Mysteries of trade

[Arnuz]

So, in this save...

1. Tauron accepts web for any of 4 spells. I just received construct catapult for web with Merlin, so I would understand that web has less value than CC, but Tauron does not accept CC for any although he doesn´t have it (check with the gifts)
2. Raven accepts only aether binding for blur, a shitty common, but more spells (even an arcane) for philosopher´s stone a quite good uncommon for AIs
3. In the meanwhile Jafar accepts the same arcane for blur

Is this only trade fatigue - I guess I traded more with Raven than Jafar - or do personalities and spellbooks matter? If spells are not on the AI research list they value them more?

Some things make no sense though. It´s almost worth it to start with blur just for the trade value.

Upload is giving some trouble, let me know if you need the file.

[jhsidi]

Spells all have a unique trade value, for some it's very high compared to others at the same tier. Web is extremely high value for a Common because of its effectiveness (it's not too much to say it can be the keystone of a game winning strategy). I'd tend to agree Blur isn't an amazing spell, mainly because it's just situational to mass combat... and kind of boring to use... -- but since it is extremely powerful for its niche uses, it's similarly given a high value.

Starting with a specific spell just to trade it is definitely a viable strategy! Although that idea is counter-balanced by things like early utility, must-have spells, research cost, etc.

[Arnuz]

Spells all have a unique trade value, for some it's very high compared to others at the same tier. Web is extremely high value for a Common because of its effectiveness (it's not too much to say it can be the keystone of a game winning strategy). I'd tend to agree Blur isn't an amazing spell, mainly because it's just situational to mass combat... and kind of boring to use... -- but since it is extremely powerful for its niche uses, it's similarly given a high value.

Starting with a specific spell just to trade it is definitely a viable strategy! Although that idea is counter-balanced by things like early utility, must-have spells, research cost, etc.

Thank you for your answer, but I knew that. The surprising bit in the web case is that after getting it for construct catapult, construct catapult is not accepted while web is. That can´t be explained by the unique trade value, as far as I can see...

[Seravy]

The trade values are in Spells.INI.
The modifiers should be in Modding.INI, there is one for spells the AI already sees on their research list, and another for spells the AI cannot research at all.

So assuming otherwise equal value, a spell the AI can already pick for researching is worth the lowest, a spell they can research later but not yet has the default value, and a spell the AI won't have in the research at all has the highest.

Then, a list of spells is shows where the spell received by the player has equal or lower trade value than the spell the AI is getting.
However there are two more modifiers :
When the player asks for the trade, if there are at least 6 players in the game, the trade value of the player's spell is decreased by 1 for each 2 players in the game beyond 4.
When the AI asks for the trade, the trade value of the player's spell is increased or decreased depending on the relation between the player and the AI, and a D3 random roll. With high relation and a good roll you can get up to 6 higher spell value than the spell offered.

In your case, this implies the AI does not have Web available for research but does have Construct Catapult.
Unless, they only have 1 Nature book in which case they can't learn uncommon spells.

[Arnuz]

That helps, thanks. Does a spell being on the human´s list affect the choices?
Is the list of players the current amount or the game setup´s?

[Seravy]

"Does a spell being on the human´s list affect the choices?"

Yes, it modifies the value of the spell received by the player the same way as the AI's list modifies the value for the spell the AI is getting.

"Is the list of players the current amount or the game setup´s? "
The game setup's.

[Arnuz]

Ah, that explains some differencies that I noted. Hmm, that´s a bit unfair no? The AI shouldn´t know what´s on my spell list... Yeah that´s kind of funny in a game that lets AIs get away without scouting but still, it makes the diplomacy part a tiny bit more convoluted without need.
Would you be open to consider removing that distinction? That can be counterbalanced, if needed, by raising the values to a level in between the two (known and unknown)

[Suppanut]

You could get away with scouting your own too. Unresearched spells in book has systemic encryption

[Seravy]

This is the consequence of asymmetric diplomacy in general.
The player gets to make the choice for both the received and given spell. The only way that can result in fair trade is when the AI gets to evaluate the value of both spells to determine which trades to accept or refuse.

However as an ex TCG player, I have to say this information usually pretty obvious, if we assume the player knows which spells they can research. Usually when trading cards in real life, you knew whether the other person wanted a card very much or just traded for it because "might as well get one of these" and it makes sense to ask for more in the former case. So in CoM that means the AI simply notices you don't really care about the trade when a spell is already in your spellbook and you look pretty desperate for the trade when it is not.

(and yes, spells are intended to be known for the player, that's why you can read your spellbook in advance. It's pretty trivial to figure out which spell is which based on category if you know which spells are in your realm(s), especially as spells are also sorted by rarity. Like if you see a unit spell that has 3 characters in the name and it's on the top, and you play all Nature books, it's obviously Web.)

[Arnuz]

It´s not about knowing, it´s about not having some options that you would otherwise have, since the higher value for spells not on your book means that you won´t ever get them offered despite the AI not being able to access it - even if the AI is offering the trade, if I get it correctly. After all when the AI proposes, the player only chooses among a few random options what to give, not what to get, so at least in that case it´s not that asymmetric.

Just as an example Sssra just contacted me to propose a pointless trade of 2 spells that I can research in 2 turns. That´s a bit wasting the player´s time. He should not know that I can only research those.

Besides to properly maximise this you should take note of all that is offered for the whole game by the AIs, to discover their spellbooks, but that´s just not fun... And even doing that the player surely can´t tell what´s on the AIs books at the first contact until there´s been quite some research and diplomacy. Maybe you could have a spell that provides that intel and modifies the weights accordingly once it´s casted? But until then, knowable and unknowable spells weights should really be equal imho. (ed: or it could be an effect of detect magic)