Current Impossible Strategy

[coldsteel]

So I continued trying to develop a 6 life 2 sorcery (floating islands) warlord/alchemy/tactician horsebow strategy that does not conquer any enemy capitals early - because sometimes you just aren't able to spike anybody with any race.  The idea was get war college ASAP and churn out an elite horsebow every turn all game.  Go to war with everyone whenever, and take their non-capital cities and raze outposts.  You can actually pull this off sometimes on fair and i presume smaller map settings and surge to the lead in the F4 power graph.  Not often enough.  On Huge it's tougher and I don't remember if it actually worked once.  My final game I tried to gear researching prayer to coincide with reaching skill 40 by not spending much on neutral buildings, but prayer still hadn't come onto the research screen in mid-05.  And only my capital was suitable for producing troops that game and you need more than one city producing a unit every turn.  And horsebows in 06 and later get fire rained on them, go up against black prayer + prayer, focus magic cockatrices, all sorts of nasty stuff.  Stacks get damaged and are worthless for multiple turns.

I did try omniscient for the first time today with life+death and that retort is intriguing.  Perhaps add a chaos book.  I suppose go with barbarians and follow the Nelphine strategy is best overall.  I wonder why Nelphine doesn't use Tactician?  It would take months I suppose to learn all that strategy, and possibly still not gain any insight to contribute that is not already known!

But I did hear that with 2 elite berzerker cities a win on impossible is in the offing.  Perhaps even I could figure out how to do that much most games by experimenting with what kills what in lairs and nodes...

[Nelphine]

I'm currently trying Tactician.

(My last game was: 4 life, 1 death, warlord, alchemy, astrologer, tactician, spellweaver)

I think specialist is probably more important than somethign on that list, for those cases you face dispelling wave. (Although dispelling wave often comes with aether binding which is why I'm not sure. Aether binding + Dispelling Wave is stupidly powerful.) 1 chaos book for flame blade is also helpful, but I'm not sure how important it is. (At the same time, flameblade + lionheart + holy weapon + crusade means that 6 bezerkers can kill 7 neutral hydra + 2 chimera in a chaos node without losing any bezerkers)

Tactician: The basic problem with tactician is that bezerkers are designed to just trash everything before it matters what damage they take. So, only spells and ranged attacks mean anything; you avoid going to war with AI until you're overpoweringly ahead, so spells don't mean anything until you can bring whatever spells you need to counter them. You bring 8 units of bezerkers (sometimes 7, if you have a warship and a hero), so you can afford to lose some. And with holy armor and ultra elite bezerkers, tactician is only a 1/8 increase in armor (from 8 to 9). Do you want to spend a pick on a 1/8 increase, especially when that increase doesn't affect the important part of bezerkers (the offensive ability to kill anything they touch)?

My '2 cities wins impossible' is probably not a good goal to go for; you really should aim for 4 cities, otherwise mistakes or bad luck can ruin you. (For instance, if you constantly have to sacrifice 4-5 bezerkers because too many lairs are against ranged enemies.)

I'm still not really sure of the best exact build for barbarians. I know it includes 4 life, warlord, alchemy. I believe Astrologer is also on the must have, but I'm very node centric and I play on max power. Beyond that, you need to pick things that suit your playstyle; I still lose games with it. Things I would consider: 1-2 death, 1-2 chaos, 1-2 sorcery, tactician, omniscient, spellweaver, specialist, archmage, cult leader. I wish I could put channeler on this list because it is phenomenal late game, but it hurts too much in the early game.

[Nelphine]

Playing against a Sorcery (6 books) Life (7 books) Alchemy Lizardman wizard.

Its 1407. My opponent has over 200 casting skill already. He's ahead of me, despite my 6 nodes and all my usual shenanigans. Impossible sorcery AI get INCREDIBLE spell casting skill.

That being said, my Myrran opponent (4 nature/4 death, alchemy, warlord, myrran, charismatic; peaceful perfectionist) this game is doing TERRIBLE. As far as I can tell, no one else got to Myrror (via astral gate or planar travel or towers). But the Myrran AI (who has warlord and alchemy and is draconian!) has the weakest army strength (less than half the 4th place player; less than 1/4 of my own), the weakest power production, and the weakest spell power. I don't really understand it.

[Seravy]

Lizardmen AI tends to perform significantly better than others. I haven't collected enough data for a conclusion but so far there was no Lizardmen AI that performed worse than 2nd best in the game in the first 200 turns. (and when they did, first was the Myrran AI - but half the time the Lizard outperformed it, sometimes significantly.)
Lizardmen might still be overpowered.

[Nelphine]

I believe it's the settler speed. For a human player it's strong, but it's not incredibly strong. For an ai, that won't use anything like ships (except in the very rare cases where the AI has no other option for intercontinental settlers) or endurance on their settlers, lizardman settler speed literally doubles how fast they get cities.

The high population growth helps, but I think all that does is give the AI vast amounts of money to buy amplifying towers with; the settler speed is the real difference.

If you could teach the ai to prefer to transport settlers on ships/windwalkers, I think the difference between lizardmen and other AI races would be drastically reduced. Lizardmen would be strong early, but the ship speed would eventually let the other AI catch them.

[Seravy]

By the time other wizards get their first ships, the lizardmen probably fills most of the map with cities.
Once I have enough data to conclude lizardmen are overpowered, I believe this might be the best way to solve it :

Swap the growth rate of Klackons with Lizardmen.

That way Lizardmen still have strong early military and quick expansion, but more average economy to go with it...and klackons get a better economy - and more people to take advantage of the "less unrest" effect. Also, bugs breed fast, would make sense. Lizards, idk but I think they don't breed any faster than other animals.

[Nelphine]

i like that idea

[Nelphine]

Why life is the best realm for impossible.

There are lots of reasons - countering chaos and death is huge, but each realm has its own advantages.

However, the reason I love life, is retorts vs spell books. What do I mean?

Retorts amplify your power. The more cities, troops, spells, summons, heroes, power, anything you get, retorts amplify that. So the strongest strategies are the ones with the most retorts.

What does this have to do with realm choice? Spells are still the most important thing in the game. The right spell at the right time wins the game. And generally very rare spells are the most powerful. From this I conclude I must have at least 4 books from one realm.

What does this have to do with life though? Two things: realm strength and realm weakness.

Realm strength: life focuses on buffs for city troops. This means you can put both production and skill into the same unit, ending up with a unit stronger than just production or just casting skill. And, it negates the fact that the realm has very few summons - you're going to use city troops anyway!

Realm weakness: very few summoning and combat spells. But this isn't a weakness. This is a HUGE strength. There are virtually no spells in the life realm that make another life spell obsolete. Angels could be considered obsolete if you get archangels, although even then angels have their moments. And prayer is obsolete if you get high prayer. That's it. EVERY other spell remains important throughout the whole game.

Compare to any other realm: every time you get a better summon, the worse ones don't get used anymore. You don't use ghouls after you get shadow demons. You don't use great lizards after you get gorgons. Etc. Similarly combat spells - you don't use fire bolt after you get lightning bolt. Even curses make other curses obsolete - the AI (besides being less bothered by curses than the human) is smart enough to dispel curses. This means curses will often only last a round or two. Which means the curse has to be dangerous in that short time, and usually you won't get a chance to stack curses - so you can only use the strongest one.

This means, if youre planning on playing impossible, you have to assume you will get bad spells. For every other realm, you have to assume some of your spells will be obsolete - which means, assuming all realms are equal, you need more spell books of any other realms to get just as many good spells as you get with life.

And then, life + retorts + city troops means a very strong early game. Which means you can clear a lot of nodes and lairs and towers (on impossible I've cleared 29 nodes and all 6 towers before the AI, and I regularly clear at least 25 nodes/towers). That means I find a lot of spells. So even though I only have 4 books in the realm, I can reliably expect to end up with at least 7 very rare spells, and I average somewhere around 8 very rare spells (yes getting all 10 very rares happens).

So life not only makes sure that I will be able to use all the spells I can get, for the entire game, but it also gives me the best chance of getting lots of very rare spells, even when I only have 4 books. Which allows me to take as many retorts as possible to amplify my power.

I've also been getting archangel often enough that I'm going to drop death spellbooks entirely. Wraithform just isn't needed often enough.

Edit: which brings up another point. Multiple realms on impossible is bad, because no matter how many spells books you have, you're still going to find (roughly) the same number of spells from treasure. If you have multiple realms, then those spells are split over multiple realms, making it less likely that you'll get the best spells from the realm you want. That being said, 2 books from a second realm isn't bad. You can usually find enough common and uncommon spells to cover 2 realms. Although the game where I had 6 nature, 1 chaos, 1 death, and I never got water walking or wraithform... That was annoying.. So yeah, less realms is better.

[Arnuz]

Makes sense, but I still disagree because I consider that once you have archangels or that level of stuff summoned you've won anyhow, and since a long time.

The most important part as you also state is the beginning, and at the beginning obsolescence is not a factor.

So I consider the multi spellbook start the best one... I have to get to writing a report of my last game, no need for rare spells if you can grab chaos spawn and other crap from lairs and nodes

[Nelphine]

As a note, most of my games I have a very rare cast before I finish researching commons. For instance, in my current game I had my first archangel in 1405. Sure this was exceptionally lucky (it could have been call to arms or life force), but I hadn't even got off my home continent before my first one was summoned (only 8 cities can fit on my home continent.)

4 of a realm allows for very rares in treasure; life is the only realm where virtually all the spells will synergize with the basic strategy.