That's actually true. Actually even better would be to start at say 6/8/8, and then scale by 1 for every extra level. Which would solve the power generation problem of heroism, but not the casting skill (although I could live with that since AI can use it effectively too and its degibitely less of a problem than the power generation.)
Noble would probably also be more balanced if it was something like +6/level. I didn't change it because it would be bad for flavor - being higher level won't make you a higher rank noble or anything like that.
They scale with level because the constant amount gets obsolete around midgame - Noble is the worst hero ability you can get on turn 200 and one of the best for turn 10.
(and yes, Heroism breaks this unfortunately, but I rather have that problem than 4 hero abilities the player will not want to see after turn 100.)
Quote:Actually even better would be to start at say 6/8/8, and then scale by 1 for every extra level.
This could work, unfortunately it's quite hard to implement. I coded the entire hero ability display procedure (all 2 -or was it 3- copies of them) with the assumption that every ability is either constant or linear.
I think even the current amount is pretty bad for the late game though. +36 power or 72 SP on a max level hero doesn't sound bad but at that time you have like 1200 power base so it's marginal - any combat ability is more valuable than these. The interesting tradeoff here is that on small land where you have more resources, these abilities suddenly become worthwhile or outright good as this amount is pretty good for Tiny land with weak nodes, but for the majority of games, they are horrible for the late game. I almost always dismiss my Ritual Master hero to make room in the late game if I don't have a slot and need to summon a champion or incarnation.
A trivial solution would be to add "the hero loses Sage, Ritual Master and AEther Master" to heroism but I don't like the idea of a buff permanently weakening a unit in Life.
Right. Although I disagree that the power generation of multiple cities (more if you don't get wizards towers) is not useful late game.
However, I'm firmly on the side of 'never use heroes in combat because they auto win'. So the only thing I use heroes for is power generation and overland casting skill.
on a Conjunction event I have some sort of multiplier applied to my SP pool.
The SP shown on the screenshot seems to be the actual value used as I do get the increase of the skill on the next turn. I've attached a save game also.
It seems the multiplier is only applied to the SP pool and is about 4x. I do have a hero with Aether Master +40 but no other SP bonuses.
Big request: Treasure spells are chosen from spells not in your spellbook (for later research). It's so disappointing to find a spell that you'd get later anyway. There are exceptions, but 90% of the time I'm unhappy to get one that's in my spellbook.
(March 5th, 2017, 05:08)Drax Wrote: Found a bug with 3.24
on a Conjunction event I have some sort of multiplier applied to my SP pool.
The SP shown on the screenshot seems to be the actual value used as I do get the increase of the skill on the next turn. I've attached a save game also.
It seems the multiplier is only applied to the SP pool and is about 4x. I do have a hero with Aether Master +40 but no other SP bonuses.
Not a bug. Right click on the Sorcery Conjuction and read the description.
Quote:Big request: Treasure spells are chosen from spells not in your spellbook (for later research). It's so disappointing to find a spell that you'd get later anyway. There are exceptions, but 90% of the time I'm unhappy to get one that's in my spellbook.
For that I would need to rewrite the entire spell treasure generator.
However if we did it, I could finally make sure the player always gets the highest rarity spell possible.
As is, realm is rolled first - if the player has 1 Chaos 1 Nature and 4 Life books and finds an Uncommon spell, they'll only have 1/3 chance to actually get an Uncommon, and 2/3 to get a common in the 1 book realms.
If this is done, spell treasure will most likely need to be significantly less frequent. Even as is, you can research all your spells first and do noding afterwards to only get new spells, and it's usually enough to get every spell even with 8-9 books. But the price - a lot of nodes are already gone when you can start. If you always get a new spell, you don't need to wait and can probably max out your spells even with as few as 5-6 books unless we reduce the drop rate of spells.
One more problem - the bumping of rarity if no valid spell is found. We shouldn't use this mechanism on unresearched spells, otherwise 7+ book wizards are guaranteed to always find only rare+ spells at the beginning of the game - they can research every single common and uncommon spell.
Furthermore, as everyone can always research Arcane spells, it means those will not appear at all.
I think this change would be far too drastic, although it would certainly benefit the AI. (no more "7 book wizard didn't get the good very rares" scenario)
I'm not sure the idea is worth it (spell rewards on in spellbook) - it lessens the relative power of selecting more books to guarantee important spells. (I'll get retort x and y and maybe 2 books in this other realm, because I may get missing spells from trading and also rewards)
My preferred algorithm would be like this:
1) chose a random spell of any realm that the player has at least one book in or arcane.
2) give a chance to learn it like: (0.15+0.10*(books in realm)+0.025(books in other realms))*4/rarity
3) if player can't learn it or already has it give mana reward instead.
The current model is basically useless. Meaning you very rarely have any use of it. It gives you spells that you most likely have and does so very infrequently. In my last 3 runs I only received one spell that ever mattered(High Prayer though) all others were either ones I had in my books and low value ones.
