Caster of Magic Release thread : latest version 6.06!

[Seravy]

No, it should be the same. 3 Farmers are 6 food, but you need 7 so that isn't supposed to work even for barbarians. So there might be a bug? I'll have to investigate tomorrow. Just a guess but the cutoff between ships and normal units might be misplaced? If yes that would mean all barbarian units don't eat food except the berserkers...

No, that doesn't add up. Then you'd have +1 food because 4 units don't need any.
So it means it's the swordsmen who are not needing the food, not the settlers.

...except, not. It seems the settler is not eating food because it's a "generic race" unit and instead of "unit ID<X", food consumption is maybe set up as "race=generic" not needing any? Oh well, I guess no point trying to guess, I'll take a look when I can. Should be easy to fix, just not at 3am :D

[Nelphine]

Potential problem with new research costs: ai with: 5 chaos, 4 death, 2 nature, 1 life, specialist, archmage, on max power, lunatic, lizardman. Has the highest power production of all 5 wizards (~450-500 power production). He hasn't been at war more than 2 years , and not at all in the last 2 years.

July 1408.  This wizard still has 4 sprites guarding his fortress (and 2 dragon turtles, a crusader, the dwarf hero, and a jackal rider).

The slowed research seems to have increased how long sprites are kept as fortress garrison's, which is just bad - it was bad enough in 1406, but 1408 is awful.

[Nelphine]

Also, on lunatic, third place AI has wraiths in January 1409. Just to give a time frame to rares. (I don't yet have rares researched.)

[Seravy]

Those are some pretty random books, that do not guarantee an uncommon creature. Maybe they didn't get any uncommon creature spell? If you upload the save I can look at it.

[Domon]

even if ai research is too slow, i think the general idea is good and we probably just need to adjust the related values somehow. like maybe -5% casting advantage, +10% research, i dunno. but the new research pacing is a solid change.

[Nelphine]

If nothing else they have ghouls which are better than sprites. They also have cockatrices floating around, but I'm not sure if they actually know the spell or if they are undead. But agreed they might not know any uncommon summons.

Either way, sprites are good super early (before 1405) because the human might not be able to deal with flight. After that, I think assuming the human can deal with flight is much safer, and other AI of course completely ignore flight. Sprites are the biggest reason AI banish/defeat each other.

[Seravy]

Ghouls and cockatrices both have a lower garrison priority than sprites. It only takes some high resistance units to beat either, like a single charmed hero. Sprites can at least do some damage regardless of resistance and fly (albeit they are weak to resist elements, but not to bless.) so overall they are a safer pick.
To replaces sprites, the AI needs to research Shadow Demons, Gargoyles, Fire Giants, Chimeras or Great Lizards considering what they might have from those books, otherwise they won't get anything until rares. In theory Cockatrices shouldn't be able to hinder the AI in researching either of these, but they might simply fail to show up in research until the AI researched several other uncommons.

[Nelphine]

Yes you can specialize against ghouls - you need high resistance, but also high HP due to lightning. And we specifically made sure there arent a lot of high resistance units early. There aren't a lot of those. Anything that specializes against fortress lightning, by definition, also makes them effective against sprites. Ghouls are great.

Yes if you make ghouls higher than sprites, then charmed heroes can take more non fortresses. In order to stop things from rampaging through sprites and banishing wizards, like bowmen, that is HUGELY worth it. Ghouls absolutely should be higher priority than sprites for garrisons in all cities (sprites are also better doomstack units than ghouls). For the fortress specifically, cockatrices should also be higher than sprites (they're better than sprites everywhere, but they're amazing doomstsck units, so I wouldn't want to lose that many by making them better garrisons everywhere.)

[Seravy]

Depends...against the human sprites MIGHT be better outside (albeit they are very likely to not have enough strategic strength to actually attack things) but if the human is not involved, ghouls are infinitely better - several times higher strategic strength, plus undead creation.

Ghouls are better against Fire Storm though which is one reason to swap the priority. I'm a bit worried there is a lot of snowball potential lost if the wizard is not using the ghouls to clear neutrals and get undead.

Cockatrices however are awful - their high strategic strength means they will stay attacking and lose the battle even if sprites would stay in a corner and let the lightning finish the enemy unit that is too hard to kill. But the main problem is this is a constant, so it can't be "fortress only".

Best might be to make the ghouls and sprites equal so the AI mixes them but how often do they even have both? I don't think the AI actually picks these two as starting picks or researches one if they know the other until directly before they get their uncommon summon....let's see. Actually the AI only picks both ghouls and sprites if they have exactly 2 or 3 Nature books which is...interesting. At 4 Nature books the AI does not pick Ghouls as it can have War Bears instead (not reliant on resistance, better strategic strength, better cost efficiency, faster overland movement, but no undead) but it never skips sprites. Maybe it should if it already picked Ghouls? Ghouls can't fly though so they aren't doomstack material and can't conquer neutrals on other continents. So Sprites are still useful even if Ghouls were picked. But which is better for the fortress is something that by game and only a human can decide (ghouls if your targets are on the other continent, sprites if they are on the same one)

[Nelphine]

The problem is weak fortresses.

Against AI, strategic strength wins. Cockatrices > war bears > ghouls > sprites.

Against humans, the human has yo defeat the fortress lightning and the units. If the human is making a fortress strike, then it becomes can the units + lightning can kill the human units first. 

Very early in the game, the human probably can't deal with fliers, while also surviving the lightning - most units that can hit fliers. But by 1404-1406, it is completely reasonable the human will attack with units that can kill fliers faster than lightning can kil them - for instance elite bowmen.

At this point the flying advantage means nothing. Which means standard survivability is more important, while still hurting the enemy.

Cockatrices > ghouls > war bears > sprites.

Sprites might be a weak attacker, but they can fly. Ghouls and war bears can't. So sprites can roam around in stacks of 9 and attack weak cities or nodes. And they're very good at it.

So no I don't think ghouls should be equal to sprites. Ghouls are flat better than sprites for all garrisons. And for the fortress, even war bears are better than sprites. Watching AI get banished by other AI halberdiers ...