As a French person I feel like it's my duty to explain strikes to you. - AdrienIer

Create an account  

 
Caster of Magic Release thread : latest version 6.06!

Here is the starting save:

https://anonfile.com/P5aaIbffb7/SAVE1.GAM

I took out some ships that might have had a spider or two, but it has been sprites and water walking war bears the whole game.

If the sprites aren't better defenders or attackers, why are they defending the crown jewels?
Reply

Because the AI follows coded rules. It can't look at what you have and make a smart choice. It literally has to be planned before the game even starts.

And basically, in some situations, sprites are better (against strong melee for instance - gnolls, high men, trolls) and in some situations the spiders are better.

But on offense spiders are always better. So if the AI literally had no idea which is better on defense, but knows spiders are better on offense, it has to go with spiders and offense. Which leaves sprites on defense by default, not because it chose them. (Against other ai this is terrible which is why I'm against it. Against humans *shrug* the human will look at what the AI has and win regardless. The human CAN choose it's attackers based on the defenders.)
Reply

I've beaten Normal and am trying Fair.  Wish me luck : )

My current build is:
Retorts:
  Archmage
  Artificer
  Omniscient
Books:
   2 Life
   2 Death
   1 Chaos
   1 Nature
   3 Sorcery
Spells:
   L:Healing
   D:Skeletons
   S:Phantom Warriors
   S:Focus Magic
Race:
   High Elf

I'm experimenting with spreading my books out to get the most out of Omniscient. Let's see if it's worth it  : D
Reply

What do you put focus magic on smile use ghouls.
Reply

but skellies make such good policemen... It's hard to choose anything else
Reply

ghouls create undead. undead are the perfect garrison.
Reply

I've got well over a hundred hours out of this mod so far and expect a few hundred more. Thank you so much for making this amazing addition to one of my all-time favorite games. In my opinion, every aspect of the original has been improved upon. Due to the AI improvements however, some serious balance issues have appeared. Since the new AI generally makes good decisions, some combinations of spells and units have become extremely difficult to win against unless you have a perfectly suited build. Since the player has limited ability to modify their picks after the game has started, I wanted to offer some suggestions to prevent unavoidably impossible situations from occurring quite as often. This post may get lengthy, but I wanted to make a case for preventing a few nasty situations I have encountered during my playtime.


How to Win in Caster of Magic
I'll start with an explanation of how I secure a victory, so it's easier to understand why some situations are so difficult to deal with. I play on Master difficulty with no reloading. A winning build in this mod needs to be able to quickly take out neighboring neutrals, nodes and dungeons with minimal unit losses and then transition into a force that can take out enemy cities one after another with minimal unit losses. I won't go into the details of the exploration phase since it's more or less balanced. There are however only three ways to win the mid to late game that I have discovered: heroes, regeneration and big summons.
  • Heroes: I'm really glad that Heroes are now something you can actively rely on, much unlike the original game. Using heroes requires at least two of Tactician, Warlord and Famous complemented by magic items from Artificer or the more chance based Charismatic + Famous combo. Almost impossible without at least one Life book for healing, but more on this later.
  • Regeneration: Trolls, Werewolves, Shadow Demons, and Giant Lizards. Get a big stack of them, pump them with spells and don't get wiped. More vulnerable than Heroes and requires at least five books for all except Trolls. Does not require life books however.
  • Big Summons: Very Rare summons only in fact. Everything else is not sustainable enough to make it. Requires Specialist, Conjurer and Sage Master. This can also be a transition from the summon based regeneration strategy. Typically requires healing to work, but with no more than two Life books or nothing but Life books to prevent the research pool from getting diluted. High risk since you can't guarantee that you'll get a very rare summoning spell at all.
Other strategies that used to work in the original to win Impossible difficulty games no long apply. 90% of the time the game is won using primarily heroes, which again I consider a welcome change, but not one without complications.


Game Stoppers
These units and spells make the game unbeatable if you didn't start with a build that counters them.

Warlocks: If the new smart AI has access to Dark Elves, it will eventually fill every city with 9 units of Warlocks. If you do not have access to Fire Storm or invisible units that the AI cannot find a way to make visible in a single turn, you will have a nearly impossible time winning that game. Doom Bolt is a strong but balanced spell when it's cast no more than a couple times a turn, 9 instances at once in every city is simply not something that can be beaten. Worse yet this situation usually happens in the late game after you've defeated a couple of wizards and went through a tower just to discover that the fourth wizard spawned as Dark Elves and is impossible to beat. I highly urge you to make Warlocks into just a strong ranged caster with same spell access as regular mages.

Air Elemental: A scary enough unit in its own right, the new battle AI has made this unit impossible to beat without True Sight or access to undead. Since Air Elementals run away when a battle can't be won and can be summoned mid-battle, you will lose EVERY battle if you don't have an answer. Invisibility in general is difficult to deal with in this mod, so I suggest adding countermeasures to a couple of other spell schools. Perhaps a see invisible mutation in the Chaos Channels pool or having an invisibility dispelling effect in Earth to Mud?

Psionic Blast: Psionic Blast is here not because the spell's broken, but for the same reason as to why Warlocks made the list. Big stacks of mages each casting Psionic Blast will wipe out any unit without access to True Sight. Although not as oppressive as Doom Bolt, this is a more common situation to encounter and frequently just as game ending as Warlocks. I suggest increasing the mana cost of Psionic Blast out of range of everyday mages or decreasing their mana pool and the cost of fireball below the cost of Psionic Blast.

Crack Calls: The biggest problem with this spell is that it only has one counter. Every Nature mage that knows Crack Calls will also know Web making flight a non-issue. I've had to drop games 8 hours in, after an opponent wizard learned Crack Calls and I had no answer. Since the damage is irrecoverable this one spell will counter every one of your win conditions unless you have access to Wraith Form. I suggest adding countermeasures to either Web or Crack Calls. Perhaps Immolation can give a unit Web immunity while Pathfinding (from either Land Linking or innate ability) provides immunity to Crack Calls?

Wave of Despair: Similar issue to Crack Calls. You need Cold immunity from either Inner Fire, playing Werewolves as your win condition or hiring exactly Cirno. More sources of Cold Immunity across other schools of magic are needed, and I propose Resist Elements or Elemental Armor as the most likely candidates.


Extremely Annoying
Spells from the original game that were once a non-issue but are a menace in the hand of an AI that knows what it's doing.

Spell Blast: The new AI loves this spell. When an opponent gets their hands on Spell Blast, any spell you cast that takes four or more turns to complete will get Spell Blasted every time. It's non-interactive and extremely anti-fun since it stops all of the big spells you've been working towards from being played altogether. Feels awful to have to research it too, since it's not something you'd ever want to cast and it's most certainly not something you'd trade to another wizard. I honestly don't think this spell is salvageable, I would love to see it go.

Drain Power: Similar to Spell Blast this is a low interactivity spell that the AI spams every turn. Sometimes you're not going to be able to make up the hundreds of points of mana that you're losing every turn and will lose the game. If Aether Flux is in effect, expect to lose 500-1000 mana a turn. This is another spell I'd love to see gone altogether. I do however think it's also possible to remove the bully aspect of it and rework it into a spell that sacrifices time to create mana similar to how Artificer and Runemaster combo worked in the original game.


Possible Re-balance Suggestions
Units and spells that I think limit gameplay in one way or another.

Heal: Currently two out of the three winning strategies I've developed including the most common one hinge on this spell. I've tried playing without Heal, using Life Drain, Syphon Life, Nature's Cures, Vampiric Weapons or just stacking massive amounts of armor. Nothings really replaces it. So I feel that unless I'm playing with a regeneration strategy, I must take at least one life book, preferably two. Of all the spells currently in the game, nothing is as limiting to pick possibilities as Heal. I almost feel like this should be an Arcane spell or have an Arcane variant that only targets heroes and fantastic creatures.

Lightning Bolt: Easy to get and can be stupidly strong for its cost. Actually outdamages Warp Lightning and Doom Bolt from time to time. Very difficult to defend against, I've had a champion tier Hero with Elemental Armor and Resist Elements active get one-shot by this spell from full health. Just lowering the damage a bit might be enough of a fix, or applying armor from protective spells like Bless and Resist Elements after the armor piercing effect of Lightning Bolt.

Mystic Surge: Very random and potentially very scary. I actually love the flavor and effect of this spell, but the lucky roll outliers here are a little too crazy. Blood Lust and Haste on pikemen can turn even a 20+ armor Hero into mush before you get a turn. There's also a lot of dud spells like Water Walking which feel really bad to get if you're the one doing the spell casting. It also seems to give Orichalcum as a unit enchantment, which isn't a spell.

Light: Light magic in general is currently in a category of it's own as far as balancing is concerned. It's the strongest realm for human players and the weakest for AI players. While the other schools of magic have a healthy mix of summons, buffs, debuffs and direct damage spells, Light is centered almost entirely around the idea of making one giant, unstoppable unit and then protecting it. It's got access to health buffing effects that no other realm can do, it's got access to True Sight which shuts down an entire realm of magic, it's got powerful early game armor buffs and movement buffs that don't come into play until much later for other realms if they do at all. On the other hand AI players with nothing but Light books are the easiest thing to bully in the game. Every other realm has spells that you feel you have to watch out for, except Light. I propose making the Light realm give up a few of its "extremely good for human players" spells in favor of a few threat spells. Maybe a common "resist or die" spell similar to Black Sleep or Confusion, a semi-scary summon or a strong direct damage spell.

Doom Bat: Much of what was said about Warlocks can be restated for the new Doom Bats. Being able to hit anything before the Player gets a turn for unstoppable damage is very difficult to navigate around. Unlike with Warlocks, Fire Storm isn't even a counter in this case. The only reason they're in this bracket instead of the the "Game Stoppers" bracket is because they're a lot more difficult to make a big stack of. Reducing their movement by even one point fixes the problem entirely. I'd also prefer if they did less damage and came in multiple figures so that you could reduce the damage in stages. Sort of like a scarier version of Cockatrices.

Encounter Zone Defenders Balance: A Node guarded by 9 Earth Elemenals requires two bowman to clear and will give insanely good rewards, a cave guarded by 4 sprites will demolish 5 of the 8 bowmen you sent to take it and give you 150 gold. I'm not sure how easy this would be to change, but the game should value flying, fast, gaze/breath attack or ranged defenders a lot higher for determining rewards.


Thank you again Seravy for this wonderful mod.
You've done a fantastic job and quite possibly created my new favorite game.
I'm very much looking forward to any future releases.
Reply

(August 10th, 2018, 23:10)Impy Wrote: Thank you again Seravy for this wonderful mod.
You've done a fantastic job and quite possibly created my new favorite game.
I'm very much looking forward to any future releases.

I'm interested in a full reply to the post, but I'd just like to echo this sentiment : )
Reply

Hello, Mr. Seravy! Thank you very much for your work! I found the following bug or unfinished moment in the standard version, in an unofficial patch and in a Caster of Magic. Building a unit's route allows you to determine the shapes of land without opening the fog of war. Building a route does not allow you to pave the way where it does not exist. On the one hand this simplifies the opening of the map, on the other hand it gives some undue advantage in the beginning of the game. Such as it is too early to determine the necessary building and routes for the units and etc. If you find it necessary, then please correct this in an unofficial patch and in a Caster of Magic.

Also I want to draw your attention to the Artificer + Runemaster exploit.
http://www.realmsbeyond.net/forums/showt...p?tid=9341

I duplicate a message in the topic with an unofficial patch.
http://www.realmsbeyond.net/forums/showt...#pid684434

Thank you!

   
Reply

Thank you for your long, and detailed post!

I admit I'm playing Expert difficulty for testing purposes, but I had success with strategies other than described. I'm unsure how these would work on Master but there is a good chance we'll need to reduce AI advantages on the top two or three difficulty levels anyway to make sure more strategies are playable. Unfortunately I still have too much rebalancing work to do to make progress in further adjusting difficulty levels.

Heroes, regenerating stacks and big summons are definitely the primary way to win past the midgame.
However I also had success with :
Spell spam strategies - Use whatever unit you can mass produce cheaply, and highly damaging combat spells to wipe out several times more valuable armies easily. Chaos with Flame Strike works best for this but Death or even Sorcery and Nature can do strategies like this (mainly through combat summoning spells, and/or call lightning.)
Chaos globals - You don't need to actually conquer anything if you can destroy them economically without doing so. A Chaos Runemaster or Specialist with high Disjunction resistance casting Meteor Storm, Doomsday and Armageddon can win without ever attacking the Myrran wizard through Spell of Mastery (or they can attack the wizard once they have been weakened enough to win).
Wolf Rush - My last Gnoll test game was successful with this strategy  build Wolf Riders early in large quantity and use their insane fast movement speed to seek out and conquer all the early hamlets of the other three wizards. You'll still need another strategy for Myrror but you'll have more time to prepare for that than usual and you'll get much more treasure as the other Arcanus players won't be in a position to get any first. Alternately, if you were really fast, you might be able to break towers and catch the Myrran wizard only halfway developed and do the same to their cities (haven't tried this way yet).
Nelphine's high strategic strength strategy - spam units that have a high unit power and possibly even buff them further, making you the stronger wizard on the charts. The AI will then try to make wizard's pacts with you and you can easily conquer them one at a time using those high strength buffed units. Due to the high strength, your cities won't get attacked as the AI doesn't have a strong enough stack to do so.
City curses These can destroy a wizard on their own, although this strategy is being partially removed as we are rebalancing the curses. It's not fun when you instantly lose because an AI has them, so the new beta has different effects that are still useful but do not completely destroy a wizard's economy. However the very rare curses : Pestilence and Call the Void will still be capable of winning the game for you.
Peaceful Spell of Mastery Hard to achieve but if you manage to maintain peace, accelerate your research and get SoM first, as a Sorcery wizard, and cast it during a Time Stop, you can win the game without ever doing a single battle. A Life wizard version, maybe with Guardian, might also work well using the new Sanctify spell, but I haven't yet tried.

There are more strategies I haven't yet tried, like the Omniscient early game strategy that tries to take advantage of the extra early resources and lots of common spells to rush the Arcanus plane and conquer all of it ASAP then break to Myrror before they are ready.

Game Stoppers
Certainly the AI's role is to prevent the player from winning, so I'd say this isn't necessary a problem by itself. It does become a problem if the particular game stopper is too hard to counter however. I didn't think there is anything like that except Spell Ward currently in the game, but maybe some of your list qualifies.

Warlocks - I think giving them the 45 MP to use freely would be worse, as it gives access to a lot more nasty spells we don't want to see. Doom Bolts are a pain but they are very predictable - all you need is to use units of 25 or higher health pool. You'll lose three of them on turn one, but then the rest of the army can wipe out the warlocks easily. So Warlocks only make is costly to conquer a city, they do not prevent it completely, meaning they are not a game stopper. I admit normal units are usually not good enough for this and you need at least uncommon summons, to reach that 25 health.

Air Elemental - This one is definitely a major pain. However they are fairly expensive to summon and weak in actual melee combat, so they aren't a threat until the wizard has a high casting skill to spam them. Unfortunately, Sorcery wizards tend to have that high casting skill. So I think this qualifies as a "Game Stopper". However if you are on offense, you can wait it out and then attack next turn when the city is already empty. On defense, Sorcery stacks generally aren't high power so a big enough garrison strength can prevent them from attacking, thus they won't be able to summon any. But expecting this from players to counter the spell is way too much.
Sadly, we can't "dispel" invisibility as it's a hardcoded native ability that was implemented in a way that doesn't make it possible to remove it unless we also change all those parts of the coding (several Invisibility checks look at the base unit type as far as I remember, not the current unit.) - we also can't have a new Chaos Channels effect as there are no free bits in the byte used to store those effects.
I don't have a good idea for this one as giving too much invisibility counter makes invisibility stop being useful. I think the current balance is good, except for the Air Elemental itself. (Albeit at very rare they get Mass Invisibility which is a lot more powerful, but at least the other units aren't 5 movement and noncorporeal flying.)
Maybe Air Elementals could be very rare instead but that would impact Sorcery a lot - Air Elementals are the primary way of not losing battles when you are playing Sorcery.

Psionic Blast - Not a game stopper, even less than warlocks. You lose some units but can still win the battle, and losing 3 halberdiers worth of units should be easy to afford. The magicians casting Confusion, Fireball or Black Sleep is a lot more damaging, the only difference here is you can't avoid this one by using high resistance/defense units. (You can still use Raise Dead to cancel the losses though)

Crack's Call - Merging also counters it. High end units have too much health to be killed by this spell, while low end units should be easily replaceable. Again, this is not a game stopper unless using heroes specifically and having no access to a Wraith Form or Merging item. (it also counters buffing strategies somewhat as buffed units are expensive and still die to the spell, but that's intended. Buffing definitely needs counters as it's a very powerful strategy.)

Wave of Despair - An army of full 9 units barely take any damage from this spell. I don't see the problem.

Spell Blast - I suggest reading the related posts on AI behavior. This spell is balanced by the AI having special rules on when and how they are allowed to use it, so you can get around that with some effort.

Drain Power - This is indeed problematic, losing 200 mana each turn at the uncommon stage of the game can easily mean having to play without mana at all. At the moment it's quite far in the AI research order so it shouldn't appear that early but on high difficulty the AI will reach it faster and it can be found in treasure too. I wouldn't mind replacing this with a new spell in the future if the problem appears too often.

Heal- the current beta has Arcane heal in it already, for heroes only. Summoned units are replaceable, they don't need it.

Lightning Bolt, Mystic Surge - yes, these are what Chaos is meant to use to have some sort of answer to high armor early units. Works as intended. (In general, chaos has the theme of "can't defend against it", so you need to rely on other ways to prevent losses : Regeneration or Raise Dead/Resurrection.)

Life - Agreed it's the strongest for human (albeit I'm really bad at using early nowadays) and weakest for AI (but Altar of Peace and Sanctify should help there as well as the uncommon Stream of Life). However we definitely do not want to make it play the exact same as other realms, so no direct damage, no strong summons, and absolutely no resist or die. (Yes, I know in Magic the Gathering, white was the color that had by far the best destruction spells of all types, but I rather not duplicate that here. Let Chaos and Death do the slaughter and destruction...)

Doom Bat - These are a rare monster, players shouldn't expect to fight one and take no losses. Keep heroes away from them, or use Raise Dead. They only deal 8 damage so you'd need to be hit at least 4 times for a strong high health hero to actually die btw. Doom is scary but it has that downside of never being able to get a "good roll" to deal more damage either.

Encounter zone balance - The value of the monsters can be changed freely, but we already went though this and set all the values as we needed them for perfect balance.  Earth Elementals in particular are worth 200 each, while sprites are 80, I think the elementals are fine (the game does need something easy to hunt to make exploration a rewarding experience and 200 isn't that much) but sprites might work better as 100 or 120?
Reply



Forum Jump: