Japan

[SuperNoobCamper]

I think we need to separate AI performance with particular civ from human performance if we are to rate/judge how powerful a civ is on paper.
Some aspects of the game are just much more exploitable in human hands and unless the AI is strictly coded to abuse it like what a human would do it would never net the same or even close results.
besides the two exceptions of Songhai and Mongolia occasionally steamrolling a continent, warmonger civs in general seem to be performing poorly in AI hands *Looking at you, Shaka* .... Neither Japan nor France seem to be performing well in AI hands no matter what the settings are *at least in my experience*.
On paper the gap between the floor and the ceiling for UA's potential seems to be smaller in Japan's case, you will ALWAYS spawn GGs if you go to war so you will always get GA, GW and GM points, however the this is not the case for France as the UA's potential floor is literally zero ... sometimes France does not capture a single city and it would be like playing a civ without a UA but the potential ceiling is not as limited as Japan's and there is no diminishing returns.

 

[Recursive]

I'm not sure what the anti-warmonger bonuses are like between the AI

Up to 50% CS.

 

[Legen]

Japan's UA and Dojo are strongly built around the Samurai, whose promotions result in fast leveling and GG generation. It's easier for a human to plan around this spike than for an AI, who may stay passive during the Samurai window and waste the uniques. It doesn't help the AI that defensive pacts start being available at late medieval. The yields on military and defensive buildings are also harder to use at the beginning, as they are tied to buildings that have a maintenance cost. I don't know if the AI knows how to deal with that without going bankrupt.

As I understand, Japan's uniques don't reward the AI for what it tends to do naturally, nor push the AI towards actions that trigger the handicap yields.

 

[Histomen]

I am Deeply not having a good time with japan on emperor and up.
What are you alls build orders and policy choices because i cant do anything but be 8 techs behind the AI and just get bullied to oblivion before i get the chance to get samurais and dojos effectively out. Its always america, atila, aztecs etc that just ruin my games and if i last they cripple me enough to make me irrelevant to push past other culture civs and to even catch up in general.

 

[stii]

Japan is slightly below average with their abilites being ok but a bit late game. I like opening tradition to get a few early wonders and tech while being defensive. Then just taking authority as the 2nd policy tree and attacking someone and making them my vassal.

 

[andersw]

I am Deeply not having a good time with japan on emperor and up.
What are you alls build orders and policy choices because i cant do anything but be 8 techs behind the AI and just get bullied to oblivion before i get the chance to get samurais and dojos effectively out. Its always america, atila, aztecs etc that just ruin my games and if i last they cripple me enough to make me irrelevant to push past other culture civs and to even catch up in general.

On emperor, I play epic speed which is a bit easier when warring.
... I settle defensive, prio Terracotta if possible, beeline Steel and then Physics (skip sailing in classic) into machinery.
If Im behind here its mainly on the upper side and it is unlikely AI has gunpowder.
This is basically my authority goto strat.

If not possible early then go into grind wars with fortified samurai and work on treb exp (you want range eventually) and the very important great general points.
God of protection on top of this (and if you are really lucky fountain of youth).

Atilla is a pain to fight early against, his horse archers can really make pre steel wars miserable, aztecs and america arent that bad in my experience.
Do you settle hills?, have you roaded so you can move defenders? what about walls?
Ranged units and 2-3 swordsmen go a very long way in early wars, skirmishers are very nice.

 

[Legen]

I am Deeply not having a good time with japan on emperor and up.
What are you alls build orders and policy choices because i cant do anything but be 8 techs behind the AI and just get bullied to oblivion before i get the chance to get samurais and dojos effectively out. Its always america, atila, aztecs etc that just ruin my games and if i last they cripple me enough to make me irrelevant to push past other culture civs and to even catch up in general.

What has worked for me is opening with Barracks -> Market on one or two cities, and Barracks -> Walls on later cities. The rationale is that a Barracks with Japan's UA and Goddess of Protection gives 1 2 3 yields for 110 production, while a Monument + Shrine + Council gives 1 2 2 yields for 195 production, with the issue that the former has a maintenance cost. Going for a Market covers the maintenance of going for Barracks and Walls in up to 3.5 cities with the merchant slot, which lets later secondary cities to skip the Market for a while.

It has worked well on any of the three social trees and is noticeably production efficient on Tradition's secondary cities. It also allows Japan to set up secondary cities with military infrastructure earlier to defend against pre-Medieval aggression, on top of Goddess of Protection's extra heal.

Neither Japan nor France seem to be performing well in AI hands no matter what the settings are *at least in my experience*.
On paper the gap between the floor and the ceiling for UA's potential seems to be smaller in Japan's case, you will ALWAYS spawn GGs if you go to war so you will always get GA, GW and GM points

Japan is slightly below average with their abilites being ok but a bit late game.

I think the Samurai should have the iron requirement removed. Japan's kit is heavily designed around the unit's promotions and suffers if you don't have a large stock of iron. I think this is especially problematic for AI Japan, as the iron requirement prevents the AI from taking advantage of its ability to produce lots of units towards the Samurai and get their faster Dojo yields and UA procs.

 

[stii]

Yeah that seems fine. Samurai are very much a long term unit anyway. I don't think letting the AI spam them will hurt much. They are weaker than a knight already

 

[andersw]

I think its fine as is since samurai mostly are defensive melee and keeps its abilities on upgrade, compare with denmark who I feel are far worse with low/no iron.

 

[Legen]

How strong does everyone rate Japan's combat capabilities? I've been reevaluating it with how the AI seems to love spamming ranged units at me. Here is the detailed description of how they work:

" Combat Strength increases with damage taken"

Spoiler :

- Currently, wounded units without a Bushido promotion gets a negative combat modifier in melee combat that scales linearly with health lost, instead of dealing less damage as in vanilla.
- Bushido virtues flip the signal of the modifier, turning it positive.
- The penalty/bonus seems to be 1/3 of the health lost (truncated, sometimes subtracting 1%); -33% at 99 health lost without a Bushido virtue, +33% with.
- The penalty/bonus reaches higher values with promotions that increase maximum health found later on in gunpowder units (Fusilier/Rifleman/etc), such as 39% for a 1/120 unit.
- The formula only considers health lost. Current health or percental health are not factored. An unit at 110/120 health will have the same 3% modifier in melee that a 90/100 unit would.
- Neither penalty nor bonus are applied when receiving a ranged attack.

Virtues

Spoiler :

- Honor: +15% Combat Strength outside of friendly territory.
- Loyalty: +15% Combat Strength inside friendly territory.
- Courage: +15% Combat Strength when attacking.
- Self Control: +15% Combat Strength when defending.
- Righteousness: +20% Combat Strength VS Wounded units.
- Respect: +15% Combat Strength VS Melee and Gunpowder units.
- Sincerity: +50% Experience gained from combat.
- Benevolence: Units in adjacent tiles receive +5 HP when healing.
* All virtues also have a "Heal 10 HP after defeating an enemy unit".

Samurai

Spoiler :

Requires 1 Iron
23 (+1 over Longswordman)
Cover I (same as Longswordman)
Quick Study (+50% experience)
Great Generals II (generates 100% more Great General Points from combat)

The bonus CS from being wounded doesn't act against ranged attacks, so an enemy being heavy on ranged units can mostly bypass it. Defensively, it only helps when taking a melee attack. If the opponent has enough ranged units nearby, which the AI has learned to over time, you end keeping your frontline units healthier than you'd like with this promotion. The bonus from being wounded is good even at relatively healthy values, though; at 30 damage (how much you expect to deal and take vs another unit of equal CS), you turn a 10% penalty into a bonus, being effectively a 20% increase over a unit without it. At extreme values, it can be worth a 66% increase, before promotions that increase maximum health. If you can keep the unit fighting against melee units without retreating from light wounds, the promotion can pay off harder than many bonuses from other civs.

On a note, I think the "heal 10 on kill" would work better as a max HP increase for Japan. The Dojo promotion needs the unit to be able to survive multiple hits to shine, and maximum health works better with it than a heal on kill.

The samurai has been nerfed noticeably from what it used to be. I remember that the longswordman was a 20 CS unit with no promotions, while the samurai was a 24 CS with free Shock I, alongside Quick Study and Great Generals. The 20% higher base CS meant that your opponents had to use knights to stand in melee against a dedicated samurai push, while free Shock I allowed those samurai to learn March very quickly, especially with Orders (start with 55xp). At the time, March didn't have the -15% CS defense malus, which allowed for very aggressive plays. Neither advantage is applicable anymore, and there isn't a comparable alternative to them. An old alternative was to pick Authority and build Alhambra, as it gave free Drill I for a fast Stalwart or Blitz. Now that Alhambra grants Jinete instead, this is no longer an option for an early tier IV promotion.

I think the samurai needs a revision. It used to be that Japan had a power spike at Steel due to the combined military boost of the dojo's promotion and the samurai's own combat advantages (20% higher base CS, almost immediate access to a tier IV promotion). This power spike isn't happening as it used to, as the samurai is now just a long term utility UU, which is not in line with what a militaristic civ needs.

 

[stii]

I think Japan is a slightly below average civ. Its abilities are fine they work together to a degree but not that well. Not sure if that rises to the level of needing changes though.

 

[Legen]

When it comes to the samurai, I'm not asking for major changes. I'm just asking to reverse the effects of the changes that happened over time that kept nerfing it. We already know how Japan performed before and we didn't find that the UU or the civ needed nerfs, so it makes sense to revert them back to what the UU was capable before.

 

[pineappledan]

At the very least, I think boosting the CS back up to 24, or even 25 wouldn't be controversial

 

[Rekk]

@Gazebo AI game derived balance changes:
2019-08-21
samurai CS 23 (from 24)
longswordsman CS 21 (20)
ie a deliberate nerf to samurai
Knight CS 24 (25)
2019-08-31
Longswordsman & Samurai gain Cover instead of Shock
2020-02-09
Longswordsman CS 22
Samurai stays the same (swordsman UUs increase)

To keep the +9% advantage Samurai has over Longswordsman, it would need 25 CS. It depends on whether we want Samurai to start winning even more against Knight than it is now (given Bushido and Virtue).

 

[azum4roll]

I think Samurai can be buffed to 24. 25 would make them the same CS as Tercio (only adding Formation I when upgraded).

 

[Legen]

(given Bushido and Virtue)

I don't think the Dojo promotion should be factored here as a reason to hold back buffing the samurai. A militaristic civ with combat bonuses having an edge over a generic civ's knight is expected; a Swedish longswordsman (+20% CS when attacking, +30% CS from great generals instead of +15%) would also win easily against a generic knight. You'd expect a longswordsman UU from a civ that also has combat modifiers to do even better than that.

Also, remember that the "stronger when wounded" part was strenghtened at the price of no longer working in Ancient and Classical eras, nor applying to all units (which included naval and air). This happened when it was moved from the UA to the Dojo's promotion:

Because of space limitations in the UA, and out of a desire to organize Japan's UA, the UA and UB will be as follows:

UA:

Gain [ICON_GREAT_WORK] Great Writer/Artist/Musician Points in [ICON_CAPITAL] Capital when [ICON_GREAT_GENERAL] Great Generals or [ICON_GREAT_ADMIRAL] Admirals are born. Recieve -50% [ICON_TOURISM] Tourism from Civs that lack a [ICON_INTERNATIONAL_TRADE] Trade Route from you, and they cannot send you [ICON_INTERNATIONAL_TRADE] Trade Routes.

Dojo:

Will affect Melee/Gun/Mounted/Armor units. Eight Virtues of Bushido promotion will include (in addition to special effect) the 'fights well damaged' element, but at a slightly stronger rate (up to 34% of CS if fully damaged, v. 20% for prior UA). So fewer units will have it, but they'll be stronger for it.

G

Holding back the samurai due to how strong it would be with Bushido against melee units goes against the intent of that change. At the price of losing that combat boost for two eras and for many unit types, Japanese land melee units are then supposed to be particularly strong against other melee units. If we are in turn going to keep the samurai's combat advantages at a minimum to counteract the Dojo's promotion effectiveness, then we are making sure that move from the UA to the Dojo was a straight nerf to Japan, rather than a refocus.

2019-08-31
Longswordsman & Samurai gain Cover

- Swordsman/Longswordsman lose Shock I, gain Cover I (Legion gets Cover II)

I remember that there was an old patch in which the swordsman line gained free Shock I, which the samurai already had. The samurai had free Shock I over the longswordsman since vanilla and was for a long time in CPP/VP what allowed it to quickly learn a tier IV promotion, far earlier than its competition.

I think it makes sense to restore that free Shock I, or alternatively, a free Drill I. Both of them are small promotions by themselves, but enable a lot on a unit that has Quick Study and the extra +5 XP on the Dojo over the Armory. In fact, this was a big part of the fun with the unit.

As a suggestion, what would everyone think of the samurai gaining both free Shock I and free Drill I, but stay at either 22 or 23 CS, a.k.a. same or only slightly higher base CS than the longswordsman? The UU would still be overall buffed, but be very flexible in how you promote it due to its quick access to the high tiers on both branches of the promotion tree. This should work really well with Japan's theme on leveling units, with the samurai's Quick Study and with the variability added by the Eight Virtues. It is also a concept not explored on other civs' UUs, who tend to at most help you with only one promotion branch.

 

[Rekk]

Interesting. Swordsmen that upgrade into Samurai will generally lose out on one of them, but would still get the other.

 

[pineappledan]

I don’t like giving so many free promos. +50% XP is already a major perk. I would just stick to higher CS. 24 or 25

 

[Rekk]

Plus, free stem promotions don't make upgraded swordsmen reach tier IV earlier, only newly trained units.

 

[Legen]

I don’t like giving so many free promos. +50% XP is already a major perk. I would just stick to higher CS. 24 or 25

Maybe, but I don't think restoring the free Shock I it had over the longswordsman is too much to ask. Buffing it back to 24 CS and free Shock I seems reasonable to me.