Spellweaver

[Nelphine]

It's far more hero centric. You could get demigods before fighting the last ai, making heroes, particularly with famous, dominate strategy.

It also has an economy benefit that warlord doesn't. 

No idea how I feel about (since I usually call heroes overpowered), although I like the economy side as being comparable to channeler.

Would also help the AI get strong heroes that were actually threatening, particularly in their stronghold.

I'm not against the idea actually.

[Seravy]

Oh I missed the halved maintenance part, well that has been rejected already before since the AI can't use it. In fact this is entirely a retort the AI has no use for.

(sorry for rejecting so many ideas but we don't want a retort we'll end up removing again later.)

So, we don't want something that's early game specific, and also don't want late game specific (would be nice but it's not very playable at 2 picks and a handicap for the AI on Arcanus while an unfair bonus on Myrror, and just generally bad design, remember when we had "Sorcery is unbeatable in the late game" as the design, it doesn't work.)

Maybe we could instead have something that is strong in the midgame? I have no particular ideas what that could be though.

[Ryzel]

Conjured requires you to use summoning spells to get the benefit, whereas my idea would get the benefit for any overland spell. Life might want to take this retort, whereas it wouldn’t want Conjurer. By redundant do you mean that they both help to ultimately reduce cost of stuff? By that logic Conjurer would have just as much overlap with Archmage.

But whatever. Other ideas...

Populous
Max growth + growth rate of your cities increased by X

Open Mind
You are able to obtain spells from outside of your book limitations, including from other schools, through trade and treasure.

Combat Summoning
Every time you cast a combat spell, you summon a creature from your school for free to place next to one of your units. The unit summoned is based off the casting cost of the combat spell. This unit disappears at the end of combat.
(Alternatively you can replace free summon with free unit enchantment)

Master Engineer
All of your units are capable of constructing roads, complete them in 1 turn regardless of terrain, and receive double the movement bonus from them.

I’ll post more later.

Is midgame defined as the point where you’re conquering your home plane?

[Seravy]

Is midgame defined as the point where you’re conquering your home plane?

Yes, I'd say it's when you fight the second and maybe third enemy wizard, like 1409-1414?

Didn't notice it's for every spell, not just summons. Still, a player's casting capacity is so much more limited than their production capacity, I don't think this is worth a retort slot. (you can cast like 150-200 MP a turn while you have like 800 hammers of production in the same empire. So getting like 100 production worth of units for each 200 MP would be not that great.

However, if we could change it to specifically boost the city with your summoning circle without having to actually do anything, that could be interesting. Like, the city with your summoning circle gains production equal to half your casting skill each turn. Problem with it, too strong for the early game, and too good for the human while not that good for the AI (who doesn't know where to focus their production).

Populous
Already decided we don't want more max population.

Open Mind
This is worth like all your 12 picks and more...

Combat Summon
Not every realm gets a combat summon, especially not enough of them to base it on costs, same for enchantments. I can see it working for enchantments more than summons for sure though, but a random enchantment isn't much of a unit and I don't see how we could allow the player to choose one.

Engineer
This breaks the AI.

[Ryzel]

Battlemage
The first combat spell you cast in a battle is free. This also applies to all caster units capable of casting a combat spell, with the exception of Doom Bolt (the wizard can cast it for free though).

[Juffos]

The experience bonus gain could work as a 1-point pick, as a cheap replacement-sort-of for warlord.

[Juffos]

If you wish for a middle game retort, it should be about built units and armies, as early game is about summons and buffs and you need to build the armies to secure your position before going for amplifiers. A global coal bonus, applied for all cities? Resource gain for military buildings? Resource rebate for built units? A +10 pop grow speed bonus for each unit of garrison? Additional resource production for garrison units?

I was considering some trade bonus gold thing but it would step on the toes of too many races.

[Seravy]

Generally, military related resources are an early game bonus as it is very well possible to do military early instead of, or in addition to, summoning for many races. They also lead to stacking with warlord/redundancy with warlord so we want to avoid that. We already discarded the resource for military idea yesterday.

As this will probably take a while (we haven't been able to come up with anything appropriate for years, not even when we tried), I think I'm going to do this as a temporal solution :

Spellweaver, 2 picks
Your non-summoning, non-artifact overland spells cost 33% less to cast.

It's easy to implement, and definitely less overpowered than what we have now, while still being at least somewhat useful.
Certainly, it has some redundancy but
-Redundancy with Archmage is out, as they remain mutually exclusive (since it's a temporal solution, I won't bother removing that)
-Cheaper recasts and dispel resistance have redundancy but they are also very distinct for certain uses, for example when you don't expect anyone to dispel you (peaceful strategies, city enchantments, change terrain/raise volcano/etc)
-Redundancy and stacking with Conjurer or Artificer is explicitly prevented.
-Might be worth the 2 picks for strategies that aren't summon heavy but still want to cast more, +50% spells is not as great for 2 picks as +33% summons for 1, but the scope of affected spell types is bigger.
-Does still benefit Death wizards so we don't need to redesign predefined wizards at all. (might want to swap wizard templates though, we probably don't want the AI to default their picks to cheaper curses.)
(affects Drain Power, Lycanthropy, Cloud of Shadows, the curses, Dark Rituals, Final Wave and the globals. Not that great, but for a default, noncustom wizard, it'll do I think. Overall, 12 spells in the realm.)

[Nelphine]

I'm not sure I agree with that. In particular, that's the same as current spellweaver, except it doesn't apply to summons/artifcats (which is fine), but also doesn't do the power bonus (which was a very heated argument when we implemented spellweaver). I don't think any of the reasons for the power portion have changed. I'd rather see what you suggested, but make it 25% instead of 33%, and still leave a portion of the power production (15-25%).

Also your own comparison to conjuror make it seem like +50% of non summon/artifact spells is roughtly equivalent to +33% summons; which means what you wrote should actually only be 1 pick, not 2. (I faintly disagree with this, but conjuror has other bonuses in addition, so it's actually about right.)

Also by lowering the overland skill portion, you could make it stack with archmage; yes it would still be strong, but having one exact pair of retorts with that limitation has always felt forced. Not going to try to convince you to do this, just saying you could.

[Seravy]

So exactly why did we need the power production?
If it's a cost reduction instead of a skill increase, you are actually paying that much less mana crystals so don't need extra income to cast the additional spells.

(Also, power production bonus stacks in a really nasty way with the spell group that likely benefits most, economy spells. Cheaper Dark Rituals that ALSO produce more power? Same for volcanoes, sanctifies, uranus' blessings, not sure what else. Sounds risky.)

Summons get +33% for 1 pick.
Nonsummons get +25% for 1 pick (50% for two), but nonsummons is more versatile and covers more uses. So I can justify it being lower. Especially as nonsummons are basically making you stronger without fighting while summons mean you need to actually go to war.