Leader memories

[RenaissanceFan]

I think Casimir, Yaroslav, Philip the Bold and other long-lasting leaders should forget both bonuses and penalties more. It''s really strange for them to remeber about a refusal some 500-600 years ago. (Sobieski appears only in 1600-1700, the other two have no replacements)

Also, when a leader switches the relationship could be partially reset.

 

[3Miro]

I think Casimir, Yaroslav, Philip the Bold and other long-lasting leaders forget both bonuses and penalties more. It''s really strange for them to remeber about a refusal some 500-600 years ago. (Sobieski appears only in 1600-1700, the other two have no replacements)

Also, when a leader switches the relationship could be partially reset.

If memory clears too fast, then the this leaves too much room for the player to abuse. I can be mean to them as much as I want and a few turns later they will not remember. I am far from the illusion that the leaderhead parameters are optimized, however, I do think that reducing memory would be bad for gameplay.

 

[Caliom]

I think it was beta 12 when the memory decays where almost halfed compared to normal RFC or vanilla BTS. And some events where added that the Ai never had forgotten before.

Chaning the memories when a new leader apperars is an interesting and flavourfull idea. Which memories would you pick? Would you only make them forget about some memories or also add some random new ones?

Is it completly random, or are there certain rules, or does every leader have some predefined "historical" memories with every other leader? The last one could replace the hidden modifier mechanics, making it more transparent and giving it the possibility to change over time.
If it is completly random it could severly hurt some of the "have x open borders" UHVs.

 

[AbsintheRed]

If memory clears too fast, then the this leaves too much room for the player to abuse. I can be mean to them as much as I want and a few turns later they will not remember. I am far from the illusion that the leaderhead parameters are optimized, however, I do think that reducing memory would be bad for gameplay.

Indeed. We have to balance the memory decays
Caliom already started this, but some of the parameters are still not optimized

 

[AbsintheRed]

Chaning the memories when a new leader apperars is an interesting and flavourfull idea. Which memories would you pick? Would you only make them forget about some memories or also add some random new ones?

I don't think this is a good idea
What if you declared war at a civ, razed cities, etc.
And then make peace the turn before the new leader appears.
Then the relations with the civs would change from -15 to +1 in one turn?
Even if it's true that quite big attitude changes very possible historically on leader changes, something like this would be very bad for gameplay IMO

 

[Opera]

Although that may be warranted in case of a respawn?

 

[Caliom]

I don't think this is a good idea
What if you declared war at a civ, razed cities, etc.
And then make peace the turn before the new leader appears.
Then the relations with the civs would change from -15 to +1 in one turn?
Even if it's true that quite big attitude changes very possible historically on leader changes, something like this would be very bad for gameplay IMO

I didn't mean that you loose all your memories. But a "you razed some of our cities" effect could be changed from -5 to -3. Jumps from -15 to 1 should be avoided.

I'm not sure how the memories are saved when leaders change. Probably the attitudes already change, because different leaders can value the same event differently. For example Isabella could reward sharing a war with +2 while Philipp gives only +1.

 

[AbsintheRed]

I'm not sure how the memories are saved when leaders change. Probably the attitudes already change, because different leaders can value the same event differently. For example Isabella could reward sharing a war with +2 while Philipp gives only +1.

Yeah, I'm not sure either how the switches happen
For example what's the case if Leader1 has memory decay of 50 turns for an event, while Leader2 has 30, and the event happens 40 turns before the switch
Would the event be forgotten on the switch, or 10 turns after it?

 

[MariusMR]

I don't, personally, see the individual leaders as anything significant, but rather view the relations as the long term relationship with that nation, i.e. the Polish people and the German people.

To me it is immaterial who the leader is and I don't think, for example, that the Cordobans and the Spanish nations would ever 'like' one another. The same with the Polish and Germans.

 

[Caliom]

Yeah, I'm not sure either how the switches happen
For example what's the case if Leader1 has memory decay of 50 turns for an event, while Leader2 has 30, and the event happens 40 turns before the switch
Would the event be forgotten on the switch, or 10 turns after it?

a memory decay of 50 does not mean that you forget the event exactly after 50 turns. Its a parameter to some kind of distribution and i'm not sure which one, but i would guess that the parameter its the expected value of that distribution. That would mean that the forgeting is expected to happen somewhere around turn 50.

In your case the new leader simply had a higher chance to forget the event every turn. As far as i know for every event every turn a check is made if its forgotten.

 

[Opera]

It's a 1 on X chance to forget. Only the ContactDelays are in turns.

 

[AbsintheRed]

It's a 1 on X chance to forget. Only the ContactDelays are in turns.

Are you sure?
From what I experienced decay values seems to be closer to an expected value, like Caliom said

 

[Opera]

Yes I'm sure. Check here if you want. They mention it's a "binomial distribution" test if that means anything to you (I certainly didn't know that).

Edit: Maybe you didn't understand my previous post though, it may be not too clear. I meant that each turn, a random number between 0 and X is generated. When it is 0, the leader forgets 1 point (so the Modiki says) of the diplomacy modifier being checked.

 

[AbsintheRed]

I don't, personally, see the individual leaders as anything significant, but rather view the relations as the long term relationship with that nation, i.e. the Polish people and the German people.

To me it is immaterial who the leader is and I don't think, for example, that the Cordobans and the Spanish nations would ever 'like' one another. The same with the Polish and Germans.

Yeah, having different attributes for the leaders is just fine tuning, don't worry
We don't want to have huge jumps in the relations when a civ switches leader
Also, there are hidden modifiers between spain-cordoba, poland-germany, and all the similar civs which historically hated or liked each other

 

[AbsintheRed]

Yes I'm sure. Check here if you want. They mention it's a "binomial distribution" test if that means anything to you (I certainly didn't know that).

Edit: Maybe you didn't understand my previous post though, it may be not too clear. I meant that each turn, a random number between 0 and X is generated. When it is 0, the leader forgets 1 point (so the Modiki says) of the diplomacy modifier being checked.

Actually I study applied mathematics
And yeah, your last post wasn't very clear, I thought you want to say something entirely different

Btw, both Caliom and I were speaking about expected values, we just wasn't sure which distribution they base it on
This means it's simply binomial
Adding x in a memory decay means that on the long run the average turns in which the AI will forget a point is (x+1) turns

 

[Opera]

Sorry, I totally missed my own point. What I said seemed obvious in my head but not at all when I read it again

 

[AbsintheRed]

What I said seemed obvious in my head but not at all when I read it again

Don't worry, this happens to me too all the time

 

[AbsintheRed]

Speaking of leader memories, I'm in the process of adding a few more unused attitude changers and memory decays to the leaders:
Accepted join war, hired trade embargo, cancel open borders, received tech in trade, and a few others

 

[AbsintheRed]

Speaking of leader memories, I'm in the process of adding a few more unused attitude changers and memory decays to the leaders:
Accepted join war, hired trade embargo, cancel open borders, received tech in trade, and a few others

Not totally tied to leader memories, but I was already setting the leader's diplo attitudes, and this is also connected to the diplo relations:
What do you guys think about adding the Show Hidden Attitude modcomp?

 

[3Miro]

Not totally tied to leader memories, but I was already setting the leader's diplo attitudes, and this is also connected to the diplo relations:
What do you guys think about adding the Show Hidden Attitude modcomp?

C++ seems straight forward. It is just some read functions and modification to the text function. If others want this, I will merge it into the .dll.