Spell Very Rare Treasure

[Nelphine]

Right. But you think summoninga Gorgon in 1403 isn't just as bad as Armageddon 1407? I think juffos point is that if we're going to ban early very rares, how is getting gorgons in 1402 (allowing summoning by 1403) reasonable?

[Seravy]

AI just used Great Unsummoning in 1412. Obviously I wasn't expecting that from a wizard who doesn't even have rares yet (just to confirm I wasn't mistaken I looked at the spell data and they only had Warp Node as researched rare, were researching Spell Blast which is somewhere at the end of the uncommons in research priority) - so I had Doom Mastery in play. So half my army gone and if that's not bad enough, they have enough casting skill to use like 4 possessions per battle due to being Myrran. I might be able to not lose the game but I don't care. I'm not supposed to have to fight against this sort of Myrran invasion backed by very rare spells on this level of difficulty (expert) this early (1412).
Several of the very rares have been designed to hold back other wizards and maintain your advantage, and those things showing up at early rare tier breaks the game completely.

I'm abandoning the game and will delay very rare spells and tower opening immediately, then start a new game tomorrow. I was considering to make the tower delay optional but I won't - if someone turns it off and then get steamrolled by this, they'll blame the game, not their own choice to disable it.

The fact that towers have by far the highest likelihood of very rares further escalates the problem by giving these spells to the AI - as a human you don't want to open Myrror when you're not ready to fight it. As towers guarantee a spell and have enough budget to get the very rare in 90% of the cases, at least 2 very rare spells per game are expected to come from the 9 towers.

[Nelphine]

Again, this is why I want a difficulty modifier onmodified on very rare spells. For very rare, 1410 is standard for very rares. For experience 1414 is. So make it 1410 +2 years per difficulty.

[Seravy]

Uh, no, our calculations for expected turn for spell tiers was difficulty independent. (Yes, if you play better you get more but the AI also plays "better" on higher difficulty so I'd expect the overall speed of expansion to be the same)
I suppose the AI does get them earlier with the research bonus but not the human.
Question is, do we want to say "Sure, the AI will research a very rare this quickly so might as well let everyone find more anyway" or say "the human doesn't have them researched yet so let's not give out any for free either".

As the original problem was "we don't want the AI let very rares before they could research any", both are compatible with this goal. However, considering you have been complaining about getting too many very rares and low book counts being far too superior due to that, I'm leaning towards the latter, not allowing the very rares anyway - which makes sure low book counts do have the disadvantage of not getting enough very rares.

Also, difficulty is only one modifier, map size and power setting are others and probably even more relevant. There is no point adding difficulty without the others, which would overcomplicate it a lot. I rather have it as a fixed turn.

(turn 200 might be too much though, it's going to be an experimental number. If needed we can reduce it to 190 or 180 but not any lower.)

[Nelphine]

Right, in lunatic, very rares are researched by 1412. If not, then either you've lost, or the AI has lost.

[Seravy]

Reviving this because I now believe it might be worth doing the same for rare spells - obviously with a much lower turn count. The replacement would be an uncommon spell and a high value item - 1600 budget? The spell "only" costs 1200 though so maybe that's excessive? Although compared to how good early rare spells are, I guess it's okay?

There are much fewer problem rare spells and the scale is nowhere near as bad as very rares, mostly because casting skill generally prohibits using them THAT early, but still, even one Flame Strike per battle is something I rather not see the AI use on me in 1403. Likewise I wouldn't want even one Wraith in their doomstack. Rampaging monsters don't get to have rare units that early either. That said, unlikely very rares, with rares the problem is more the fact a human player can find these as early as the phase when no one has uncommon spells yet. In theory the AI might get lucky and do the same but I don't think I have seen that happen yet, at least not with a spell that was a game changer. While rares have much less raw power, some of them still win the game, like Magic Immunity, as Sapher is demonstrating in his current game.

However, at the moment, this is just an idea, and not decided yet. So let me know your opinion.

Edited below:

As towers no longer have very rare spell treasure in them, at most rares, I'm also more willing to scale the turn count down by difficulty on those. It does make more sense to find them a bit faster on higher difficulties. Probably by 12 turns for Master and 24 for Lunatic or something like that. We also have no reason to keep the base amount at exactly turn 200, although as far as I remember that was the cutoff point between the rare and very rare phase in our economy calculations? So the main question is how much earlier than the expected time for those to get researched, should they show up in battle? A year? Two? Probably no more than 3...
The thread does say turn 200 was experimental. I'd say it worked reasonably well, found very rares still are a big deal when they happen most of the time, and they almost never happen "too early". One exception was Hadriex finding Time Stop and winning but that was the problem with time Stop being too good. It shouldn't be nearly that bad now. Maybe the difficulty scaling can start one difficulty lower, at Expert? I generally do research my first very rares in the 1413-1418 time interval depending on how well I'm doing, but I do play Expert difficulty and test strategies that are average nowadays.

Reading the thread, the idea of making found spells add the spell as researchable to your book (most likely without guaranteeing it's the next due to space and timing issues) is now doable. The current treasure generating function places a predefined quantity of each type of treasure instead of randomly generating each lair individually, so redesigning costs of these no longer really matters. However, finding a spell in treasure that you won't actually research until 100-150 turns later is pretty disappointing, so this is probably a bad direction to take?

[Sapher]

let me know your opinion.

I have already said that I don't like any restrictions by the turncount
AI could try to protect his lairs by turning hostile if my units enter his territory (like during wizards pact but also without it). May be also during alliance? It might need a few exceptions thought like protecting a node and may be a tower. It might be a bad idea, i am not even sure myself

[Seravy]

Players have consistently demonstrated they are greedy enough to get into wars for a node or lair, or even a small patch of medicore territory for a settler. Also, cities have overlapping tiles so the owner of territory isn't even well defined. It wouldn't work. In fact, considering the potential loot is too powerful for the early game, and often good enough to win those wars, it doesn't really do much except making the player win even faster.

If the problem was the player getting all the lairs and not leaving any for the AI, the solution would be to improve the AI's ability to take lairs. I see two main ways to do that :
1. Make the AI cheat against neutral like the original game used to.
2. Change the AI combat so that specifically against neutrals, losses are maximalized by focusing damage on units one at a time instead of diluting it.
The first is plain unfair to "normal" players who clear out lairs at a normal speed. The second is risky - It does help the AI kill the node by multiple weaker attacks, but if the node had regenerating units, it doesn't work and it'll be an endless pit of doom that keeps destroying AI stacks without the AI ever clearing it.
Unfortunately, even if we successfully implement either, it won't fix the problem, it will only escalate it. Instead of the human getting too powerful spells too early, now the AI gets them. The AI casting rare spells in the early game when most players have only commons will be a frequent occurence. That's even worse for the game than the human player getting the same spells too early - instead of winning the game, they'll lose.

Edit :
We have 2 separate things to decide :

1. Are rare spells too powerful in the early game? Or is the high casting cost enough to make them fair? (Before 1406-1409 depending on difficulty?)
If the answer is yes, we need a way to prevent anyone from finding any too early. I don't see a good solution other than turn count for doing this, why the other ways don't work is explained in the other thread.
If the answer is no, we don't need to do anything with this.

2. Is the AI too slow at clearing their lairs? Do we want the AI to get a larger percentage of lairs?
If the answer is yes, we need to make them faster. Probably the only safe way to do so is to give them a difficulty based modifier that applies to lair hunting, for example all levels of difficulty above Advanced reduce lair strategic strength by 15%.
if the answer is no, we don't need to do anything here either.

[Sapher]

I am still not convinced that making AI defend his lairs is bad. If you are playing a treasure hunting strategy you really dont want to fight AI. All those draconian bowmen  and heroes that only can run away and shoot are very vulnerable to a simple fire bolt. If entering AI territory is bad there is an option to count attacking a lair in AI's territory of influence as an act of war.

[Seravy]

Okay, that can be question #3.

3. Should we implement a new diplomacy penalty for clearing a lair or node in someone else's territory?

Personally, I think it's partially redundant with the Wizard's Pact (without one you'll usually get attacked anyway, and if you have one, you shouldn't be going into their territory), makes minimal sense for flavor (ok so I kill these huge dragons that threaten the citizens in your country so now all your people can live safely, why are you angry again? Sure taking treasure is part of it but not having to deal with rampaging monsters from the lair is a benefit for them, and a big one.) and will make players angry (the AI will take their lairs anyway so it's highly assymmetric, and unlike wizards' pact, this doesn't give any benefit for the player in exchange. It's a choice between starting war, or watching as the AI clears your lairs as well as their own. Neither is really a good option.)