used obsolete air defences but stubbornly refused to lose most of them and persisted for months. 
We (the West) should think hard about whether we exaggerate the usefulness of air power in depth. It's obvious that a near-peer defence could make very good use of defenders' advantages.
The one thing that I don't doubt about air power is that it handily knocks out rail traffic. Rail bridges in particular are a nightmare.

The fashionable way to think about infrastructure attack is to think about malware, of course. And yes, this could be a major issue for days or weeks. This could happen exactly in the most important phase of a conflict, the first couple of weeks. Nothing of this sort can stop a brigade on a road march, though. You don't need any electronics for that. Most military vehicles are too old to know the meaning of the word "internet" anyway. The two things that could really mess up such a march are fleeing civilians jamming the roads (this could be provoked by propaganda) and busted bridges. Which is why I place such an emphasis on military bridging for Oder and Vistula.

#10 Enduring the problem. This appear to be a most underrated approach. It runs 180° counter to the armed bureaucracies' self-interests (unless the activity required to counter the problem is super-unpopular, like convoying for civilian shipping or naval mine countermeasures), which alone suffices to suspect a systemic neglect of this option.
I'm suspicious of the extreme expenses required for BMD. Ballistic (and soon hypersonic) missiles are so darn hard to intercept that the interceptor is usually more expensive than the hopefully intercepted munition. I'm very much in favour of limiting theatre of war BMD to few small footprints. Feel free to try protect a couple super high value targets against all incoming munitions, but we should just endure the other (quasi)ballistic missiles. American cruise missile strikes and the bomb damage repair experiences of WW2 offer plenty evidence that you can very well live with much more direct hits than the the size of threat arsenals has on offer.
Another thing we should endure is "offensive cyber warfare". Every tracing back of malware attacks to a perpetrator nation is a scam unless evidence was collected in that supposed country of origin itself through multiple sources. There's some Russian comment text in the malware? Oh really? North Koreans never heard of the Russian language? I remind you they share a common border. All that talk of offensive cyber by evil-this and evil-that only serves as cover for the offensive cyber that is publicly attributed the Westerners (and frankly, that's not beyond reasonable doubt, either). The pursuit of offensive cyber abilities in the West leads to an undermining of "cyber security", and thus an undermining of our "cyber defence". In other words; stop the "offensive cyber" BS and stop pretending that we can attribute "cyber attacks". We cannot and offensive cyber is worthless BS. Let's endure the problem and focus on having proper security in our own systems instead of backdoors and encryption weak spots galore.

As long as part II grew, I feel I still couldn't properly convey the core message about how changing circumstances may lead us to stick to an outdated mix (structure) of our defensive posturing, with an improper emphasis on some options and neglect of others. I didn't feel like throwing this away, though. Maybe someday I can make my point succinctly.

*: To be honest; my model does not cover "Raumverteidigung", and it doesn't because it's an outlier that doesn't seem to fit to all three domains (land and air at most). I'm disappointed that no comment pointed that out before I discovered it by myself.
**: Patriot can only cover a 110° sector well due to its crappy radar design that was meant for a stupid SAM belt. IHAWK batteries were meant to cluster as 360° protection for a Patriot battery. NATO knew since the German reunification that IHAWK was practically useless in face of Russian ECM, so Patriot was still a mere target on a silver plate, just now isolated due to clustering rather than as part of a line.
***: Pre-WWI statistics told about an unstoppably superior mass of Imperial Russian divisions. The  Western powers looked feeble by comparison. The real Russian Empire of WWI proved to be one of the weaker great powers even on land, two leagues behind the German land warfare machine. This wasn't merely about lack of industrial base for heavy artillery and munitions. Officer corps quality, NCO quality, infantry training, morale and techniques were all inferior. The Red Army of WW2 didn't change much of this.



Ten options when facing a threat of specific geographic origin (I)


Today I shall present a model of how armed forces deal with a threat that has a geographic origin. In other words, a model that brings all of warfare into an abstract level and looks at it in a particular way. It's a framework of sorts for looking especially at inter-state war. Be warned; this is going to be long.

- - - - -

Let's use a simple case, the case of a fictional European naval war around 1900. Two navies oppose each other, one is supposed to raid commerce and coastal settlements, the other is supposed to defend against this.

The point of origin of the threat is one particular naval base. The attacker's warships should leave that base, find targets on or at the European waters and the world's oceans, harm or capture them and return. Rinse repeat. We'll see there are stages at which the threat can be countered (or "options with which the threat can be countered") by the defending party:

- - - - -

The first stage is to prevent the attacker's shipyards from building more warships. Air power was not yet capable of that in the scenario, but political and blockade measures to limit the availability of materials was possible. So this is the first stage of defence: Reduce the reinforcements and replenishment to the attacker's base of operations.

The second stage is to attack the base itself, or the attacker's fleet inside the base. This was doable at the time with a naval bombardment or small boats entering the port with explosive munitions.

The third stage is to intercept whatever attacker's forces leave their base of operations within a defined geographic area. This may be an offensive naval minefield, but it could also be an actual cordon of patrolling scout ships and a patrolling or sortieing battle fleet.

The fourth stage is to intercept gradually. It's a bit misleading to count is as fourth, for it could happen in front of the third stage, behind it (against ships that leaked through the cordon) or instead of the third stage. You may thus alternatively think of options instead of stages.

The fifth stage is to patrol the seas and hunt for commerce raiders with hunting groups far away from the attacker's base of operations.

The sixth stage is to protect maritime transportation with a convoying system, providing escorts to convoys.

The seventh stage is to secure certain important yet relatively small areas, for example at straits. This would not intercept the threat in general, just for a small area. Another facet of this stage is to provide coastal defences to coastal settlements.

- - - - -

All of these stages were used in naval warfare, often times many of them in parallel and thus likely wasting resources by lack of focus. Some examples are

(1) Most naval blockades had this effect, also diplomatic efforts to reduce neutral countries' deliveries of essential raw materials.

(2) Battle of Port Arthur 1904, Attack on Mers el Kebir 1940, attack on Alexandria 1941, attack on Pearl Harbour 1941, attack on the Tirpitz 1943

(3) Offensive minefields were used much during WW2, and the NATO maintained an anti-Soviet submarines cordon North of Europe from Greenland to Iceland to United Kingdom (GIUK gap). The Entente also laid an enormous naval minefield barrier between Scotland and Norway in late WWI after already establishing one against submarines in the Strait of Dover. The German minefields between Estonia and Finland in 1942-1944 also fit.

(4) This fits to Royal navy actions in the Eastern Mediterranean 1940/1941, and also to the American submarine campaign against Japan in 1943-1945.

(5) This is exemplified by the Royal Navy hunting groups for German commerce raiders in 1914 and in 1939-1941. The anti-submarine hunting groups in the Bay of Biscay in 1944 do also fit.

(6) Convoys were widely used, examples were the Spanish Silver fleets of the 16th and 17th century and naval convoying in late WWI and WW2.

 (7) coastal defences all around the world, Soviet naval "bastions"

 - - - - -

There are actually at least two more stages (or options). They just don't quite fit into the scenario, but I shall add them as #8 and #9:

The eighth option is to shadow attacker's forces as they leave port and go on a patrol. This is mostly feasible during peacetime (today's SSNs trying to shadow SSBNs, "fishing boats" shadowing surface warships) and becomes important at the start of hostilities, but it also happened to the Bismarck during wartime when inferior but fast enough British cruisers attempted to keep track of its course till battleships or carriers could engage it. The attacker side can also use this, example being the German submarines shadowing a convoy as "Fühlunghalter" in wolfpack tactics.This could be re-interpreted as a possible action of a defender against an invasion fleet

The ninth option is to cause harm to infrastructure so the attacker's forces could not operate well. This would be possible by damaging or blocking canals, but also by going after replenishment ships in addition to countering the raiders themselves directly. This happened to German replenishment ships in 1940-1942, and also to German refuelling submarines in the Atlantic during WW2.

- - - - -

The ninth option makes it obvious that all of this could just as well be applied to land warfare. Let's look at examples from the land domain:

(1) Allied air attacks on German armament industry in 1943-1945

(2) Artillery strikes on marshalling areas before frontline breakthrough operations. This was a huge thing in WW2 and often caused more casualties than the much shorter duration breakthrough operation itself.

(3) A static frontline with field fortifications, or a siege army besieging a fortress

(4) Counter-reconnaissance ambushes and patrols, mobile land warfare

(5) I mentioned an example of such defensive efforts here

(6) Protected land convoys in Indochina/Vietnam, Iraq, Afghanistan. Actually the Romans already lost legions in escort duty for food resupply convoys against the Parthians

(7) Security efforts at important bridges, at airbases, headquarters, mountain passes before they're part of a front line

(8) Mostly done by guerillas

(9) Scorched earth tactics including the burning of Moscow, destruction of rail lines, blowing up bridges or cratering of roads at bottlenecks with explosives during a withdrawal

- - - - -

I mentioned that the model will be applicable to practically all warfare, so here are examples from the air war domain as well:

(1) Allied air attacks on German and Japanese aviation industry in WW2

(2) Air attacks on airbases; Battle of Britain, Pearl Harbour, Six Days War

(3) Kammhuber Line, British heavy anti-air artillery dispositions in Kent until the V1 attacks, NATO's SAM belt of the mid and late Cold War in Germany

Kammhuber line - German night air defences WW2

(4) combat air patrols

(5) OK, this one is not  a very good case for historical analogies because aircraft have little endurance and extensive patrolling is thus difficult. The Fernnachtjagd of WW2 comes to mind, though. German nightfighters without radar flew offensive combat air patrols over Southeast England around 1941 to shoot down returning British night bombers.This does thus not quite fit into #3 or #4, as they targeted returning bombers. It's not a perfect match because the activity was close to the base.

(6) Civilian aircraft were as far as I know never escorted by fighters, but convoying has very much been used to protect bombers and support aircraft, from escorted photo reconnaissance aircraft of WWI to bombers through most of WW2, Korean War, Vietnam War and Iraq Wars.

(7) Point air defences, such as for airports or capitals (Baghdad, Berlin in WW2 and Moscow in Cold War were extremely well-protected with rings of air defences)

Mosocw air defence rings as of 1965 per CIA

(8) Shadowing was used against American bombers in 1943/44 to help track the bomber fleets over Germany.

(9) Attack on support aircraft (tankers, electronic warfare, reconnaissance), energy/physical attack on satellites and their signals

- - - - -

Let's now name the nine options with all three domains and much of history in mind:

  1. Economic attack
  2. Base strike
  3. Cordoning
  4. Mobile warfare
  5. Hunting patrols
  6. Convoying
  7. Secured zones
  8. Shadowing
  9. Infrastructure attack

- - - - - 

One final word before the summary lines: There is a 10th option. This option is to NOT counter the threat ("10. Enduring the problem"), but to endure it as the lesser evil compared to the effort of countering it. This may seem counter-intuitive, but it is and always was an option. I mentioned an example here regarding the non-necessity of ASW for the Italian Navy.

One could also come to the conclusion that all the craze post-9/11 did hurt the West A LOT more than the attack itself. Maybe doing nothing in return would have been suboptimal, but the very thought of choosing to NOT do something might have been a helpful thought and kept the West from harming itself in a myriad of ways.

- - - - -

These abstract nine options of countering a threat of specific geographic origin were thus shown to be very much applicable to the warfare domains of sea, air and land. They were observable through much of history.

Part II will cover some lessons that can be easily seen and drawn from this nine options/stages model.




Link dump April 2021





This is sadly believable to me, for it fits to how bureaucracies and especially LE bureaucracies appear to work all over the world. This case is but a minor extension on the ordinary. LE usually uses the mantle of a union speaker to manipulate public opinion.

It's telling that parties which supposedly mistrust bureaucracy (if not government as a whole) make an exception with those bureaucracies that can be used to bully people.

- - - - -

[German] "Mogeln ist ansteckend"

[German] https://www.tagesspiegel.de/politik/anti-transparenz-bei-der-cdu-wenn-datenschutz-zur-ausrede-wird/27004512.html

[German] https://blog.fefe.de/?ts=9eb7ee5f




About unguided torpedoes (addendum)


I wrote about unguided torpedoes kinetics in 2013, and after a couple years it dawned on me that I had missed an explanation for the use of early torpedoes on capital ships.

Books about Pre-Dreadnought and Dreadnought ships as well as armoured cruisers usually do not appreciate their torpedo armament as essential, I've even seen it called "useless" in subject-specialized books.

Those warships of the roughly 1890...1910 era did indeed have a torpedo armament that was unlike the torpedo armaments used later, or on light warships and boats. They used fixed underwater torpedo tubes, facing port, starboard, bow and possibly aft, though not always in all these directions. The torpedoes had an abysmal propulsion, light weight, small diameter, small warhead, contact fuse and an abysmal range.

They were typically only a few knots faster than the ship itself and ran for a few hundred metres only.

You can find data on such torpedoes here:


The archetype was the Whitehead torpedo, so let's look at the Royal Navy's model of the 1890's:

It had about 700 m range at 26.5 kts mobility and a 59 kg TNT equivalent warhead with impact fuse.

So why were they installed, and quite often so?

My opinion is that you can actually see by looking at for example the most iconic capital ship of that period in question:


HMS Dreadnought

Do you see the bow? It's no clipper bow by far, clearly not optimized for seaworthiness. It's a ramming bow, with a protrusion below the waterline like ancient Trieres had.

Now imagine one such ship trying to ram another (in case of the Dreadnought, such a scenario would be a night or foggy scenario, for it was meant to fight at long distances instead).

The targeted ship would typically be able to dodge such an attack unless it's already crippled. So during the dodging the two ships pass very closely past each other. THAT is when the port or starboard torpedoes could be used by both the attacker and defender. They would act as deterrent, even in small quantity and with a small warhead (which would not face much in terms of underwater protection against torpedoes in that period).

The forward-facing torpedo tube could be used to shoot at one possible dodging route of the targeted ship before the captain of the attacking ship steers to strike on an alternative other dodging route.

The rear-facing stern torpedo tube could be used by a ship running away from a dodging attempt, shooting at the pursuer as the pursuer cannot reach it with a torpedo.

Now why were torpedoes mounted below the waterline with all the complications this causes? Later era warships (1920's and later) installed torpedoes on the top deck, same as the torpedo boats of the 1890...1910 period.

The reason was probably that the top deck was simply too high. I have no data on this for the ship-mounted torpedoes, but some early aircraft-launched torpedoes (dropped from up to 30 m altitude) needed about 400 m to level to the correct depth. Let's assume a torpedo launched from a battleship would be stable on the preset depth after 200 m. That might very well be a longer distance than to the target during one missed ramming attempt. The correct depth was important for two reasons:

(1) too deep means no hit whatsoever, for the torpedoes of that time had contact fuses, not magnetic influence fuses

(2) too shallow means that the water pressure at the (small) hole in the hull would be rather small. The water would flow into the target ship much more slowly, and the sailors would have more time for easier countermeasures.

Torpedoes eventually gained much better mobility in the 1904-1910 timeframe, but they were very expensive munitions (kind of like really expensive missiles today; battery-electric torpedoes of the late 1930's were the first quite affordable torpedoes). The shipbuilding practices of the great powers did not adapt to this immediately, and they kept using the old underwater torpedo tube arrangements into the First World War.

A LOT more about unguided torpedoes (though not this above) can be found here:





Naval gunnery 1890 to 1945 and lessons drawn


Naval gunnery was in a dismal shape during the last decade of the 19th century. The old extremely smoke-producing blackpowder (not really a "powder" any more at that time) propellants had recently been replaced by smokeless propellants (such as the hazardous cordite in the Royal Navy) and breech-loaded guns were being adopted for good, but the gunnery and artillery ship design was in a horrible state of affairs.

Expected useful firing ranges were short (less than 4 km against ships), as optical range-finding was in its infancy and there was no central fire control yet. The coal-fired boilers produced terrible amounts of smoke and many guns were placed so low even on battleships that fairly normal rough seas made them temporarily unusable as waves crashed into them.

The typical armament of an armoured ship of the time consisted of two or four large calibre guns (8"...13.5" calibre), with a substantial secondary battery of  usually 4...6" calibre and soon some smaller guns specifically to shoot at the small (about 200 tons) and nimble torpedo boats.

There was a break in rate of fire between 6" guns and heavier guns because 6" has the heaviest shells that can be loaded by hand throughout a long battle by the non-roided men of the 1890...1945 period. The heavier guns used more machinery to load shell and propellant and this was slower, not faster (and still is, though it's faster than manual loading for the modern naval guns of up to 130 mm).

This huge drop in rate of fire from about 5...15 rpm for 5...6" guns to 1.5...3 rpm for most heavier guns was in part cushioned by the reality of long-range gunnery post-1910, but this rate of fire and the huge cost, size and mass increase past 6" calibre led to a curious situation at first:

Ever since the 1860's there were many warships with a primary artillery of few impressive big turret guns and many more smaller and somewhat hidden casemate guns up to 6". The public perceived the primary guns as the defining firepower. Those guns may have penetrated heavy armour protection on other ships, while the smaller guns did not. The reality was almost certainly known to most warship-commanding naval officers and naval gunnery officers: The secondary guns were the main armament and the big guns were for show and possibly useful in coastal bombardment, or as a deterrence.

The small guns fired so much more often that they were bound to score about tenfold as many hits at the then-relevant combat ranges. The heavy guns fired so rarely that corrections after observation of fall of shot by the gunner were obsolete by the time the heavy guns were ready to fire again.

Smaller than 8" shells were also less likely to be duds when they hit some unarmoured part of a ship where a very large calibre shell fuses often failed to trigger even as late as during the Second World War.

Royal Sovereign (1891) 13.5" twin barbette
A most egregious example was the Royal Sovereign class, which had four 13.5" guns in twin barbette mounts. Those were unarmoured to save mass high up on the ship, to improve rolling behaviour and general seakeeping. (The ships of this class still rolled badly when introduced.)

These 13.5" guns fired one broadside salvo every 135 seconds only, (0.44 rpm), as they could only be reloaded in fore-and-aft position. They might not have fired more than four or five shots per barrel in a battle, for the unprotected crew would likely have been cut down by fragmentation and even shrapnel (that was still a thing in navies at the time and has recently kind of returned with the 35 mm Millennium gun and its AHEAD munition) by many small calibre hits at that point. Meanwhile, the Royal Sovereign's protected 6" gun crews could have continued the fight with 25 or more rpm in total (port or starboard).

You may think that those 6" shells wouldn't do much damage, but most parts of any warship were unprotected or at most protected against fragments and shrapnel. Fires could be started (there was still much flammable material and even oil lamps in use) and the heat and smoke could render the burning ship combat ineffective (secondary fires could even doom it).

Most direct damage done - even by hits of the heaviest calibres -  would still usually be superficial, as the shells up to the First World War period usually used impact fuses without the delay required for explosions where they hurt the most (boiler rooms, turbine rooms, magazines deep inside the  hull). Secondary explosions were rather flash fires in turrets or where turrets and their munition hoists if not magazines were very close to the side of the hull. The extremely delicate boiler rooms with all their pressurized containers and pipes were so rarely harmed that even badly-hurt ships could make it to home port if not finished off by a magazine explosion or torpedo hit.

A bit more about rate of fires; by the Second World War, these had been improved to maximum slightly more than 3 rpm for (German) 15" guns, about 15 rpm for some (American) 6" guns, about 20 rpm for (American) 5" guns and also for 4" guns. Army 3" guns had reached about 30 rpm before 1900 already, but many early naval 3" mounts fell well short of such  rate of fire. The practical rates of fire on a warship in a gunfight were mostly MUCH slower.

- - - - -

Let's pause with this military technology history wonkery and look at a still-relevant lesson:

The public was very much fascinated by and interested in its battlefleet (at least the middle and upper classes, especially in urban areas). Such huge armoured ships did cost a fortune each, and were a heavy fiscal burden even to the richest countries. Yet all this attention did not reveal the terrible flaws of the capital ships of the period, much less lead to civilian oversight pressure to work on those deficiencies. That push came largely from some high-ranking officers who especially in the Royal Navy thought the time was due to become an actual fighting force again. (The newcomer German navy had been in technology catchup and modernisation mode anyway and no such awakening experience.)

- - - - -

Thus improvements were made after the Royal Sovereigns; later capital ship primary artillery always had properly-protected turrets (many gun crews in the 4"...6" gun range especially on cruisers remained inadequately protected even against fragmentation), and a long time after (after the First World War) the last navies adopted proper and high-enough mounted secondary artillery turrets instead of casemates. There were also experiments with really high-mounted casemates in the superstructure. 

The biggest change happened with regard to fire control and quantity of the primary artillery guns. The individual aiming was hopeless at long ranges, and thus fire control became centralized with gunners in turrets merely following orders about elevation and training as well as shooting the guns.

It was understood that simply giving more gunnery practise did not suffice; you needed to observe the fall of shot and correct accordingly. This required salvo fire and all guns aiming the same (or, if spaced much, at least using the same gun elevation). A mast top-located observer would see whether the water fountains were too far left, right, long or short and issue corrections. Eventually, the salvoes would straddle the target (shells impacting on or around the target) and the salvoes would continue without further corrections at highest possible rate of fire until the salvoes were observed to be off again or the target turned.

Mechanical fire control computing aids were introduced, communication devices were needed between fire control and turrets and optical rangefinders were introduced and improved until they could in theory measure distances to the horizon. Such (except the computers) was the art of gunnery by the First World War. This centralised aiming and firing depended so much on observation of the fall of salvoes that a rule of thumb began to dominate warship design; it took a salvo of at least six shots for a proper observation, and the guns better not be spaced very much (this led to a few capital ships with all primary artillery on the forecastle; Nelson, Rodney, Strasbourg, Dunkerque, Richelieu, Jean Bart). Some capital ships had 12 primary artillery guns and were able to shoot at two different targets with full fire control process simultaneously. This helped alleviate the issue of long spacing between turrets. Many dreadnoughts and super dreadnoughts were on the other hand unable to fire with a full 6 shell salvo in (nearly) all directions, although this had already been achieved with HMS Dreadnought (and the prototype battlecruiser HMS Invincible shortly after). They were thus rather a kind of ship of the line, optimised for broadsides only.

The so-called Pre-Dreadnought battleships with their mere two or four primary artillery guns were thus obsolete with the rise of centralised aiming and firing, and the appearance of HMS Dreadnought in 1906 is commonly considered to be the signal for this paradigm change. 

Finally, by about 1910 there was no reasonable doubt any more that the primary artillery was actually the main armament of battleships. The other guns turned into self-defence weapons until battleships were turned into heavy (anti-air) escorts by the rise of the aircraft carrier in 1942 (then the secondary or even tertiary guns in the 40 mm...133 mm range became the main weapons).

Now back to why large calibre gun rates of fire weren't THAT terrible during the World Wars as it seemed (despite shooting even more slowly in practice than nominally able to): The cycle of flight of shot, observing, reporting, calculating, transmitting directions to turrets, turrets training and elevating according to directions and finally firing a salvo together took longer than the reload even of a 16" gun. Practical rates of fire were even much worse than nominal ones. The Bismarck was capable of shooting its 38 cm (~15") guns in 18 second intervals. That should have sufficed for about 45 salvoes in its battle with HMS Hood and HMS Prince of Wales. It fired off 13 salvoes only.

The nominal rate of fire was really only important at short ranges that required no observation and corrections, such as night combat at ranges of less than 4 km. Such ranges also devalued armour very much; horizontal armour would not be tested and vertical (belt) armour would be more easily penetrated as it was meant to protect at a greater distance only. This quick fire superiority scenario of a short range night battle allowed the really quick-firing American 6" guns to shine in the Guadalcanal campaign. Even the Japanese battlecruisers were inadequately protected against 6" shells due to their narrow belt armour.

(Now keep in mind that a ship such as the HMS Royal Sovereign of the 1890's would have made full use of its rate of fire, for it would not have the shoot-observe-correct fire control cycle. It would have made full use of 2 rpm rate of fire if it had had such a rate of fire. Its 0.44 rpm broadside rate of fire was thus exactly as bad as it sounds. The 2 rpm of some First World War-era main guns was not as bad, as the fire control process slowed them to less than 2 rpm during most of the battle anyway.)

- - - - -

Fast forward to the Second World War period: What was the peak of warship gun duel gunnery then?

The peak in daytime would have been like this: An aircraft or small and nimble escorting destroyers would lay a smoke concealment to blind the enemy fire controllers (see excellent collection of photos here). 

The own ship would use its radar and fire control computer for a first salvo or set of bracket salvoes, and a spotter (float)plane would issue the corrections by radio instead of the ship's own observers in mast tops. Top quality radars could even detect the impact slashes of big shells and shells had dyes of different colours so fire controllers could tell which salvo had been fired by which ship. Depth charge throwers ("K-guns") could be used to deceive enemy spotters to believe that the target was straddled when it was not (this was greatly complicated by the use of dyes and dependent on seeing the salvo being fired). It was also possible to group the artillery of one calibre into two or three groups that shoot at different elevations in order to quickly find the correct elevation by observing which impact group was more close to the target (bracket fire). Knowing the correct distance to the target and even the two ships' movement vectors accurately and doing all calculations correctly (including coriolis force correction) would not necessarily suffice to get the elevation right on first try: The temperature of the propellant, barrel temperature, air temperature, air humidity, how much the barrels were worn and wind also had an effect. A radar-only fire control without impact splash observation would have failed at medium and long ranges with the state of the art of even 1945.

Secondary artillery could use the same sophisticated fire control process as primary artillery.

The targeted ship would make evasive manoeuvres when straddled, degrading or resetting both the fire control process against itself and its own fire control process (if it has any and isn't completely blinded).

This combination of fire control and countermeasures rarely if ever happened in perfection. Smokelaying is tricky in windy conditions and some navies had good-enough radars only late in the war (Japan) or never (Italy). The battle in the Java Sea saw Japanese cruisers using the smoke+spotter plane combination, though. Radar-controlled fires were used as well, albeit sometimes with surprisingly bad results (the Bismarck was very poor at getting the elevation right, for example).

- - - - -

An aircraft carrier as a mere support ship without any torpedo bomber or (dive) bomber could still have had decisive impact on a battlefleet engagements of the 1930's by providing smoke, spotters (replacing the more fair weather-dependent floatplanes and flying boats) and fighters to chase off or down enemy spotter planes. This begs the question why some navies (Germany, Italy and mostly also France) neglected the aircraft carrier so very much. It was already a decisive asset without high performance aircraft (Morse radios/wireless telegraphs were already tested in aircraft during the First World War).

- - - - -

Yet again, I saw no trace of public pressure on navies to correct such a deficiency during the 1920's and 1930's. The naval discussion in Germany was about the expenditure for the Deutschland class (the later so-called pocket battleships*) until dictatorship took over and public discourse was muted. A similar excuse can be given for Italy, albeit as far as I know their dictatorship would have tolerated pro-carrier enthusiasm  while censoring negative critique. Their navy was barred from having aircraft in favour of their air force, so I suspect their naval top brass simply passed on a critical naval asset because it wouldn't have been fully theirs (the carrier aircraft would not have been navy-operated). I see no excuse whatsoever for France, which had a carrier arm that badly stagnated with the Béarn**.

So why were such severe shortcomings not corrected under public pressure in general? It wasn't just the dysfunctionality of dictatorships. The story of anti-air artillery during the 1920's and 1930's was a farce all around the world. The ridiculously poor quality and also ridiculously poor quantity of anti-air guns in that period was in stark contrast to the known threat of torpedo bombers of the Interwar Years. Their heavy and still very short ranged (low muzzle velocities) medium anti-air artillery was even too weak to protect their own spotter planes from enemy planes even only overhead themselves. Even ships such as heavy cruisers were often only equipped with four weak heavy anti-air guns and a couple machineguns. Moreover, several exercises thoroughly embarrassed navies, as they failed to hit the aerial target for extended times.

It's a general problem with (relatively) highly technicized armed forces in peacetime. Warfare against peer or better opposition may reveal their deficiencies, but most of the time armed services can hide their shortcomings behind the veil of secrecy.

It's a cautionary tale regarding how much top brass can be trusted by the public.


 In kind you want yet more video about naval gunfire control:

*: The Spanish España class deserves this title much, much more.

**: Which also had poor quality aircraft during the 1930's.  



Drone art of war


Drones are becoming smaller, lighter, cheaper for the same performance. They're also very likely to become smarter - smart enough that small, cheap, quantity-produced drones will exceed today's autonomy of killer drones and be able to identify and classify targets and make not only an attack decision, but also tailor an attack manoeuvre to the situation.

Drones will not only gain time to find targets by loitering, but also by simply sitting on vantage point from whence they can approach the next location or a target by ground or by air. There might even be drones that will use inland waters to hide.

What's going to be the art of war in an age of autonomous killer drones?

(1) A seemingly eternal rule in war is that no novelty is total. Never did one novelty become the only way of fighting.* Armies did not transform into all-machinegunner armies. Not all field army vehicles are tracked and armoured. Submarines did not become the only warships, Jules Verne's opinion didn't matter. Missiles did not replace aircraft fully. It's in my opinion a good call to expect autonomous war (killer) drones to co-exit with traditional forms of warfare.

(2) Specialised assets rise as a form of warfare becomes more sophisticated, but generalist assets prevail and in parallel. The benefits of specialization are huge, but limited and the uncertainty about the face of the next war makes versatility an important risk management approach.

(3) A look back at the early days of military aviation may help to anticipate how the flying autonomous drone armies will play out:

First, planes were used for observation and a little ground attack. Then the countermeasure of fighters appeared. Observers were armed for self-defence, and some evaded the fighter threat by flying very high. Bombers were armed as well, also armoured later, some also evaded to greater altitude or into the night and generally they received armour. Fighters responded by flying faster, higher, getting more armour, more firepower and sometimes sensors for night intercepts. Specialist aircraft appeared in WW2 for electronic warfare and pathfinding at night. Eventually, there were dedicated electronic reconnaissance aircraft, area search radar aircraft and tanker aircraft.

We might very well see reconnaissance drones that are semi-autonomous (radio link to user preferred, but able to scout autonomously until radio link is re-established) first, soon complemented by optionally autonomous killer drones of varying sizes.** Fighter drones with optimization for finding and killing other drones will appear, and the killer drone swarms will have to adapt with self-defence (not so easy for kamikaze drones, they can mostly counter by being cheaper than the fighters), better camouflage and concealment, more speed, agility and acceleration. 

Specialist drones may appear for radio relay function, electronic reconnaissance, electronic attack, maybe even entrapment (hunting other drones like spiders with webs). Picket (early warning) drones may also come into existence, forming a detection corridor and a screen around human troops. Confirmation drones may appear that hurry to inspect and confirm a suspected hostile contact with better sensors than kamikaze or even sniper drones have (or by going real close while being cheap). Command & control drones with superior AI may serve as forward commander to a swarm that's got no reliable radio comm to a HQ with humans.

There might even be civilian interaction drones that advise and guide civilians to safety, maybe even prisoners of war.

Drones may differentiate into high altitude drones (flying so high that no missile or gun can economically kill them), above-treetop altitude drones, below treetop drones and drones which can even enter buildings and do their job indoors.

USAF strike packages often had no more than 40% of their aircraft carrying munitions to strike the actual target. All else was support and escort. Drone swarms may in my opinion range from 10...90% support, and the difference would come from the task; a terrain control swarm would have few strike drones, whereas a main effort swarm meant to annihilate large forces in a small area might 'zerg' with huge quantities of strike drones either saturating or held in reserve.

(4) It would be nice if we would simply ban the use of autonomous killer drones like NBC weapon use became taboo, but so far the governments focus on what advantages drones offer to themselves. Loitering munitions with partial autonomy are blurring the transition to autonomous killer drones.

Armed bureaucracies have a great potential for conservatism and sluggishness and are not good with money. They have a high risk to resist de-humanisation of war out of selfishness and overestimation of human capabilities. They also have a high risk of focusing resources on basic types of drones (scout & strike), and becoming overmatched by an aggressor that invested in more sophisticated swarms. This may even happen after they understand the issue, as pre-war funding for a drone war is all but guaranteed to be too small given the establishment self-interest of other force structure components.

(5) Force structure and tech isn't everything. Concepts of operation matter greatly. 

Drone swarms could span the entire theatre of war or be segmented into multiple swarms. Either way, rules for interaction, cooperation, coordination and geographic limitations are necessary. Rules of engagement would be all-important.

You could also apply the concept of main effort, for example to gain supremacy in a geographic bottleneck, or at your capital, maybe at an important port or to gain dominating heights for heavy ground-based electronic warfare equipment to use. Too great concentrations of drones might be countered with area effects, especially EMP.

We know how to give orders to human troops, but the best technique for giving orders to more or less (semi-)autonomous drone swarms may be very different. I like to think that my own idea/concept about zones with different level of ambition (regarding tolerance of hostile presence) and different allocation of resources would work well.

(6) There will be drone war prophet celebrities that will develop and incessantly repeat catchy buzzwords and simplifying models. Some of them will be Americans, earn some good money with books and lack substance at closer inspection.

*: Firearms kind of achieved this, but it took many centuries and wasn't anything close to what some early firearms proponent of the 14th century AD might have envisioned.
**: Meaning general target drones, not those specialised on radio frequency emitters. Those already exist.