Armies and alternative units

[Cassius Critzer]

I've been modding the Sengoku Jidai 2.0 for personal use for the last ten days. It seems to me that two things would most benefit the unit composition.

First the AI badly underdefends its cities and thus loses them unrealistically. Thus I altered some units to have no movement and higher defense to act in a garrison capacity. But in addition, this is not purely comprised of one unit, but mixed composition to achieve that goal.

Secondly, there are no armies in CIV2 yet they are sorely needed. You don't dispatch certain units to attack and defend when dealing with giant armies as occured at Sekigahara. Recall that Japan coud field 2-3 times the number of soldiers that France could during the same period. While Sekigara was raging with 200,000 soldiers, there were other battles simultaneously going on which caused delays in the soldiers arriving to Sekigahara.

Using the standard that a:
Platoon equals 15-45 soldiers
Company equals 80-150 soldiers
Battalion equals 300-800 soldiers
Regiment equals 1,000-5,500 soldiers

Then the following creates battalion level armies from platoons and companies of mixed units jointly forming to fight in some realistic manner.

So I considered that mixed armies could be created with defensive armies might have ashigaru totalling 500 men, then 150 ashigaru archers, then 25 no dachi infantry shock troops, and 20 yari cavalry. All in all totalling 695 soldiers. Similarly creating 640 soldiers in an offensive capacity. This would be occur after certain technology was researched and at great cost. In one example, a largely developed city could make such units after 13-15 turns while they would otherwise commit to making individual units of 15 or so in number to oppose such an army. One using orchestrated tactics and might have special abilities to manuever and have scouts to coordinate manuevering. The other would mixed tactics to defend and oppose the army.

Such armies make defending a city realistic. Adding terrain defensive bonuses can make taking such cities extremely difficult and why field battles were so prevalent as beseiging a fortified city might take many months or longer.

Then at higher levels, a medium army is achieved by technology with better stats and abilities, and a superior army at even higher levels. Both similarly are constructed to fill either a niche for defense or offense.

 

[Cassius Critzer]

Consider the most excellent map in Sengoku Jidai 2.0 of most of Japan excluding Hokkaido. Japan roughly is the size of California, so let's say minus Hokkaido it's 3/4 the size of California. With single armies, then these mixed units are in the same map coordinate point and do not risk the dreaded stack kills. Individual units coming to oppose the army are scattered over a hundred miles over many map coordinate points. It may seem like the individual yari spearmen and no dachi and heavy cav can easily envelop the single army,but it's largely offensive and defensive values make it more than a match for them.

The individual units can act like hyenas taking down a lion. Sure most of the hyenas will die, but they each have movement points that far exceed the armies limited number. They then can take a bite from the lion when carefully coordinated by ganging up on it.

With its very high stats, the army can take a very defensible position and intimidate nearby cities and there is little the individual units can do but take up similar defensive positions and then impede the progress by screening. Defensive individual units can block and engage the army and then tie it up. That limits the army's manueverability even more and then allows offensive units to use guerilla tactics on the interloper.

 

[Cassius Critzer]

Consider this. Civ2 is a strategy game over a vast region, often the world, yet is using tactical units of single composition to achieve victory. In fact, a very interesting squad level game might consist of using certain little utilized terrain like glaciers and making them very high shield and food and defensive bonus tiles. Then make the map into a realistic battle map.

In other words, have seven single cities or more on the periphery generating squads of tactical units fighting some famous historical battle like Agincourt. The single cities are the battle tents where soldiers get healed and from which they are deployed. The inner map is where the famous battle took place and considering alternatives where it might have changed the outcome to engage from there instead.

Thus you could deploy tactics like flanking, wheeling, pinning, pincher moves, etc using these single composition units. This would make more sense then individually fighting one on one battles to take over a city.

An army has a synergistic strength from working together that obviously is stronger than single units of one composition. It makes zero sense to send a cavalry unit to beseige a castle/city as that would be futile,but on a battle map of a realistic size, then cavalry is essential.

 

[Cassius Critzer]

Here's one problem of armies. It isn't really much different than the logical flaws built into the game engine. An army of various defensive and offensive strengths and movement rates and special abilities all have a support cost of one. Well that is absurd. If the turn base is set to a year duration and the construction time is 13-15 years on the most fully optimized city, then still once the army is built it has a support cost of ONE.

Meanwhile the fully optimized city can construct 13-15 heavy cavalry level 3 in the same time period in Sengoku Jidai 2.0. But that means support costs of 13-15 units. Interestingly when tested, the AI elects to make either individual single composition units or various armies depending upon the pressing need. And even though it could make offensive armies it largely built defensive armies. In testing, I made the armies to have a limited movement rate of one as early on the lack of roads in 14th century Japan would impede deploying the army. The most advanced armies might have movement rates of 2 or 3 and can cross arduous terrain like mountains.

Now consider that in the vanilla game making a spearman or a nuclear submarine both have a support cost of one as well. It's not much different, right? For example, the crew size for an Ohio class nuclear sub is 155. The crew size for a Nimitz class aircraft carrier is over 6,000. See the valid problems with support costs but a battalion sized army of 650-695 is nothing by comparison and significantly adds to the thrill of the game, especially one focused on developing the arts of War.

In history, one could successfully evade as a ronin (masterless samurai whose leadership had been defeated) by running up into the mountains and entering the domain of another clan, as it would largely take too long to orchestrate an effort to catch him, and not send a whole force as that would weaken the defensive abilities of a city. By doing so, that city could then be attacked while on the fools errand.

 

[Tanelorn]

Interesting.

 

[Cassius Critzer]

The only way I could make it work versus the size of the map and movement rates, was to alter the rules to gain enough movement but then still reduce movement in the terrain editor, but then meticulously create roads across Japan.

Then there are now three seasons per year so units and armies can be recruited in a logical length of time and have proper movement rates to actually traverse appropriate lengths of Japan and have actual armies end up fighting each other in central Japan as happened in history. It only works for patient players wanting a game mod with 400 to 600 turns. Do the math. If you do not watch battles being resolved but only watch your own, that is a huge time commitment.

The problem is how long mod takes and tech acquired over many hundreds of turns and numerous battles. The AI loves defensive and offensive armies with very high values and that takes awhile to defeat as they are so strong. Playing it through, one clan has two defensive armies and the other has one and has almost created a second and then the attacker wins...just barely. That leaves that newly conquered citadel open for a new attacker or the defender to redeploy and take it back.

By adjusting the naval defensive values to reflect "being boarded", now the defenders can often win but be so crippled and down to a scant value and should head back "for repairs". The AI is foolish and seldom understands this.

 

[Cassius Critzer]

I was very naive and began the mod in 1125 and planned to go to 1750 at three turns per year and it was GLACIAL to meticulously create all infrastructure, economy, and defend each new city, while the AI fought stupid battles when the Japanese map is huge (despite being only the size of California). The only logical way to have units was to start in 1448 with all the economy and infrastructure and base units created. Only then could I make it work.

Civ 2 is just not designed to actually develop nations of 25-30 fully built cities with seven different nations. Instead the AI will battle it out and therefore not develop armies even with a great tech tree. You essentially have to build the base to overcome the flaw. It won't build complete infrastructure and will have perpetually rioting cities.

In my opinion, it actually works better for unifying a nation with seven clans plus barbarians instead of a global map because few nations other than the USA, Russia, and China have adequate land mass to make up more than 10 urban regions. Great global mods work best when infrastructure is already developed and the economy with adequate soldiers FIRST so the players can fight.

Some mod like Ancient Rome only works by conquering weaker less technologically advanced rivals or vice versa destroys a bloated weakened late Rome that can be neutralized.

A global map logically makes no sense versus redploying armies and holding whatever is gained.

I don't think even the most ardent "turtler" would enjoy such glacial play over 700+ turns. At least start with each nation having 20+ cities. That leaves 36-20 so 16 new cities to turtle as 7 nations trying to create 255 cities is the maximum. In a perfectly balanced mod with a large map, then 36 cities are fully built with most infrastructure, then epic battles using both units and large powerful armies can work. That would take a minimum of two weeks to clone the cities with the bulk of time on the 21 grid terrain for each plus roads to connect them properly.

 

[Cassius Critzer]

Say you have a late era army with very high defensive and offensive values, hit points and firepower and a logical movement rate per the turns per year. Since each city has a maximum of 21 tiles to generate trade and shields and food, then a smart commander can send low value units to "take one for the team". 20 units occupy the land and ocean terrain. This causes "attrition" as each turn nothing is being generated as food or being constructed nor income. The AI fights with its strong army by taking the field and using up its movement points. Now the player sends in their army to defeat it in the field and take the city.

So ideally the AI generates two defensive armies plus individual normal units and recruits an offensive army. That is a superior mod. The problem is the AI is likely to uselessly waste individual normal units by en mass attacking the individual units of another nation before it develops at least defensive armies. Then just like history, a third nation laughs and clobbers the two foolish ones. The AI, if not properly set up, will expend massive casualties with its individual units by head on combat.

Meanwhile the AI being not entirely stupid will "disband" individual units into the newly recruitable army to rapidly assemble a defensive or offensive one. I've watched this happen plus when when the AI has a robust treasury, cashflow, infrastructure, then it will outright buy an army when a city is imperiled by an approaching invader.

It then will recruit low level fast defensive individual units to sort of protect the originating city.

 

[Cassius Critzer]

The outstanding effect of dealing with the technology of one nation versus several means having very nuanced unit progression. Typically unit progression is rather generic because of dealing with multiple eras plus nations so there are very few slots unless each nation is essentially creating the same units which is not historical whatsoever. But in a focused historical mod based upon the time period, then if many hundred turns, you have low level drafted farmers all the way to semiprofessional to professional to shock troops to elite units. That is marvelous.

In such a focused mod, then just as you have armies of various attributes, you can have many different kind of navies with escorts albeit limiting how many units can be transported within that fleet. This eliminates the annoying stack kill issue and yet then you have huge defensive and offensive values as an amalgam.

And think, this honestly is no different than higher value units whatsoever. It also helps the foolish AI who fails to return to port for repairs.

You could have specialized mixed infantry units (both highly defensive nagamaki wielding samurai plus shock troop nodachi samurai) with shorter movement rates as a specialized battalion or a mixed horse archer and heavy lance cavalry and mounted infantry as a battalion as well.

In the modern and postmodern era, tanks battles are not one on one but multiple tanks engaging multiple tanks. The same is true for infantries. Armies and fleets helps depict massive scale battles from history. That goes for ancient Roman battles in Gaul as well.

Or in the industrial to atomic age have bombers with escort fighters. Or you could have a naval task force with destroyers and aegis cruisers. No more naval stack kills by trying to coordinate a group of 5-6 ships as one unit is simulating all of them with very high attributes and essentially calculating fighter pilot escorts. Then that naval task force carries bombers and or a full compliment of fighter pilots in a transported air unit.

With mixed armies and fleets, and especially with limited slots, then you could simulate the Imjin War involving China, the Joseon dynasty (pre-Korean), and samurai clans of Japan versus struggling to represent true units from that time period. When actually it's superior to have them jointly coordinating the attack and defense with high value units.

 

[Cassius Critzer]

Besides the AI being smart enough to disband low level units to build their first defensive and offensive armies, the AI also realizes in a battle rich environment that the chance of success using merchants (ten percent or less so a bad return on investment for a trade deal that would take take 20 turns to achieve) is small when dealing with true urban regions with high population with armies. So it seems the AI is disbanding merchants to use those shields to hasten army completion just as it would do for wonders.

 

[Cassius Critzer]

After letting the fully populated cities build up adequate storehouses of grain, having the economy humming, prebuilding proper infrastructure and adequate soldiers, and especially if the terrain and roads are fully realized with adequate fortresses, since the AI chooses to "sally out" anyway even with defensive units, then these fortresses ensure their survival.

Then as long as they have both high offensive and defensive cavalry with proper movement points, the offensive cavalry (heavy cavalry 2-3) will "dash through" enemy neighbors to meddle with shield production. That is smart. Then the defensive cavalry (imperial horse archers) can even sit on an enemy fortress that is unocccupied. I saw both happen over and over. With highly defensive infantry (nagamaki samurai and naginata samurai 2 and 3) and shock troops with high offensive values (nodachi samurai 2 and 3), the AI properly occupied enemy fortresses that were unoccupied and on highly defensive terrain. They are realistically brazen.

Then with armies being researched and constructed with naginata samurai 2 and 3 internal to the citadel [usually at least 3-4 but often 5] , they protected their own lands and even in frontier regions had massive peasant infrastructure going on and protected by highly defensive infantry. I was glad to see that. In this case they often used naganata 2 and 3. They were aggressive and didn't keep offensive troops "bottled up".

Sengoku Jidai already had musket units that could settle and/or do critical infrastructure, so I made the Yari Ashigaru 3 do that too. This greatly facilitated this way earlier. This actually is correct as ashigaru were not actually samurai but farmers looking to send extra sons to maybe earn some money and return or end up being promoted as very low level samurai.

Meanwhile they all build the robust defensive army that may serve in an offensive capacity and hurry that along. Starting in 1448, I predicted it routinely happening by 1473 (16 x 3 so 28 turns later), but some achieve this as early as 1457, then by disbanding early units and merchants, then by 1459 have them.

They seemed to do this without tech trading too which is exactly what I want. They are sufficiently hostile so less likely to trade techs.

 

[Cassius Critzer]

The way Sengoku Jidai is originally configured, Democracy is based on "reformers" which actually happened late after a very established Tokugawa Shogunate. The mod starts with Feudalism and progresses to a Shogunate (Republic). In my opinion, the AI foolishly gets nervous and switches back and forth between Shogunate (Republic) to Reformation (Democracy) to sometimes Fundamentalism (which actually is the most stable naturally as there were rigid Buddhist coupled with simultaneous Shinto and Confuscianism.

Fundamentalism has a minor effect and is far superior due to low corruption, whereas Democracy has horrible corruption losses to trade even with heavy infrastructure to offset it.

The reformers truly are happening post1750. I'm not sure that democracy has a place in a Japanese mod of this time period except with the UN wonder and peacekeeping by a true Shogunate like the Tokugawa.

 

[Cassius Critzer]

Scratch that.

The way the governments are configured currently follows.
Anarchy during transition

Despotism

Feudalism (the beginning government as Monarchies ie essentially barons)

Shogunate rapidly follows which is Communism.

Fundamentalism ends up happening just from common study of budo and bushido. It is an absolutely fine government with no corruption and a clever human player can use "sensei" specialists to offset the AI using a penalizing tax rate.

Reforms which is actually a Republic...and not historical until about 1750+ AD. [Watch Twilight Samurai as the protagonist genuinely is concerned his wealthier friend is caught up in political turmoil due to potentially being labeled a "reformer".]

Then in 1542 AD the possibility of acquiring various Christian technology from the Portuguese and largely as a consequence of firearm technology. That is Democracy. This actually is interesting as it did cause political instability. [The dirty historical secret is Japanese slaves were sold to Portuguese and exported to colonies in the Philippines and this incensed Oda Nobunaga). That ends up causing the Christian purges under the Tokugawa Shogunate ]

No other nation on the Earth had more firearms than preTokugawa Shogunate Japan. They even trained female teppo units. While the archers greatly exceeded them in accuracy and firing rates, the teppo blast caused FEAR and Dread and penetrated most armor. There actually are preserved samurai cuirass with big indentations as armor was tested and then naturally battle tested.

Teppo at Nagashino caused a decisive victory as the units were not in the open but fired behind bamboo bundled together against a fence grid that idiotic cavalry charged toward and got mowed down.

 

[Cassius Critzer]

The mod is set during the Sengoku Jidai period which begins after the Muramachi era ends and in history when the Onin Wars began. Then it goes forward through the exposure to Portuguese technology in 1542. That breaks the stagnation that has halted Budo (war tech) and allows both better navies and armies which will cause significant potent armies and a really strong ship to emerge. The armies and new units consist of two these of firearm units of various levels of ability (three levels).

And since there are artificial limitations implied, though there was persistent war, it wasn't a time of wasteful aggression just for the sake of aggression. It was a time of realistically building up sufficient military power to repel invaders and develop enough soldiers to then exploit weaknesses that resulted in failed invasions with counter-invasions.

Massive numbers of soldiers in a small nation (say by comparing Japan to China or Russia or the USA today) in which ten percent of the entire Japanese people were samurai of one kind or another. They were not just soldiers but the aristocracy responsible for the culture and economy and for clan justice and governance. There is no Japan...yet, but warring clans.

So since 20 cities were built up with maxed infrastructure based upon the technology learned thus far, then that left 16 for the tribes to build and wisely so such that they could be the roads and improve the terrain of those remaining 16 (since 36 x 7 = 252 and since the maximum is 254 cities), and develop sufficent soldiers to defend those 36 cities.

That worked on a tribe setting of rational/military perfectionist. ..even though every tribe has been in a constat state where in nearly every season from March 1448 through March 1470, every tribe has been at war and often by up to four others based largely on that struggle to gain those remaining 16+ cities.

The AI managed to fully build the tech tree for Budo (the ways of war) and now is aggressively building Bushido (the ways of chivalry) despite the fact that there is no clear "payoff" in terms of better units or infrastructure. The AI looks at the values in the tech tree including the modifier and determines that there is a worth to keep on doing research.

This Bushido will have special spawned units and events like more money and aggression. This is tobe determined.

It already has the capability since about 1457 of building very strong defensive armies and units, then learned how to build offensive armies beginning in 1465, but thus far as the defensive armies are so powerful as single units, has switched from defense with naganata level 3 to defenders (which are the first armies).

Takeda elected to wait on defenders and stuck to making a significant Wako fleet and naganata units and advanced more on technology while all others (Mori, Oda, Imagawa, Hojo, and Uesugi) built up far more defenders. I'm playing the Shimazu and considering where the best fortresses should be prebuilt as well as roads and terrain. With the Uesugi having their flank protectedand naturally the Shimazuhaving theirs protected as well, either would have the advantage. In history, just the Shimazu maintained their autonomy during the Tokugawa Shogunate. (To some degree and in some ways they had less demands placed on them, and then consider theyhad not committed large troops to Sekigahara thus history might have turned out far differently).

The AI has amassed 100,000+ in their treasury despite all of this. I want that so the AI does not have a cash flow issue. It started at 20 cities and 30,000 each and that ebbed and flowed down to 9,000 as it had to cope early on with constant war. It started out with diverse available troops with half of the tech unlearned and so had sufficient offensive and defensive capabilities, then built better units to see what things it could try to defend and gain new territory.

None of the 20 citadels were lost but some new cities exchanged hands and multiple fortresses went back and forth. So far so good.

Now I'm waiting for the AI to stop making defensive armies and to switch to Attackers (which are the first offensive armies).

The next upgrade won't be until 1544-1545 with those attackers, then firearm units in the same period, then 1548 should see the first atakebune ships, then maybe true armies around 1560 or so. Then ultimately imperial armies post 1570.

I'm hoping to see mixed tactics and strategy to gain citadels from 1548 onward. It seems to be the tribe parameters of rational, military, perfectionist may have to be adjusted some. Originally it was aggressive, military, expansionist which caused insane levels of pointless causualties. Now it's reasonable but may be too reasonable on the highest level to deal with human players.

For play testing I have elected not to start new aggression but to defend as my position is nigh impregnable as all of Kyushu is under my dominion as the Shimazu. In history that was the Shimazu and the Otomo.

With a conversion to ToT and using the ToTTP software, the Uesugi far to the NE and with the rich, though hilly frontier terrain, might be able to keep on expanding and become the most powerful as the rest fight over existing terrain.

As the 254 city threshold was reached, that left a large border "no man zone" where battlefield conflicts are taking place. That seems historical versus the constant sieges in vanilla Civ 2.

Allegedly the defensive value of the city tile is not a factor in the battle calculations. I wonder as each largely began as extremely strong values with many being on a terrain defensive value of six. It seems to me that that is preventing cities (at the highest level citadel) from being taken without either extreme casualties from regular units or until the appearance of attackers ( offensive army 1). It may take until 1548 with offensive army 2 to appear to take those while right now in 1470, the AI may begin taking the middling 16 cities they each built.

 

[Cassius Critzer]

Since their discovery in the tech tree in 1457, Defenders were mustered in 1462. And since that time, by March of 1472, only an average of 2 per tribe have been lost as casualties. Most tribes have built at least 15 and often 20.

Defenders
Build cost 85 Movement 8
Attack 65 Defend 75 HP 8 Firepower 8
Can see two spaces, ignore zones of control (need to be surrounded), x defense versus horse (they possess lots of pikemen), and can see ninja.

No attackers built yet which is a mystery as human players would have begun building them.

Attackers
Build cost 90 Movement 9
Attack 85 Defend 65 HP 8 Firepower 8
Can see two spaces, ignore zone of control, x defense versus horse, can see ninja.

A typical tribe has in excess of 100,000. They have great cashflow.
36 cities with 20 maxed out on infrastructure.
Cost 1390
Income 7860
Science 1788
Discoveries every 13 turns so with seasons that's 4.2 years. Most have nearly maxed out the tech tree with two or three Bushido levels left. Then no tech is needed and General Research has a modifier to greatly discourage investment until 1542. By 1475 then, most need not do any research until 1542. A human player would have modulated this better.

Most have switched to Fundamentalism as the science penalty is very minor and so there is little rioting with some hunger in a few of the new villages. The AI has aggressively been building roads and irrigation and farms in the new 16 villages they are turtling so some hunger can be expected. The yari ashigaru have an engineer quality as in history they were farmers and so the AI is sending them out to do needed infrastructure while protecting them with Naganata and Defending armies.

Attacks are mostly:
Border clashes by probes
Denial of resource attacks by attacking infrastructure upgrades to terrain
Fortress acquisision
Contested territory issues where there are border disputes with overlapping tiles

All of that is excellent. Now it's waiting for the AI to elect to build the Attacking army based on threat and causalties.

The ninja units were tapered back as their firepower was absurd (equivalent to cannon attacks prior now halved). They are treated like missiles as they are killed on completion of their suicidalmission. They have very strong attacks still. They are wisely attacking garrisoned high defense units like the nagamaki.

I've been combing nagamaki who are excellent seasoned samurai defenders with nodachi which are great attackers, essentially impetuous and making risky attacks so they have less defense. Such garrison units coupled with heavy cavalry 3 or imperial guard (advanced cavalry archers) or elite cavalry (best in the game with high attack and defense), means an excellent ability to "sally out" with invasions to then weakened enemy infantry. This especially works well on highly advanced terrain infrastructure plus fortresses.

This sallying out replicates traditional cavalry usage to rout or mop up remnants of infantry.

Switched the tribes from:
Rational, Militaristic, Perfectionists

To:
Neutral, Militaristic, Perfectionists.

I think the border daimyo should be rational (Shimazu and Uesugi), but the Mori, Oda, Takeda , Hojo and Imagawa should be neutral. Hopefully this causes Attacker armies and causes a chain reaction of recruitment as Defender armies did.

 

[Cassius Critzer]

https://totalwar.fandom.com/wiki/Category:Fall_of_the_Samurai_ships

Because of the hard work of many modders depicting European navies, and the samurai postPortuguese influence began altering their stagnant naval technology, then with available slots of late recruitment and due to obscelescent largely semi-professional units, then later stronger naval units can appear post 1542. An artist who does pixel art for units could make small changes in existing European units to depict the ones in the link above. Regardless they are close enough to just use existing units. The problem is the movement rate is maxed out and there is little as attributes to make a nuanced naval battle possible except for attack and defense and hp and firepower. Things like morale and boarding and range and crew complement would make for thrilling naval combat.

 

[Cassius Critzer]

Advanced armies can act as marines and so they could attack a ship ie boarding but would die in the process so that is a waste as you can't seize a vessel. So all you have is ship to ship combat.

At advanced technology, your ports can have naval bombardment as in history there was the Otomo using Ozutsu shore batteries.
https://en.m.wikipedia.org/wiki/Artillery_of_Japan

Very high in the tech tree after extreme expense you have essentially safe naval transport from city to city having those transport facilities (like an airport). And so since ninja are treated both like submarines and air attacks, you could have ninja bases with garrisoned defenders, say on islands ie airbases.

When the Shimazu control all of Kysuhu, their landholdings are significant BUT given the location of the Uesugi, their lands may be double what the Shimazu own but the 254 city limit sadly makes them not as powerful as they could be. So with so much land under their domain, such fast and safe transport wouldbe essential to Uesugi success. Likewise, just to traverse that far if the Shimazu and control all cities, might take the Shimazu from 1448 to 1700+ if not for safe transport.

 

[Cassius Critzer]

I declared war against all but the Takeda as they already were at war with three clans. That caused chain reactions and broken alliances across all but the Takeda and Shimazu but still Attacker armies (the first level of offensive armies) were not being recruited. The clans were smug and now I see why. Remember this is July 1472 through landing July 1473.

I see why with testing. The Defensive army is so powerful as configured versus the offensive armies. In testing it took seven imperial armies (basically a level three offensive army) to take five defensive armies plus three naganata level 3. Then as multiple budo tech causes partisans, this spawns a true headache with up to 20 imperial guard (advanced cavalry archers). Their unit stats are:
Attack 5 Defend 6 HP 2 FP 2.

Imperial armies can do just about anything but cost 125 shields.
Attack 98 Defend 98 HP 12 FP 25 which I think are maximums.

To realistically travel and do a naval invasion from the closest point in Kyushu to Imagawa historical territory on their borders with the Oda, that took three turns ie a year. Any naval battle halts movement so you're trying to avoid engagement, deploy ground forces and maybe do a naval engagement (as taking the city may then erase those ships). And likely you can best occupy all armies at once by sending the entire ship in to "dock". Obviously you are looking for a vulnerable port city.

This means at a bare minimum you need three probably atakenbune class ships as wako class might end up taking out boarded armies in transport and or greatly delaying the naval invasion by a total of five turns (a year and 8 months).

At Totomi, the major Imagawa had 10+ defensive armies so likely you would need at least 14 armies to take it not counting dealing with partisans.

So attacking meant instead of seven imperial armies (and realistically far weaker Attacker units would be used so likely 14 of them) plus 2-3 more to deal with partisan imperial guard. These cities are strongholds.
That means many more realistic wako in a fleet like 8 or so to both transport and deal with enemy naval forces.

So realistically
Defenders
5 Defender class armies
3 Naginata
(Plus up to 20 spawned elite cavalry archers)
(Possibly each Defender is 670 some samurai plus naginata companies 3 x100 plus up to 20 x 100 partisans)
Subtotal Defenders 5650 samurai

Attackers
Seven imperial armies so 7 x 1000 with 2 more to deal with partisans.
Subtotal 9000 invading samurai...not including the navy.
[Because taking a city eliminates the redcruited units, if other units are not there from another supporting enemy city, there is a very high chance of 20 partisan elite cavalry archers.]

Realistic Attacker class armies have 640 and would have required double so 14x640.
Realistic subtotal at least 8964 samrai but might not succeed. I'd say a 30-60% chance of success with many things that could go wrong.

This seems very historical.

Those up to twenty partisans meant a a large urban city went into a negative with severe hunger meaning you just about must have extra armies to get them off the tiles so farmers can get back to work plus trade can be restored.

It seems very realistic to history as massive armies were needed to deal with major garrisoned troops. Then also in-game deal with issues with stack kills.

I don't want to weaken the Defender army, but might. A naval invasion commander might instead use attrition by occupying the upto 20 tiles around the city,thus causing severe hunger and forcing the defenders to redeploy their own military forces to break the siege. If they foolishly sally, the invaders likely win.

Conclusion: armies seem not to be overpowered but require realistic amounts to defend and attack similar units after many hours calculating the necessary unit attributes.

 

[Cassius Critzer]

Now combine the last two posts.

To have an army fielded close enough to the front to launch a fast invasion, post 1600 AD, one might have occupied enemy fortresses and accumulated them after many invasions and redeployed them over and over incrementally.

Or used the fast transport (airport in the vanilla game) but at huge infrastructure costs.

And or used these ninja strongholds ie airfields to act as temporary staging locations and most likely on islands. I may have to add some islands on the map.

 

[Cassius Critzer]

By July 1473, and considering that a Defender class army has 650 samurai in it of various kinds, then here is the census for the following and not including their standard units:
Hojo 200 or 130,000 samurai
Mori 95
Oda 115
Imagawa 178
Takeda 135
Uesugi 164

Each is playing very differently based upon their map location even though they are set to the same style now of neutral, militaristic, perfectionists.

These are staggering numbers. Each tribe except the Shimazu and Imagawa (both used extensively in testing) have 36 cities.

There probably are 900 other units besides these for individual units with those tallying 80,000 to 90,000 samurai plus these.

I'm curious when the units will slow down how fast the turn rates are as they deploy. Tribes could try to conquer then leave their own lands defenseless as each clan has tremendous power. It should be very fun for multiplayer.

If I did the math correctly as I've been working round the clock on little sleep, that's something like 754,000 samurai. I know it sounds absurd but imagine the fun.

The peak census is supposed to be 15 million and based on percentages but excluding wives and children, that does come out to 750,000.

Per one document, this excludes those under age 15 who cannot be used as runners and in similar duties, and not support crew, and excluding the infirmed due to advanced age as even old samurai were still used post age 60+.