Sprites are still overpowered

[Catwalk]

As exhibit F (or however far I've gotten to), I submit this save game using version 3.3:

http://www.filedropper.com/save1

I've just banished Kali in August 1402 using a stack of 9 sprites, lost 3. I financed them by taking out two lairs with unicorns and guardian spirits. I'm about to breed halflings like rabbits across a lush continent, and each city will be producing double power from all religious buildings. That's (2+4+6)*2 = 24 power per city. They also produce double taxes, courtesy of Inquisitor. I've taken a single sorcery node, a nature node with 1-2 giant spiders and 1 great wyrm will go down in less than 10 turns.

This is the third game in a row where I've gotten off to an amazing start using this setup (I've played until 1403 or so).

Setup:
6 Nature
2 Spellweaver
1 Cult Leader
1 Inquisitor
1 Conjurer
1 Specialist

Inquisitor and Cult Leader are mostly for fun, they don't help me get off to a faster start but they help ensure a dominant mid-game. There's no way this game can be lost.

[Seravy]

A suggestion.
Try playing this exact same map but summon War Bears instead of sprites (or replace nature with other books and try ghouls - I don't think any other common summon is strong enough but two stacks of 9 of skeletons or hell hounds might work) and send a 9 stack of those into that capital. I'm pretty sure you'll still banish the wizard. Most likely also take the Sorcery node.

I'm not saying sprites are not good - they obviously are, and this isn't even a question. Every spell in the game needs to be good, or that's a problem in itself. The questions we need to answer are :
-Do they provide significantly more advantage than other common (or uncommon - you can push research and get spiders on turn 10!) creatures you can summon using the exact same retorts?
-Can they do that on their own, or do they need specific retorts or other factors for it to happen. If yes, are those other factors overpowered instead?
-If we decide to change sprites from a node hunter role to something else, what will be their intended role? As is, they are only good for that and nothing else. If we take away flight, they'll be useless. (ghouls deal more damage, have more health, and make undead, plus have lots of immunities - the only thing sprites have over them is flight. )
-Maybe the ratio of summoning vs normal units in the early game is the problem? For the first 20-30 turns, most players don't produce nonsettler units, but they can summon creatures - dozens of them even.
-If we do any changes to Sprites, how badly does that hurt the AI? As of now, Sprites are the most desired units for defending capitals, and one of the AI's main ways to leave their starting continent using military units in the early game. They are also very effective at harassing the human player and punishing them for leaving new settlements insufficiently defended.
-Is the AI's capital still too vulnerable too early? I think it's unrealistic to expect they can defend themselves in 1402,  the sawmill, settlers and fighter's guild probably takes that much time to build on their own. By definition, the game pretty much starts after 1403, the AI cannot even attack before that. Again, might be the problem that you can summon way earlier than build units.

Edit : let me answer at least one of those questions.
If you pick Conjurer, Specialist, and Spellweaver, which results in the ability to summon 3 times the normal amount of creatures per turn, it obviously breaks the summon to normal unit ratio of the early game - you get 3 times more summoned units than intended (You pay 0.5 cost but have 1.5 skill and 1.33 income). And these retorts pay both the mana and casting skill cost for those creatures. I have no good idea how to fix this though aside from outright forbidding the combination of these retorts. But the normal unit to summoning ratio of the early game might be broken even without retorts - A common summon can be done in 2-4 turns, which at best gets a swordmen and requires sacrificing "sawmill first" and/or "settlers first".

[Catwalk]

This would be completely impossible with War Bears:
1) I had two nodes early on with Guardian Spirits and Unicorns. I could whittle those down with just 2 sprites, giving me the crucial early economic boost.
2) Ditto with the sorcery node, I was able to take it with just two Sprites. I would have needed far more War Bears, and I would have had casualties.
3) Going from there, I could certainly win some battles with War Bears. But there'd be more casualties. With Sprites, I keep cruising with a mostly intact army that grows stronger with each turn.

You definitely need retorts for this to work. Conjurer and Specialist are the most important ones, taking Sprites from casting cost of 80 and upkeep of 2 to casting cost of 48 and upkeep of 1. But these retorts are not unbalanced. Weaken them and they'll be relatively uninteresting for other strategies. Spellweaver is also great for this, and Inquisitor + Cult Leader guarantees me long-term power. If anything, this strategy is even stronger than the previous one and it works regardless of node strength.

Changing sprites boosts the AI immensely. The AI is terrible at using them, and the human player can get huge advantages using them. Early flying + ranged is just a terrible idea. I can live with draconians having it, although I might be tempted to suggest removing draconian bowmen. Building units is slower than summoning units, but much of the same can be accomplished and the units get massively more powerful after turn 40-50.

No other summoned unit can get you these crucial early game advantages. I'm not going to replay this game with War Bears because it'd be pointless, but I certainly invite other players to do so. The best fix is just to make them Uncommon. Giant Spiders are also very powerful units (and an excellent way to consolidate your power after using Sprites in the early game), but they're balanced because they're uncommon. That's my preferred solution, and I'm guessing that's more acceptable to you than removing Flight.

[Seravy]

I'm not going to replay this game with War Bears because it'd be pointless, but I certainly invite other players to do so.

Okay, give me the turn 1 save file and I'll do it.

As I was editing the post while you replied, I'll copy what I added meanwhile here :

Edit : let me answer at least one of those questions.
If you pick Conjurer, Specialist, and Spellweaver, which results in the ability to summon 3 times the normal amount of creatures per turn, it obviously breaks the summon to normal unit ratio of the early game - you get 3 times more summoned units than intended (You pay 0.5 cost but have 1.5 skill and 1.33 income). And these retorts pay both the mana and casting skill cost for those creatures. I have no good idea how to fix this though aside from outright forbidding the combination of these retorts. But the normal unit to summoning ratio of the early game might be broken even without retorts - A common summon can be done in 2-4 turns, which at best gets a swordmen and requires sacrificing "sawmill first" and/or "settlers first".

Make them uncommon? That raises these questions in addition :
-Which uncommon Nature spell goes down to common? Last time I checked I found none that can be reduced.
-Taking away the AI's access to an early flying ranged creature massively weakens it. If it's uncommon, they can't pick it as a starting spell anymore, meaning they won't have it until, like, turn 40-50 at earliest. The AI doesn't spend on research in the first 30 or so turns at all. Not having a flying creature means they can't conquer neutrals on other continents and hit weak lairs on other continents during these turns. (yes there is water walking but there is a difference between summoning creatures which is done often, and buffing with ww which is done much less frequently.)
-As uncommon, what stats will be raise on the unit and by how much, and what additional abilities do we add? The current sprites are worthless for an uncommon, by that time a 5 hp unit is as good as dead against anyone.

[Nelphine]

I'm on the fence about sprites. I think catwalk is right, but I haven't been able to make them work with myrror. The difference in plane shouldn't be THAT large. But 3 games, and not enough advantage in a single game to make sprites worthwhile, since I can't make a strong mid game with them. My best game with this tactic was actually my 4th game, using maxed out draconian bownen, not sprites at all, and even then, I couldn't stand against the first impossible AI I fought - I was only winning maybe 50% of my battles and he was producing far faster than I was. And he was dwarf, arguably the worst myrran race to fight flying ranged units.

This I think actually points to myrran races being more important than I thought, which is a good thing.

[Catwalk]

I agree that this is not a strong strategy with Myrran, but why is that a bad thing? That's like saying this is not a good hero strategy You need to base Myrror strategies around your race. Also, Sprites will have fewer targets available on Myrror.

It's actually fun with heroes, I'm picking up a bunch of items. And heroes are so much better balanced without Heroism.

[GermanJoey]

I'm on the fence about sprites. I think catwalk is right, but I haven't been able to make them work with myrror. The difference in plane shouldn't be THAT large. But 3 games, and not enough advantage in a single game to make sprites worthwhile, since I can't make a strong mid game with them. My best game with this tactic was actually my 4th game, using maxed out draconian bownen, not sprites at all, and even then, I couldn't stand against the first impossible AI I fought - I was only winning maybe 50% of my battles and he was producing far faster than I was. And he was dwarf, arguably the worst myrran race to fight flying ranged units.

This I think actually points to myrran races being more important than I thought, which is a good thing.

Sprites aren't very good against wizards in the general case, I think, unless you're going against Gnolls or some such. They're more for just snowballing into a huge lead by cleaning up Nodes and Dungeons etc that should be far too difficult to kill without flying ranged units. Not every node is doable like this of course, but all it takes is one Great Wyrm + Earth Elemental node to break the game wide open.

Drac Bowmen are a lot better against wizards (especially since you can put 'em on airships) but Magicians (so awesome w/ Nature) and Halbers are much better, I think. And Doom Drakes, of course.

[GermanJoey]

I'm not going to replay this game with War Bears because it'd be pointless, but I certainly invite other players to do so. 

I agree. Killing off one AI early is a pretty insignificant event by itself; it's more about the fact that you probably steamrolled the guy and can thus continue to build up an unstoppable snowball by looting nodes, lairs, etc .

[Seravy]

As you haven't posted your turn 1 save file, I went and tried my own.
First try I didn't find any enemy wizards using about 6-8 Earth Lores - I had to conclude there is no other wizard in a reachable distance to conduct the test and started over.
Second try, my 9 War Bears banished the nearest enemy wizard in 1401 November, while also conquering a neutral barbarian city on the way. I wasn't even playing it well - I didn't pick Alchemy to covert gold at a more efficient rate, didn't pick Elves for the extra 2 mana, and even forgot to start summoning the next bear a few times. This wizard was playing pure death - the hardest possible enemy to beat using bears as bears have poor resistance.
I believe war bears this early using these retorts are unstoppable - ironically the only exception is if the AI has sprites in their capital, as that's the only common unit bears cannot attack and kill effectively, and casting 2 webs isn't enough to bring down a capital with 4-5 sprites.
The capital had 4 ghouls, a skeleton and a nomad spearmen, and Darkness was cast - I think if I had 9 sprites here they would have lost the battle, ghouls deal about the same damage as sprites but take a lot more to kill, and they get the first turn, plus there is the lightning.

There is an adjacent Sorcery node with "many phantom beasts", if I wasn't aiming for the enemy wizard, I could have conquered it - for a test I sent the same 9 bears there instead of the wizard and they killed 5 beasts and 1 naga, leaving only 1 beast and 2 nagas. So about 12 bears can clear the entire node, which are easy to summon - They only take about 1 turn to summon each. (6 beasts with 3 nagas is the highest value Sorcery node sprites can attack, btw, so any sprite compatible Sorcery node is compatible with Bears. Not true for Nature though.)

   

btw Bears are more than enough to beat unicorns and guardian spirits. I feel you are severely underestimating them.

The fact the AI had 4 ghouls in their city is enough proof that I could have also won if I had ghouls instead of the bears - 9 ghouls beat 4 ghouls.

...so, I'm awaiting suggestions on how to fix the "early summon spam" balance problem - sprites are merely a part of it - the retorts are fine on their own but together, they triple the unit output of a player. If that wasn't bad enough, summoning can be done on turn 1, but producing units can only start after building a sawmill which takes like a dozen turns, so even at an even cost summoning has an advantage of an entire year. Furthermore, summoned creatures are stronger by definition than normal units of the same tier, escalating this problem.

Oh, and I think I could have done it another ~6 turns earlier if I maxed out my bear summoning capacity - with Inquisitor for double gold, and/or Alchemy for no conversion loss, I could have had enough mana to fuel my entire casting skill every turn instead of "only" having enough for about 2/3 of it.

(Note - I didn't bother to reload when I misclicked on an enemy unit while moving, so they attacked the neutral city I had and took it. The yellow spot in the middle is that one.)

[zitro1987]

The easiest way to balance out the early summon spam is rather game-changing and quite unpopular I think, but I have quite a bit of experience with it:

the idea of starting normal units with +1 to hit and magic weapons not granting the hit bonus. While it balances certain concepts (mainly alchemist retort and barbarians) and essentially solves the early summon spam from experience with tinkering with this, it can introduce other issues, though mostly taken care of with lots of unit statistics/cost tweaks.