So difficulty now increases your own unit upkeep as well. What are peoples opinions of it?

[Warpfire]

So I started a coop campaign yesterday and noticed that the upkeep of units did not match at all with what was displayed in the building browser.

After some experimenting it seems that ridicilously high upkeep increase on higher difficulties. (For reference in Queeks army (Who has a 50% upkeep reduction but in another lord's army they also seemed to have about doubled upkeep on very hard) a unit of stormvermin costs 100 upkeep on normal, 150 on hard and 200 on very hard)

At the same time like in the previous game the AI has a upkeep reduction on higher difficulty levels (Probably around 50% on very hard given that a skink priest costs 90 for them while an agent for the player on very hard costs 400)

In the previous game Clan Angrund had the unique modifier that all units required 50% more upkeep untill they captured karak eight peaks and this was a major influence on their playstyle yet it seems that now very hard has a larger upkeep penalty then that by default with no way to get rid of it.

In effect it means that the AI is able to field roughly 4 times as many units as a player could (or potentially even more since they probably also get an income modifier),

I suppose it is still somewhat manageable but it seems like too much of a fake difficulty when any random Lizardmen AI could field a unit of Saurus Warriors for a unit of clan rats and still have some income left over.

What are other peoples opinion of it?

Edit:
I did some checking and it seems less bad then I thought its just that the logistics system seems quite a bit more punishing though I can't remember it listed as such yesterday though it is listed correctly today.
Numbers seem to be:
Normal: 10% upkeep reduction
Hard: 10% upkeep increase
Very Hard: 30% upkeep increase
What caught me is that the first additional army already has a 15% increase in logistics costs (though I could swear it was listed at 2% increase similar to Total War Warhammer 1)
The stormvermin example is still accurate with 0 addtional lords (60% upkeep reduction of 250 is 100, 40% upkeep reduction is 150 and 20% reduction is 200).
The 400 upkeep for agents was with 2 addtional armies so presumably 60% addtional upkeep costs over 250 upkeep for 400 upkeep for the agent.

[burningmime]

引用自 Sai Kyouji：

引用自 burningmime：

The AI plays the same on every difficulty. They just get bonuses (and you get nerfs).

Really? I heard they are a lot less aggressive below Hard.

Oh right there's the "avoid player" and "target player" things. I meant that they don't play any *smarter*.

[Raviollius]

I have bad news for you: the battle bonuses for the AI are still there.

[Knight-du-Lys]

The only battle bonus I saw that the AI gets is actually a debuff on player army of -4 leadership during battle on VH difficulty.

[StarofTanuki]

引用自 YxTHE_BUTCHERxY：

The only battle bonus I saw that the AI gets is actually a debuff on player army of -4 leadership during battle on VH difficulty.

no dude on very hard AI unit get +12% most stat but -12% most stat for player . CA already confirm that. so you can test in custom battle

[Red Bat]

I don't understand why people always complain about the AI cheating in strategy games. Making AI that doesn't cheat and is still challenging is so time consuming to do that when Galatic Civilizations did it, it was considered the game's main selling point. It's not like you just spend a few hours coding "Do this if the player does that" for a few dozen lines and you'll get a smarter AI. It doesn't work like that. The more complex the game gets, the harder it is to make the AI good. Then once you get an AI that does play well, you run the risk of making an AI that's actually borderline unbeatable because it plays like a player but without human error. For a game like Total War, progressive AI boost and player penalties are pretty much the right way to go about in terms of difficulty, and the devs actually did a pretty good job as the AI keeps its cheating quite subtle. It's not like in Shogun 2 where you wipe out a 1 province clan's army, and 3 turns later they come at you with a full stack of samurai.

Worth pointing out that the AI on the campaign map is HUGELY improved compared to past Total War games.

[Panfilo]

I really don't like the idea of disguising difficulty in the form of player handicap. Leadership buffs/penalties, public order buffs/penalties, and upkeep cost increases don't so much make the game harder per se as they just penalize the player more.

For example, on Easy/Normal it is viable to use Chaos/Vampiric corruption to cause attrition to enemy units, use abilities that wreck public order to trigger revolts in enemy cities, and use leadership reducing spells/abilities to cause enemies to run more often. But on Very Hard/Legendary these become almost useless which kind of takes away from that faction's theme.

[Red Bat]

引用自 Panfilo：

I really don't like the idea of disguising difficulty in the form of player handicap. Leadership buffs/penalties, public order buffs/penalties, and upkeep cost increases don't so much make the game harder per se as they just penalize the player more.

For example, on Easy/Normal it is viable to use Chaos/Vampiric corruption to cause attrition to enemy units, use abilities that wreck public order to trigger revolts in enemy cities, and use leadership reducing spells/abilities to cause enemies to run more often. But on Very Hard/Legendary these become almost useless which kind of takes away from that faction's theme.

I agree that this is a problem, but unless you have a solution to balance difficulty levels better that wouldn't involve quite a bit of dev time to implement, it's kind of a moot point. There are SOME things I think the AI shouldn't get a boost to (public order, corruption reduction, etc), but at the end of the day, player difficulty penalties and AI boosts are the only realistic ways you are going to have difficulty levels here. The question ends up being, which numbers to adjust and which ones to leave untouched?

I tried designing a simple wizard fighting video game a while back as a hobby. Trust me, designing an AI that is intelligent but not unbeatable or exploitable, was such a frusterating endeavor that I ended up just making the AI act randomly and giving each subsequent wizard better abilities to compensate, and then scrapped the project because the entire point of it was a proof of concept for differing AIs and thus the project no longer resembled what I intended. This wasn't a particularly complex game either, but if I had to finish it, designing the AI's would have taken several times longer than designing characters and balancing game mechanics. For strategy games the difficulty in implementing a good non-cheating AI that is still fun to fight and balanced with campaign objectives is exponentially harder and more time consuming to implement.