Caster of Magic Release thread : latest version 6.06!

[ChongLi]

Hey, I have a question. If it's been answered already, I apologize. However, I can't find the information in the wiki nor in the CasterOfMagic.html file. Do Sorcery/Nature/Chaos Mastery harder to dispel effects stack with Rune Master? Does Rune Master dispel effects stack with Dispel/Disenchant/Disjunction True? Do any dispel multipliers stack?

[Seravy]

Hey, I have a question. If it's been answered already, I apologize. However, I can't find the information in the wiki nor in the CasterOfMagic.html file. Do Sorcery/Nature/Chaos Mastery harder to dispel effects stack with Rune Master? Does Rune Master dispel effects stack with Dispel/Disenchant/Disjunction True? Do any dispel multipliers stack?

Yes to all of the above, at least I think so, I haven't intentionally changed how these stack together.

[Seravy]

A question :
Defense is a pretty powerful stat and heroes with enough defense are pretty much invincible to any attacks in the game. What does everyone think, is the amount of defense available reasonable, or not? I'm not sure but I think it might be too much.
I could change Agility to grant only 6/9 DEF instead of the current 9/13 on max level, and I could reduce the DEF available on items from the max 6/8/6 to, let's see, 4/6/4 or something like that. 8 DEF plate mails are a bit crazy, and 6 DEF accessories might also be overkill.

[Drax]

Normal combat would take a lot of time to execute and isn't an option, even if it was, it would result in 25 minute long AI turns.

Yes I know but didn't realize how bad this auto combat worked until now. Yesterday had one AI hero(Barbarian) with a priest take a sorcery node with 5 Phantom beasts and a couple of Air elementals.... hero was captain or commander level.This is ridiculous. I also have the feeling individual figures act as a multiplier in this current way of simplified combat.... which leads to other stupid cases with ruins.

About simplified combat there was a way used by Warlords 3 I believe which basically lined up both armies and units were fought one after another... which didn't work too well but could be done in the context of MoM and work a lot better than what it is now. Performance wise should not be much worse than what it is now. Tell me if you think any change in this area is possible and I could lay out my plan to make it work well. no really need to bother discussing it if not possible.

---

Speaking of performance there is huge difference between 1.33 and 1.35. 1.35 is waaay slower in my experience. I did not have to wait much at all in 1.33 and I did have some 30-40+sec waitings on turns in 1.35. hope that might be a clue. I do use DosBox at 20000 cycles. There is some complexity weighting over it... some nasty "n".

---

I also noticed something odd yesterday. A demon lord casts a spell and also summons a demon in the same turn(actually every turn if the spell chosen doesn't use all the mana). Is that a new feature?

---

I also have a question about spell loot. In old MoM it basically gave random loot which is compatible with books... which had some issues of course as you could get some great spell from a minor location and win the game with it. Now there seems to be a limiting factor of the rarity of spell and value? of item. What I did get as a result however was very often pretty terrible - basically my cheap uncommon spell being currently almost done researching or the really crappy items often accompanied by plenty of mana or gold. Now crappy items I do accept as plausible from low level(or any) dungeons, but insisting of giving you the same spell(I did try a dozen of saves to test on one occasion) that I'm in 1-2 turns of researching is pretty bad. Dungeon doesn't know what I'm researching so it's not supposed to give me that as the lowest possible value spell there is in my realm.

So my question is: is it possible to implement level of spells?
example: warlords battlecry 3 has basically 10 spells in each realm. When you are level 1 in that realm you can cast the first spell. On level 2 first and second and so on. There is also chance involved but ignore that for now. What happens when you go past level 10 to level 11 is that your first spell is level 2 now and the next ones are still at level 1. and the level 2 spell is more powerful.

Which leads to my general idea of being able to have like level 0 of spells(may be limited to common+uncommon) in all realms and these could be cast at limited skill by wizards not having the books. Dungeons could be a source for any spell and you could not cast it if it's a rare or cast it at a penalty if it's a common. For example costing twice mana maybe. This will work well with the current idea of limiting spell rarity I think.

My logic is that as a wizard you might not be good in certain school, but it is wrong to assume you're absolutely incapable at it. Similar penalty could apply to dark wizards casting life spells and white wizards casting death spells which is now possible.

Think about it

---

Next question Excalipoor... while some of the items I've found are certainly "interesting" I have not found any good item in CoM yet. The way I used to play MoM before was that I created a bunch of really powerful items in itemmake and these were the great loot I could find maybe in my games. The rest I made mediocre and there was a huge number of them so it was not certain you'd get any of the good ones at all. The really crappy ones I just removed since they are for all intents and purposes useless. My logic was to have a chance to get some good effects from other realms through items. I kind of expected something similar in CoM as the remaining of the game fits this. Similar concept is in Warlock 2(which is heavily influenced by MoM) where you can't make items as good as the bunch of high level ones that are found.

Now what's the concept behind items design in CoM? I seem to be missing the point. They seem like jokes, which is ok as long as there is something else besides the jokes... which I have not found yet. I've only found a couple of ok ones. Nothing good at all or comparable to what you could create yourself.

[Seravy]

Yesterday had one AI hero(Barbarian) with a priest take a sorcery node with 5 Phantom beasts and a couple of Air elementals.... hero was captain or commander level.This is ridiculous.

Did you check if
-the hero had enough artifacts to actually win?
-there were no powerful invisible units stacked with him/her?
-the wizard had a high enough casting skill to actually pull off a victory on magic power alone?

Tell me if you think any change in this area is possible and I could lay out my plan to make it work well. no really need to bother discussing it if not possible.

Not really possible, although, well...I've done crazier things already so...throwing out the entire old process and writing my own is certainly an option. If I decide to go for it, I'll figure out how on my own, I have ideas.

Speaking of performance there is huge difference between 1.33 and 1.35. 1.35 is waaay slower in my experience. I did not have to wait much at all in 1.33 and I did have some 30-40+sec waitings on turns in 1.35. hope that might be a clue. I do use DosBox at 20000 cycles. There is some complexity weighting over it... some nasty "n".

The AI now considers sea to land attack options, which is a lot of extra movement to check. Either this or other similar additions, there have been a few. The new code to find the way the units move (which both the AI and human uses) is also slower, but it's better than having bugs in it that cause units to be unable to enter valid map cells.
I've managed to make the game run at an acceptable - in fact, faster than before - speed by setting core to dynamic and cycles to 100,000 without having graphics and sound problems. I guess the number depends on your actual CPU power though, I had problems at 200k but 100k was more than enough speed. Core on auto couldn't even support 100k properly.

A demon lord casts a spell and also summons a demon in the same turn(actually every turn if the spell chosen doesn't use all the mana). Is that a new feature?

Are you absolutely sure it isn't the summoned demon casting the spell (they can)? If yes then that's a bug, maybe Summon Demon doesn't set remaining movement to 0.
The ability does exist though, but I don't remember giving it to Demon Lords, it's exclusive for the Illusionist hero.

Dungeon doesn't know what I'm researching so it's not supposed to give me that as the lowest possible value spell there is in my realm.

It isn't doing that, it selects a random spell of the chosen rarity and realm and if is already researched, cycles through until it finds one available. You are having really bad luck.
What you describe does exists when trading with the CPU, the spell you are researching is rated lower, so the AI is more willing to trade it away, but this does not affect treasure.

So my question is: is it possible to implement level of spells?

I'm happy with the current system. If you have 1 book you can find commons, 2 allows uncommons, 3 rares and 4 very rares. You'll never get a spell of lower rarity than what the treasure is supposed to contain unless you already know all spells of that rarity in the rolled realm. However, Rare and Very Rare spells are what the name implies. The chance for the treasure to contain them isn't very high (actually it's 1/4 for each rarity but when the treasure isn't valuable enough to proceed with it, it gets rerolled so only stronger areas have a chance to contain them)

Now what's the concept behind items design in CoM? I seem to be missing the point. They seem like jokes, which is ok as long as there is something else besides the jokes... which I have not found yet. I've only found a couple of ok ones. Nothing good at all or comparable to what you could create yourself.

I believe you are finding way too low value treasure. CoM items are better than the original and the game puts more effort into giving you a predefined item as well as having a higher value cap on treasure so you can actually find stuff over 3k value which you couldn't in the original. On top, quite a few item powers cost less and the treasure value is boosted by a multiplier of, currently, 1.5x.
If you still only find bad items, either your expectations are too high or you are fighting too weak dungeons.
As a general guideline, you can expect ~1.5-2x the total summon cost of guarding monsters in item value but of course the treasure might be split up between gold/books/items/spells etc, no guarantee all of it goes into a single item.
If you are expecting 5000 mana items you have to kill like 3 sky drakes for that and get lucky to have all that treasure go towards one items.
What difficulty are you playing? Higher difficulty puts more monsters in places, which results in more treasure.

About Excalipoor, while it isn't an endgame item, it's pretty powerful for the first half of the game. Four chances to reduce attack power of enemy units to zero each combat it pretty awesome. Even more so on a hero that isn't a spellcaster and was only in the battle for the leadership ability. I used it in my current game and it was great up until a certain point.

[zhantul]

Speaking of performance there is huge difference between 1.33 and 1.35. 1.35 is waaay slower in my experience. I did not have to wait much at all in 1.33 and I did have some 30-40+sec waitings on turns in 1.35. hope that might be a clue. I do use DosBox at 20000 cycles.

There may be an easy solution to this problem. Using the front-end D-fend, you can choose values higher than 20.000 cycles if you input them manually. (Or directly in a dosbox game config file type: "cycles=fixed 60000"). I'm currently playing the new version at 60.000 cycles with 4 opponents and I haven't had an enemy turn lasting longer than 10 seconds. (not counting turns where I could see 50+ units moving). Nearly the same as in the unmodded game.

I've had no stability issues playing like this but if I go too high in cycles (150.000 for example) the game slows down a lot and stutters. The ideal amount may depend on your computer.

[Drax]

Did you check if
-the hero had enough artifacts to actually win?
-there were no powerful invisible units stacked with him/her?
-the wizard had a high enough casting skill to actually pull off a victory on magic power alone?

13 attack 6 def hero. and Hell hounds. Life/Chaos wizard does not have enough spellpower to handle a decent sorcery node like this. I'm thinking node bonuses are not applied to the defending units and somehow the 0def of Phantom beasts is abused for this to be possible.

Not really possible, although, well...I've done crazier things already so...throwing out the entire old process and writing my own is certainly an option. If I decide to go for it, I'll figure out how on my own, I have ideas.

I would propose lining up units of the two sides and making first ones fight. Ranged units could attack and go to the back of the list once. Last winning unit could attack next. This is the simplest logic of a fight queue that I think might work much better then the current. If you decide to look into this here is a link to a video showing this type of combat in Warlords3 https://youtu.be/ibtIE2D-tok?t=1207

I've managed to make the game run at an acceptable - in fact, faster than before - speed by setting core to dynamic and cycles to 100,000 without having graphics and sound problems. I guess the number depends on your actual CPU power though, I had problems at 200k but 100k was more than enough speed. Core on auto couldn't even support 100k properly.

Dynamic core with 100k cycles seems a good setting. Thanks!

Are you absolutely sure it isn't the summoned demon casting the spell (they can)?

Is a demon able to cast black prayer and wrack? If it is then it might have been.

It isn't doing that, it selects a random spell of the chosen rarity and realm and if is already researched, cycles through until it finds one available. You are having really bad luck.
What you describe does exists when trading with the CPU, the spell you are researching is rated lower, so the AI is more willing to trade it away, but this does not affect treasure.

Well the result of this logic in case you have been actually researching spells and you have at least 6-7 books in only one realm... is the spell you're likely researching right now if you have a rarity restriction like this. It's basically guaranteed to happen.

A possible solution: select not one but say 4 spell candidates. If you find 4 spell candidates select one of them at random and give it as loot. If you can't find as much then obviously they have been researched and then maybe give a small chance at current and +1 rarity or give the loot as mana. It is really bad loot to find spell that you will have researched next turn. it's about 0 value.

I believe you are finding way too low value treasure.

well the treasure is large as a whole. 3 items of 400ish value + gold and mana is quite often a total way above 3000... but the thing is when you find 6 crappy shields/armors and 8 crappy axes in a game you can use maybe 3-4 of those and only until you decide to craft as they are well bad. Maybe the diversification of treasure is way too much.

CoM items are better than the original and the game puts more effort into giving you a predefined item as well as having a higher value cap on treasure so you can actually find stuff over 3k value which you couldn't in the original. On top, quite a few item powers cost less and the treasure value is boosted by a multiplier of, currently, 1.5x.

Not sure why I'm not finding them then. Merchants did bring a lot of items in one of my famous games but nothing remarkable.

If you still only find bad items, either your expectations are too high or you are fighting too weak dungeons.

1500-2000 value can produce pretty good items. I've found very few worth keeping in that range so far.

As a general guideline, you can expect ~1.5-2x the total summon cost of guarding monsters in item value but of course the
treasure might be split up between gold/books/items/spells etc, no guarantee all of it goes into a single item.

Yeah I've got retort+book+3 bad items. Not much further you could find in one place as a total.

So I'm not complaining about total amount... just the way it is attributed doesn't make much sense to me. Finding 3 crappy(+4ish total stats) artifacts are basically more mana.

What difficulty are you playing? Higher difficulty puts more monsters in places, which results in more treasure.

Extreme. The amount of monsters is quite substantial at places.

About Excalipoor, while it isn't an endgame item, it's pretty powerful for the first half of the game. Four chances to reduce attack power of enemy units to zero each combat it pretty awesome. Even more so on a hero that isn't a spellcaster and was only in the battle for the leadership ability. I used it in my current game and it was great up until a certain point.

I don't mind the Excalipoor to exist at all it's ok. Don't particularly like shatter but that's not the point. The point was I found 10s of such weird and extremely conditional items and basically not much of balanced or good ones which you'd actually wish to make yourself. I'd never make a item like the ones I've been finding. That's why I was wondering what was the logic behind the design.

Just wanted to know what to expect so I can plan accordingly, which you basically answered. Thanks!

[Seravy]

Is a demon able to cast black prayer and wrack? If it is then it might have been.

Yes, they have 40mp if the summon creates the spellcasting type, there are two types.

and Hell hounds.

A hell hound is roughly as good as a phantom beast.
it's 12 health ,beast has 20.
It's 4 times 3 melee and 2 ranged = 12+8 = 20 attack, phantom beast has 18 I think.
It's 3 defense vs beast's 0 which gives the hound a +75% multiplier to health while the beast gets none.
My numbers might not be exact but it sounds reasonable for the hounds to win if there were enough of them.
Node bonus exists but the beast is single figure so it doesn't benefit much from it, it gains defense to even out the hound's, but the increase from 18 to 02 attack is not going to do much.
Air Elemental is the same deal, probably even weaker (less health and attack).
The result would be different in manual combat due to Illusion and defense on the beasts and elemental.

I would propose lining up units of the two sides and making first ones fight. Ranged units could attack and go to the back of the list once. Last winning unit could attack next. This is the simplest logic of a fight queue that I think might work much better then the current.

A queue is one thing I absolutely want to avoid. Units dying in order instead of all of them taking damage is both somewhat unrealistic (the defender will try to move to keep them alive) and bad for the AI as it'll suffer heavier losses when fighting itself or neutrals. It would kill all the heroes too.

Well the result of this logic in case you have been actually researching spells and you have at least 6-7 books in only one realm... is the spell you're likely researching right now if you have a rarity restriction like this. It's basically guaranteed to happen.

I would say it happens because you find like 10-15 spells in the game and most of the time both your research and your found spell will be low rarity. If you are researching a common and it's the last one, you'll always find that if you would find a common.

A possible solution: select not one but say 4 spell candidates.

I think you don't understand the process properly.

It's like this, let's assume we need a Sorcery Uncommon :

1. Roll a random number 1-10. (there are 10 spells of each rarity and realm in the game)
2. If the spell rolled is already known, add 1 and repeat this step until either a valid spell is found or all 10 numbers were checked. (loop around to 1 if 10 is reached first)
3. If we didn't find a spell in the previous step, increase rarity by 1 and try again from step 1.
4. If all rarities are already checked, try another realm the player knows.

"valid" refers to spells the player does not know yet and has enough spellbooks to receive.
I might be wrong and it might loop realms first, rarities next but the result is almost the same. If there is a spell to get, you'll get one.

I guess an option here would be to make spells being researched "invalid" but that isn't a very good idea.
I can imagine plenty of situations where I'd love to get the spell I'm researching because I need it as soon as possible. It might take another 30-40 turns to complete it otherwise.

I'd never make a item like the ones I've been finding.

If you were always finding items you want to make there would be no point having the spell to make them. That doesn't mean there are no items in there I would want to make, it's just not all of them. There even are a few you can't make yourself, like -3 spell save accessories, or magic immunity items. The overall ratio of expensive to cheap items is much higher than in the original.

You can always run itemmake.exe and see what is in there, or even change them if you desire like in the original game.
If you do so, be warned that the AI can and will use them against you and they are likely to find them first.

[Drax]

Thank you for your answers!

A queue is one thing I absolutely want to avoid. Units dying in order instead of all of them taking damage is both somewhat unrealistic (the defender will try to move to keep them alive) and bad for the AI as it'll suffer heavier losses when fighting itself or neutrals. It would kill all the heroes too.

It's always possible to provide retreat action to the queue. Not sure current system does anything better in the fair department. Idea is to basically reduce 2D combat to 1D in order to be more fair/realistic for the automated fights. Also AI has bonuses to producing so losing more units is actually already compensated...

[Seravy]

Thank you for your answers!

A queue is one thing I absolutely want to avoid. Units dying in order instead of all of them taking damage is both somewhat unrealistic (the defender will try to move to keep them alive) and bad for the AI as it'll suffer heavier losses when fighting itself or neutrals. It would kill all the heroes too.

It's always possible to provide retreat action to the queue. Not sure current system does anything better in the fair department. Idea is to basically reduce 2D combat to 1D in order to be more fair/realistic for the automated fights. Also AI has bonuses to producing so losing more units is actually already compensated...

The player does not, and they can also use automatic combat if desired. I believe a battle is supposed to create a mixture of wounded, dead and undamaged units, not only dead and undamaged. The original game used such a queue and it wasn't good, weak enemies could kill your best hero in the stack if it got selected to be first in the damage queue.