[Vote] (5-19) Japan Adjustment Proposals

Approval Vote for Proposal #19


  • Total voters
    97
  • Poll closed .
Status
Not open for further replies.

Recursive

Already Looping
Moderator
Supporter
Joined
Dec 19, 2017
Messages
6,257
Location
Antarctica
Current bonuses on Japan:
Shogunate:
When a GGeneral or GAdmiral is born, gain 50% completion towards the next Great Artist/Writer/Musician in your capital
+1 :c5culture: and :c5faith: to all military training and defensive buildings

10 buildings are boosted by the UA:
  • Military training: Barracks, Armory, Military Academy
  • Mainline Defensive Buildings: Wall, Castle, Bastion, Arsenal, Military Base
  • Extra Defensive buildings: Minefield, SDF
Dojo
5:c5science:
2 flat supply in city
25XP to all units trained in city
All Melee/Gun and naval melee units gain Bushido promotion (10HP on kill, fight better damaged, 1 of 8 randomly-picked extra bonus)
Gain :c5culture::c5science: in this city when a unit built here gains a level (gains yields equal to 4 * (X-1)^2, where X is the new level gained)


VP Congress: Session 5, Proposal 19
Discussion Thread: (5-19) Japan Nerf
Proposer: @pineappledan
Sponsor(s): @pineappledan

Proposal Details
Spoiler :

1. Reduce the buildings boosted by the UA from 10 to only the 5 Mainline Defensive Buildings: Wall, Castle, Bastion, Arsenal, Military Base
2. Remove Bushido promotion from naval melee

Spoiler Reasoning :

  • According to both player and AI test feedback, Japan too strong right now.
    • They have a flexible and powerful kit that gives well-rounded yields
    • Their playstyle facilitates clean transition between domination, culture and science victory focuses late game, and they have a safe early game too.
  • They are too safe at founding.
    • This is because
  • The +1:c5culture::c5faith:on 10 buildings is too big
    • they get an early 2:c5faith::c5culture: in all cities. This is enough to give a major leg up on the competition.
    • This bonus continues to scale up to 10:c5faith::c5culture: in all cities, which is very substantial
    • This is a really, REALLY boring bonus. Even if it were a policy it would be lame, but it's a UA. It takes up too much oxygen for how lame it is.
  • Bushido shouldn't be on boats
    • Balance:
      • This bonus is contributing Japan's power at sea. The Bushido promotion bonuses are better on boats than they are on land units
      • Human players have reported making great use of it, especially in combination with Authority's 15 HP on kill. heal on kills is really good on boats, because they can't heal outside friendly lands normally.
      • Fight better damaged is another thing that works really well on ships that can't heal anyways. If you can't heal up, you may as well hit harder
    • Update history: This is the most recent buff given to Japan. Last hired, first fired. They were doing well before this, they certainly don't need it.
    • Design overlap:
      • This forces an overlap between Japan and Denmark. Denmark is the civ with bonuses to land/sea melee already.
      • Imperialism also gives melee boats fight better damaged, and the two bonuses don't overlap. Japan's bonus partially overwrites a policy bonus
    • Historicity: Giving Japan a naval melee bonus is bad history flavor.
      • This was added because the Japanese did boarding actions as their main naval tactic during the Mongol invasions. Japan's boarding actions were extremely minor in that war. Not noteworthy, much less decisive. The first Mongol invasion died in a storm with no engagement whatsoever, and the second was turned back in a land battle, and then they lost their invasion again due to weather. The Mongols chained their boats together to counter possible Japanese boarding actions, but that mainly made the storms worse. The Japanese can't take credit for that.
      • Also cited was the Imjin War, and this is where the idea gets a bit silly. The Japanese were SLAUGHTERED because of their dependence on boarding actions in that war. They lost almost every naval engagement in that war against better-made Korean ships that emphasized cannons. In 1 engagement, the Japanese lost more than 200 vessels against 16 Korean Panokseon Frigates. The Koreans didn't lose a single ship. The Koreans created their own anti-boarding ship, which is a UU in the game already: The Turtle ship. These ships completely countered the Japanese. The Koreans never lost a single battle where a Turtle Ship was engaged, and they didn't lose a single one in the entire war. The Imjin war can be summed up as follows: Japan invades and quickly wins the land war and conquers all of Korea, but then fails to protect their ocean supply lines, causing their occupation to collapse against partisans and guerilla remnants. And we want to highlight this atrocious failure as a UB ability? How is a historic naval humiliation supposed to be the basis for a bonus?
      • The Japanese naval presence was not normally composed of warrior-poet bushido-following Samurai boat boys. They were a rabble; thieves and peasants with rowboats and guns. Oda Nobunaga's "Admiral" was a pirate. Japan only professionalized its navy with the Meiji reforms in the 19th century, and they won their greatest victories with long distance shelling. Ranged Naval. And after that they were at the forefront of the Aircraft Carrier revolution. Their preceding time spent depending on naval melee engagements were marked by Japan's amateurism, defeat, disgrace, and isolation.
      • Japanese shipbuilding was awful. Half the reason the Japanese were so dependent on boarding actions was because their ships were so flimsy they couldn't handle the weight or recoil of cannons. They build massive floating fortresses called Atakebune, and they weren't seaworthy at all. Then for the next 300 years, building large combat-worthy ships was outright banned by the Bakufu.




VP Congress: Session 5, Proposal 19a
Discussion Thread: (5-19a) Japan Tweaks
Proposer: @Legen
Sponsor(s): @Legen

Proposal Details
Spoiler :

Spoiler Dojo stats (changes from the base Armory underlined) :
Cost: 300 :c5production:
Science: +5 :c5science: (from 2 :c5science:)
Culture: +3 :c5culture: (from 0 :c5culture:)
Maintenance: -2 :c5gold:
Experience: +25 xp (from +20 xp)

Doesn't require a Barracks in order to be built.
Supply: 2
:c5war: (from 1:c5war:)

All Melee, Mounted Melee, Naval Melee, Gunpowder and Armor Units trained in this City receive the Eight Virtues of Bushido Promotion.
When Units are promoted, gain :c5culture: Culture and :c5science: Science based on their current Level.


-1 unhappiness from :c5food::c5production: Distress.


Spoiler Formula for Dojo's "yields on leveling" mechanic :
Current formula: max{1, 4 * (NewLevel - 2)^2}

Examples:
Upon reaching level 2: 1 :c5culture::c5science:
Upon reaching level 3: 4 :c5culture::c5science:
Upon reaching level 4: 16 :c5culture::c5science:
Upon reaching level 5: 36 :c5culture::c5science:
Upon reaching level 6: 64 :c5culture::c5science:
Upon reaching level 7: 100 :c5culture::c5science:
Upon reaching level 8: 144 :c5culture::c5science:


Spoiler Japan's current UA (Shogunate) :
UA: Shogunate
+1 :c5culture: Culture and :c5faith: Faith from Defense and Military Training Buildings. When a :c5greatperson: Great Admiral or :c5greatperson: Great General is born, gain 50% progress toward a :greatwork: Great Artist, Writer, and Musician in your :c5capital: Capital.


Spoiler AI difficulty handicaps :
From Recursive's 1-05 proposal, relevant lines in bold:

Difficulty SettingWhat does it do?SettlerChieftainWarlordPrinceKingEmperorImmortalDeity
AIStartingDefenseUnitsExtra Warriors (or era equivalent unit) at game start00011111
StartingMinorDefenseUnitsExtra Warriors (or era equivalent unit) at game start for City-States00112233
AIWorkRateModifierWorkers complete improvements x% faster+0%+0%+0%+25%+35%+50%+60%+70%
AIUnhappinessPercent% modifier to Unhappiness from Needs+10%+5%-0%-0%-0%-0%-0%-0%
AIUnitCostPercentDiscount on unit maintenance100%100%100%90%85%80%75%70%
AIBuildingCostPercentDiscount on building maintenance100%100%100%90%85%80%75%70%
AIUnitUpgradePercentDiscount on unit upgrade cost100%100%100%90%80%70%60%50%
AITrainPercentModifier to combat unit production cost110%105%100%90%85%80%75%70%
AIConstructPercentModifier to building production cost110%105%100%90%85%80%75%70%
AIPerEraModifierAdditional reduction to building and combat unit production costs (multiplicative) for each game era that passes-0%-0%-0%-4%-6%-8%-9%-10%
AICivilianPercent (new)Modifier to production cost for non-combat units110%105%100%100%100%100%100%100%
AICreatePercent
AIWorldCreatePercent
AIWorldConstructPercent
Modifiers to production cost for Projects, World Wonders and World Congress Projects110%105%100%100%100%100%100%100%
AIFreeXPFree XP given to units (except starting pathfinder), scaling with game speed0001015202530
AIFreeXPPercent% increase to XP gain from combat+0%+0%+0%+20%+40%+60%+80%+100%
VisionBonusExtra sight for AI units (except scouting and trade units) in # of tiles00000012
AIResistanceCap (new)Maximum anti-warmonger fervor AGAINST the AI64%48%32%40%50%50%50%50%
AIBarbarianBonusAI Bonus VS Barbarians0%10%20%20%25%25%30%30%
DifficultyBonusBaseSee AI Periodic Yield Bonuses, below00047101214
DifficultyBonusASee below000320335350360375
DifficultyBonusBSee below000190210230240260
DifficultyBonusCSee below000100123145160180


Proposal
  • Remove the Dojo's "yields on leveling" mechanic.
  • Remove the yields on Defense and Military Training buildings from Japan's UA.
  • Japan's UA gain the following effects:
    • :c5greatperson: Great General and Admiral birth also grant 250 :c5culture: Culture and :c5science: Science, scaling with Era.
    • Free :c5greatperson: Great General at Iron Working.
Spoiler How the tweaked UA would look like :
UA: Shogunate
When a :c5greatperson: Great General or :c5greatperson: Great Admiral is born, gain 250 :c5culture: Culture and :c5science: Science, scaling with Era, and 50% :greatwork: Great Writer/Artist/Musician progress in the :c5capital: Capital. Free :c5greatperson: Great General at Iron Working.



Rationale

These changes aim to replace a component that has issues with AI difficulty scaling, and reinforce the UA's GG/GA birth aspect instead. It also tries to address an issue with the Defense/Military building mechanic, which departs from its original intention.

The core issue of the "yields on leveling" mechanic is its volatile performance with the AI handicaps. This mechanic has a history of easily getting out of hand in high AI difficulties from back then, when Gazebo experimented placing it in the UA; there were reports of Deity AI Japan about 10 techs ahead when everyone else was at Classical. This mechanic scales exponentially with unit level by design, and the AI handicaps has two parameters, AIFreeXP and AIFreeXPPercent, that can push this mechanic to extremes. I suspect this mechanic is affecting AI Japan's winrate at high difficulties again by the results of the 3.6 AI Emperor/Warlord tests, in which AI Japan ranks 3rd place in Emperor and 36th in Warlord.

Notably, Japan's wins at Emperor has a significant portion of scientific victories, despite Japan's kit being mainly tailored abound cultural/domination victories instead, and AI Oda having an overall low scientific flavor (conqueror personality, the lowest science/spaceship flavor of the four types, and no personalized adjustment towards either). With the only extra science yields in Japan's kit being in the Dojo, and the leveling mechanic having a prior case of getting out of control to extreme results, there's reason to believe the same happened in those tests. Meanwhile, these yields do not perform to such degree in human hands. It is generally treated as a fun extra that quickly falls off. Human players don't generally aim for a scientific victory with this mechanic in mind, as its values are not currently tuned to carry a victory type.

So, there's a strong mismatch between how this mechanic has to be balanced between human and AI, as well as between low and high AI difficulties. This makes it a problematic mechanic on this civ's design, as what would be balanced on one end is likely to be too powerful or underwhelming on the other. Until the AI combat handicaps are reworked, it is probably healthier to refrain using leveling as part of Japan's design and focus on their other core mechanic, a.k.a. the GG/GA birth rewards.

The "yields on Defense/Military buildings" has a different issue. Originally, it was intended to cover Japan's early game, as much of Japan's kit was being moved to Medieval (the reason why the Dojo has so many effects stacked on it) and AI Japan was struggling to keep a decent pace with other civs in the first two eras. As implemented, though, this mechanic extends its usefulness to much later eras. There's also complaint about it not being particularly fun or interesting, as it just reuses a very common effect in policies and wonders with no particular link to the UA's theme. It is also tied to costly Ancient buildings; players would sometimes report struggles to find a decent build order for Japan due to the unusually high maintenance costs and unorthodox tech priority that comes from prioritizing both Barracks and Walls.

The tweaked UA aims to cover for the yields from those two mechanics, tying those to the :c5greatperson: Great General and Admiral birth instead, and reinforcing this core mechanic for Japan. Moreover, the addition of yields on :c5greatperson: GG/GA birth opens the possibility of using the UA as an early game boost through a free Great General; here, the proposed UA gives one at Iron Working, a tech that used to grant one though the Heroic Epic (instead of a Great Writer) in earlier versions of VP. This should cover the original intention behind the "yields on defense/military buildings", and expands Japan's early game with Tradition (who can take advantage of the early 50% :greatwork:GWAM progress and normally struggles with setting defense/military buildings) and Progress (who benefits greatly from early :c5science: science), giving some extra variability over the straightforward Authority start.

The free :c5greatperson: Great General and the yields on :c5greatperson: GG/GA birth should also be insensitive to AI difficulty, neither being particularly affected by the AI parameters and only giving one extra injection of handicap yields (gold and GAP) early on. Overall, the tweaked UA should place a much closer balance parity between human and AI performance for Japan than what we currently have with the "yields on leveling" mechanic, and hopefully address the complaints regarding the yields on defense/military buildings.



VP Congress: Session 5, Proposal 19b
Discussion Thread: (5-19b) Japan Nerf
Proposer: @azum4roll
Sponsor(s): @azum4roll

Proposal Details
Spoiler :

Proposed Japan UA:
+1 :c5culture: Culture and :c5faith: Faith from Defense Buildings. When a Great Admiral or Great General is born, gain 50% progress toward a :greatwork: Great Artist, Writer, and Musician in your :c5capital: Capital.

Defense buildings include Walls, Castle, Bastion Fort, Arsenal and Military Base only. Current boosts to Military buildings provide 2 :c5culture: :c5faith: in Ancient and make early culture/religion game too easy.

Proposed Dojo:
Remove Bushido on Naval Melee
Remove yields on level up
Gain 1 :c5science: Science and :c5culture: Culture per XP in this City when Units trained here gain XP (note: does NOT include training XP)

Comparison (assuming only Barracks and Dojo are built, and no handicap XP):
Level/XPCurrent Dojo Total YieldsProposed Dojo Total Yields = XP Gained
3/305N/A
4/602120
5/1005760
6/150121110
7/210221170

Comparison (assuming only Barracks and Dojo are built, and Emperor AI handicap XP):
Level/XPCurrent Dojo Total YieldsProposed Dojo Total Yields = XP Gained
3/30N/AN/A
4/60210
5/1005740
6/15012190
7/210221150

As you can see, the proposed version is difficulty neutral, and is actually a slight buff for low difficulty AI and humans at low unit level assuming no Teocalli/Brandenburg Gate/Military Academy are built.
It also no longer has yield jumps in specific XP thresholds and gives a much more consistent stream of yields.

As for the removal of Bushido on Naval Melee, see pdan's thread.
 
Last edited:
Please vote no to all of these. There are other ways to de-tune without fundamentally changing what is an interesting and fun civ to play. How the hell did we get here??
 
???

The only one of these that changes any mechanics is 24a. You described your own changes to Japan in the discussion thread, and they are far more ‘fundamental’ changes than what is proposed in 24 or 24b.
 
And I would answer the same thing to anyone who ask others to vote no : it is not yours to decide alone, don't try to impose your vision of things to others
 
Please vote no to all of these. There are other ways to de-tune without fundamentally changing what is an interesting and fun civ to play. How the hell did we get here??
Please don't. I don't want any of these passing either, changes to the defensive/military building bonus and bushido boat removal are both unnecessary and ineffective for what this proposal wants to accomplish, but just directly asking people to vote no is a terrible idea and more likely to get them to vote yes than anything else.
 
???

The only one of these that changes any mechanics is 24a. You described your own changes to Japan in the discussion thread, and they are far more ‘fundamental’ changes than what is proposed in 24 or 24b.
On reflection, it was a lazy comment on my part as I was in a rush. I used the word fundamental, because I assumed changing the yields from leveling require changes to the DLL for 24a and 24b. I tried changing these years ago and wasn't able to, perhaps that has changed?

The second reason that I'm against these proposals is that I haven't reached a conclusion on bushido on naval units.
24 and 24b both remove these changes that are fairly recent. I'm not familiar with exactly when that change was made or the discussion that took place that brought that change into the game. However, I'm trying to give it a fair shake and rebalance with the assumption these changes were discussed and voted on by "the process".

In terms of my own balance changes, I realized after my last comment on the Japan thread that I don't have any way of turning off the yields for scouts, since I don't make changes to the DLL. So, I'm back to the drawing board. We at least seem to agree that defensive building bonuses are boring and overpowered. I believe that removing science from Dojo will normalize the higher level AI results, and having it available earlier is fun, which is an element we should be considering more often.
And I would answer the same thing to anyone who ask others to vote no : it is not yours to decide alone, don't try to impose your vision of things to others
Who sounds like the autocrat here...? Silencing dissent.

Please don't. I don't want any of these passing either, changes to the defensive/military building bonus and bushido boat removal are both unnecessary and ineffective for what this proposal wants to accomplish, but just directly asking people to vote no is a terrible idea and more likely to get them to vote yes than anything else.
I've mentioned above that it was a lazy effort on my part, but I still don't understand the spirit of this comment. I did provide some rationale for why I thought it was a bad idea.

As far as I can tell this all spawned from Japan winning a few too many science victories?
 
As far as I can tell this all spawned from Japan winning a few too many science victories?
It's more that Japan just wins a lot:
Spoiler AI test game results :

1690240121750.png


Japan closed out 350 AI test matches on the standard balancing settings with a 26% win rate. Games are played with 8 civs, so win rates should ideally be around 12.5% (green line).

There are 6 civs in S tier right now, with a big 5% gap in win rates separating them from everyone else. The other 6 civs in that grouping have all come under varying amounts of scrutiny. Austria has 5 different proposals for how to nerf her this congress as well. There has been at least some discussion on the forums and on discord about what to do with the other 4. It's not that Japan is being singled out, it's just that people really like Japan's kit, and are really worried about him being made less fun in the name of balance.

The science wins aren't a big deal. It's his 3rd-most common win type, and he has 2 different :c5science: bonuses in his kit. Seems right to me. Brazil won almost as many games with SV as Japan did with even fewer SV bonuses, after all.

I assumed changing the yields from leveling require changes to the DLL for 24a and 24b. I tried changing these years ago and wasn't able to, perhaps that has changed?
It does. The proposed change in 24b involves changing how the instant yields on levelling is calculated. Azum posted a comparison of the yields, and they are a modest downgrade compared to what they are now. It is a complex change that requires going 'under the hood', but the end result is just a number tweak.

The second reason that I'm against these proposals is that I haven't reached a conclusion on bushido on naval units.
You can see it here. It didn't generate much discussion.
The Bushido promotions are much stronger on boats than they are on land, because boats don't have many ways to heal themselves outside friendly lands. So any ability that gives healing on boats has a large impact.
 
Last edited:
It's not that Japan is being singled out, it's just that people really like Japan's kit, and are really worried about him being made less fun in the name of balance.
Yeah, fun matters!

The science wins aren't a big deal. It's his 3rd-most common win type, and he has 2 different :c5science: bonuses in his kit. Seems right to me. Brazil won almost as many games with SV as Japan did with even fewer SV bonuses, after all.
1690243118058.png

If you remove the 4 SVs then the win rate might only be a little above the average. A nerf to science should also affect the other victory types as well.

Also, do you know if the proposed changes require changes to the DLL? A big part of my reaction is that I don't want to have to go through the trouble of setting up the environment for changing the DLL (which was hellish the last time I did it.. multiple versions of old MS .NET installers that required mysterious incantations). If I can make SQL changes to continue playing Japan the way it's been for years, then I'm less inclined to make a fuss.
 
Last edited:
Maybe it's more clear when you are looking at the raw numbers, but from the graph you've provided : if you remove the SVs then the win rate might only be a little above the average. A nerf to science should also affect the other victory types.
Japan won 4 of its 58 games via SV (7% out of 26% win rate). If you deleted all their SVs they would be at 11 wins out of 54 games, or 20%, which would still put them in 7th place, ahead of Poland. That is a helpful demonstration of how big the gap is between this S tier and the rest.

But of course deleting all of Japan’s science bonuses wouldn’t just remove his SVs. It’s not that clean. Even Polynesia won an SV with no science bonuses. Also the yields on levelling is the thing I like the most about the civ, so I would much rather look for other things to remove. That’s why I’m mainly opposed to 24a, because it removes a fun and rewarding mechanic and replaces it by doubling down on the GG/GA birth ability. It deletes an entire mechanic and just piles more yields onto existing ones.
Also, do you know if the proposed changes require changes to the DLL? A big part of my reaction is that I don't want to have to go through the trouble of setting up the environment for changing the DLL (which was hellish the last time I did it.. multiple versions of old MS .NET installers that required mysterious incantations). If I can make SQL changes to continue playing Japan the way it's been for years, then I'm less inclined to make a fuss
24 involves no dll at all.
24a is a big change that has entire new abilities that will require DLL coding
24b is basically the same as 24, but also includes a change to how Dojo’s levelling yields are calculated. It will work the exact same, and scale similarly, but the yield amounts will be lower overall.
 
Last edited:
I thing to point out on proposal B....the chart there suggests that there is roughly equal parity to a nerf between the curent model adn the "XP = science and culture" model. And in terms of absolute yields that is true.

It is important to remember the speed at which you gain those benefits though. For example, it takes 40 XP to go from level 4 to 5. So the new proposal could have gotten 35ish science and 35 culture while the leveling mechanic hasn't given you a single new bonus yet until you actually get to level 5. So over time the new proposals would be generating yields earlier than the current proposal, and yields now > yields later. Is that enough to make the new proposal yield neutral or even actually a buff....its hard to say at first glance, but I wanted to note that as the chart can be a little misleading.
 
Would be nice if someone can plot an actual graph for me.
 
Would be nice if someone can plot an actual graph for me.
You summoned me?

screenshot.png


This is starting with barracks + dojo (aka 40 xp).

Effectively what this shows is that at most XPs the new proposal will generate more yields than the old. That changes mostly around ~210 XP when the majority of the level Proposal B is trying to catch up.
 
You summoned me?

View attachment 667978

This is starting with barracks + dojo (aka 40 xp).

Effectively what this shows is that at most XPs the new proposal will generate more yields than the old. That changes mostly around ~210 XP when the majority of the level Proposal B is trying to catch up.
Just what I wanted! The red line will also shift right with more starting XP while the blue "curve" stays in place. It's definitely a nerf for the Emperor AI.
 
Last edited:
Who sounds like the autocrat here...? Silencing dissent.
There is a difference between arguing about why you think those are good proposals or not, and directly asking to vote no. Starting your post as you did, is -in a way- considering that everyone should have the same way of seeing things that you have, disregarding their own capacity of taking decision.
Explaining why you personally dislike changes, instead of bruteforcefuly asking others to vote no, would have been better.

I'm convinced this is the wrong way to change Japan. There are other ways to de-tune without fundamentally changing what is an interesting and fun civ to play.
This is a rephrasing which would have been entirely fine in my book. You don't try to force your vision on others and clearly state that this is your opinion.

Edit : typo
 
Last edited:
This is a tough one...
There are no other proposals to nerf Japan, are there?
There was a big japan/france discussion thread in the main forum if you want to check it out.

There were lots of ideas. A few total overhauls were proposed, like bringing back the sakoku stuff, removing all the faith, removing all the science, etc. The ones that Azum and I proposed are the ‘safer’ suggestions to come out of it, because the community members that gave their input generally like Japan’s current schtick, and don’t want bigger changes.
Edit: This one might fix the xp problems https://forums.civfanatics.com/thre...aps-for-one-that-works-only-vs-humans.684630/
Which makes me like the initial proposal most.
However, that vote is going toward No at the moment quite heavily.
That proposal basically removes all XP bonuses to unit training and vs other AI and then gives double XP vs humans. I think if the changes were more modest, like adding a little extra XP vs humans and just reducing the XP bonuses on training and on AI vs AI combat people would have been more receptive. As formulated it is a radical change.
 
As far as I can tell this all spawned from Japan winning a few too many science victories?
The thing going on about Japan is that the AI is performing really well with the civ at high difficulties, but really poorly at lower ones. The scientific victories at Emperor are an indication of what is going on, rather than the problem itself.

Japan's leveling mechanic has an exponential growth; the yields grow exponentially based on the new level, while the experience needed for each new level grows linearly. This design was intended to push Japan to focus on having a small number of elite units, instead of just throw a large army at the enemy. However, it also means that AI Japan becomes exponentially more powerful with difficulty, as each extra difficulty level increases both AIFreeXP (extra starting experience) and AIFreeXPPercent (extra experience in combat). At high levels, AI Japan can field a large number of elite units, instead of the intended small core of elite units a human can, and reach those higher levels with ease long before a human player can.

We already saw this happen before when Gazebo experimented placing the leveling mechanic on the UA, hoping to address AI Japan's poor performance before Medieval. And we had reports of Deity players stating how AI Japan was 10 techs ahead when everyone else was at Classical, making it the most broken AI by far at high difficulties. At lower difficulties, AI Japan was not performing anywhere close to that. And, in human hands, it was underwhelming. Eventually, the mechanic moved back to the Dojo, at a lower potency, and the current defense/military building mechanic took its place to cover Japan's early game.

The science victories are an indication that the leveling mechanic is acting again, just like it did before, since the Dojo is Japan's only source of extra science. And AI Japan performing poorly at Warlord (ranking 36th out of 43 civs, from 3rd at Emperor) is another indication in this direction. Moreover, AI Nobunaga has the lowest science/spaceship flavors in the game (meaning this AI isn't even trying to get a science victory, unlike AI Sejong or AI Nebuchadnezzar), yet wins SV at a comparable rate to scientific civs (visible in the first published 3.6 test), indicating that this leveling mechanic alone is performing better than the entire kit of dedicated scientific civs in AI hands at higher difficulties. And, of course, that also means AI Japan is getting way too much culture than intended as well, not just science.

24a is a big change that has entire new abilities that will require DLL coding
It is very easy to implement, purely with XML/SQL.
 
Neither 24a nor 24b reward the "elite units" direction that Japan currently has.
 
So here's how I'm leaning right now.

I think Proposal A might be a buff in nerf's clothing, hard to say, but the idea itself doesn't do anything for me.

The OG proposal is the most straightforward one. A couple of key nerfs, nothing fancy, no possibility of a stealth buff. It just does the job.

Proposal B is also a question of.... is this actually more a buff than an nerf in some ways? As my graph above shows, there are several points where you would make more in the system than the old. However, it would address the AI bonus XP scenario.


For me it comes down to this.... if you think Japan is legitimately too strong, I think the OG proposal is the right approach. If you think Japan is fine in human hands, and is only OP in AI hands with their extra bonuses, proposal B seems a way to fix.
 
Status
Not open for further replies.
Back
Top Bottom