Experimental version

[Seravy]

Also, the 'guaranteed uncommon on turn 1' only shows as a non-intuitive 'available'. 'Early Research' may be more intuitive for player.

It has to be the same text for guaranteed and guaranteed first turn spells.

Your 'experimental 3' files version allows you to choose up to 9 common starting spells, but you get 5. It will need to be fixed on the selection display.

Confirmed. I thought I updated the commons but they were the old amount for some reason.
Oh... the editor I'm using has a "starting commons" field so every time I update the data, it restores the old numbers. I have to change this in the editor's data file too.

[Seravy]

As I don't expect to have the time to play a full game in the next day or two, I started a quick test game for sprites instead.

10 books, Specialist, Conjurer.


My observations :

-New sprites are boring. With the old one, you needed to know what is safe to attack and what is not, which rewarded better players. The current ones you can throw at anything any enemy wizard has except other sprites and win, as long as you have superior numbers.
-Unlike treasure from lairs which is just an advantage, they now conquer and eliminate wizards effortlessly. Not much to do against that - it's the "top" unit in the early game - comes from the best summon realm for the highest cost. A stack of 9 of these must be better than other summons. Previous sprites were specifically meant for neutral targets and didn't have this problem - fortress lightning chewed them up and they couldn't take on ranged units at all, not even with slightly superior numbers. War Bears don't have this problem - while they are even more powerful technically, they are walking and melee - wall of fire, city walls, other bears, nagas and ghouls with their poison hurt them and oceans slow them down, plus they can't attack sprites - so they are not unstoppable. Their resistance is fairly low so common spells can kill 1-2 each combat as well.
-The unit confuses the AI. It has too low attack to trigger "leave walls and attack" even if it's a stack of 9. Instead the units wait to be slaughtered because they don't consider a ranged of "2" enough of a ranged threat to ignore walls.

I'm heavily leaning towards rolling back to the old sprites unit - if everyone agrees the new ones are no good, I can do that for EXP4.
With all of that said, I only defeated one wizard using them - by the time I reached the second with my 9 stack, they had too strong defenses. While the sprites effortlessly melted away ultra elite halberdiers buffed by holy armor for a total of 11 shields, they still lose the battle as they didn't have enough ammo to kill everything, and they most likely wouldn't have been able to damage the 15 shield hero even if they did have ammo left to try. This is only for the capital though - any other city or stack of any wizard I could have destroyed if I wanted to.

Oh and I didn't even try lairs or nodes.

(btw, "fairy dust attack" is already coded - we might want to use it for some other unit if not Sprites - or maybe not, it's quite...powerful. We can have it use other direct damage spells at other strength levels, though)

[Nelphine]

I'd consider dropping the sprites normal ranged attack to 2, and drop the fairy dust attack to 1 per figure. I'd also drop HP to 1, and I don't understand why slingers have lower armor than them. Increase slingers to 2 or drop sprites to 1. Sprites defense is flying - any ranged attack should wipe them.

[Catwalk]

The old unit is just as broken. Rolling back doesn't fix anything.

[Seravy]

I'd also drop HP to 1, and I don't understand why slingers have lower armor than them. Increase slingers to 2 or drop sprites to 1. Sprites defense is flying - any ranged attack should wipe them.

Then we are back to the old unit - the one that cannot be used against wizards. Except, it would also be weak in lairs - fairy dust (especially at 5) can only kill weak multifigure units. If the supposedly best common summon in the game is unusable against other players (yes, they work in overwhelming numbers but not if equal) and performs barely acceptable on lairs, that's very disappointing.

The old unit is just as broken. Rolling back doesn't fix anything.

We don't have an immediate good solution and the new unit is far more broken, so it shouldn't be left this way.
We'll be back to when we was - waiting for your analysis on treasure obtainable by the old version, and trying to come up with a solution in case it's needed. (I hope not, at this point I'm pretty sure one does not exist)

[Nelphine]

Why try to make sprites the best? Make war bears the best, since they are ground melee. Make sprites able to beat war bears because they fly and have ranged. Tune sprites specifically around that. Make them cheaper than war bears, so that it makes some sense that it takes 8 shots from sprites to kill war bears. Then they become a good niche unit - ai can use them on defense, but they don't destroy everything due to flying + ranged shenanigans.

War bears beat hell hounds, ghouls, nagas. Hell hounds and ghouls beat sprites. Nagas should be around the same as hell hounds or ghouls (make nagas poison immune if they aren't already). Sprites beat war bears. Sprites and nagas are intercontinental.

[Seravy]

Why try to make sprites the best? Make war bears the best, since they are ground melee.

Wait, are you telling me to make War Bears even stronger?
Or just keep them the same and make sprites cost 40 and be completely worthless?

Make sprites able to beat war bears because they fly and have ranged

How, exactly? If sprites are cheaper and weaker, they won't win - bears have armor. Even as is, sprites won't take out a bear one on one - bears have too much hit points and armor. You need multiple battles or more sprites. If we make them weaker (and maybe also make bears stronger), they'd be outright ineffective against them.

Then they become a good niche unit - ai can use them on defense,

Erhm, no. If Sprites are weaker than the old version, it won't be an effective defending unit anymore. An army of 9 bowmen would kill them as they don't do enough damage.

[Nelphine]

Right. Against bowman, you use war bears. 'what, the ai can't tell what's coming after them?' True. Get the AI to have a nice mix. And by that I mean war bears + at least 3 ranged units.

But yes, make War Bears cost 100, with 2 maintenance. But put most of it into hp and resistance. maybe even go back to 2 figures so you can increase attack power. Make Sprites around 60, with 2 maintenance. (Or you can leave war bears at current stats, but drop sprites down to 60. Since they fly, they can still be much worse. And with my suggestion of 2 attack and fairy dust strength = to figures, 2 units of sprites would kill 1 unit of war bears; yes it would take multiple combats to properly win, but the war bears could never beat the sprites. Which would give the split i'm looking for. War Bears can't beat sprites, sprites can't beat war bears with equal expenditure.)

You don't want a 'one common unit wins all the time' unit. That means that unit defeats AI. You specifically want all the commons to effectively negate each other, so that the human can't do early rush tactics. (For instance, even though war bears beat naga and hell hounds and ghouls, they should do so by such a small margin that a very weak, say 10 casting skill, lightning per round, will beat the war bears.)

[Seravy]

Get the AI to have a nice mix. And by that I mean war bears + at least 3 ranged units.

That's not doable. The AI simply picks the unit with the lowest "garrison rating" and kicks it out every time it produces a 10th unit. Such a tactic wouldn't be universally good either, only for this one case.

And with my suggestion of 2 attack and fairy dust strength = to figures, 2 units of sprites would kill 1 unit of war bears; yes it would take multiple combats to properly win, but the war bears could never beat the sprites.

no, this isn't how it works!
In reality, you lose one sprite in the first combat to a fire bolt (or life drain, or fairy dust, or psionic blast or star fires or call centaurs or...I think you get it now.), then lose the other in the second combat because it failed to kill the bear in the first. You suggested 1 hit points so assume a unit of sprites die in every combat.
Such sprites are unusable against wizards whatsoever unless you use a very large stack and can afford to ignore the loss of 1-2 units every combat - even against spearmen.

[Nelphine]

Right. But that has nothing to do with the war bears. That's just the wizard supporting them. I don't see a problem with this.

The Sprites, as they stand, can't be hurt by the war bears (or many other things). Against lairs, this means auto wins, and with strong attacks, fast auto wins, that allow you to expand very quickly.

Against wizards this means that the wizard must have a combat spell capable of defeating them, or the sprites will eventually wear down the war bears. AI know how to do this; not all players do (or can, depending on realm.) When the AI attacks with sprites, or the player attacks sprites with war bears, the player will be punished for it; but the player will not be able to abuse it. The player won't be able to simply defend with sprites and destroy everything; but if they are being attacked by war bears (or some other strong melee ground unit, such as my vaunted buffed halberdiers) the sprites will be immune to the units, and only vulnerable to the wizard.

Sprites don't become a unit that should be regularly seen. They become a scout, they become something that can take weak lairs, they become a unit that punishes player mistakes (so the AI should still summon some of them, but not a lot.) Basically, War Bears are nature's common summon. Sprites are a utility spell, similar to earth lore, or focus magic. It just happens to be that Nature uses a unit as a utility.


On topic of how garrisons work - I'm aware of this. But, in a game this complex, any time you simply say 'this unit is the best, so the ai can just make them', then the players can abuse that, especially in the early game. This game has to be a game of checks and balances. If you can't teach the AI that for garrison defense, then I think you need to teach them to research better units faster. It will be risky in it's own way, but right now, despite the AI summoning vast armies of units, it's relatively simple to avoid most of those meaning anything. Summoning less things to research faster won't make a large difference to the player, most of the time. Most of the games I play (even the ones I lose, as long as I don't lose super fast), I have very rares in play, and the AI still consistently summon nagas and unicorns and ghouls constantly. Those cheap summons are not what defeat me, they mostly just clutter up the map; the AI can afford to try to get better, faster.