• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

binTravkin

Annoying Latvian
29 Badges
Aug 18, 2004
3.243
19
  • Europa Universalis III Complete
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Death or Dishonor
  • Hearts of Iron IV: Together for Victory
  • Crusader Kings II: Reapers Due
  • Hearts of Iron IV: Cadet
  • Crusader Kings II: Horse Lords
  • Crusader Kings II: Way of Life
  • 500k Club
  • Victoria 2: A House Divided
  • Victoria 2
  • Semper Fi
  • Europa Universalis IV: Res Publica
  • Europa Universalis III Complete
  • Crusader Kings II
  • Heir to the Throne
  • Hearts of Iron III: Their Finest Hour
  • Hearts of Iron III
  • For the Motherland
  • Europa Universalis IV: Call to arms event
  • Europa Universalis IV: Wealth of Nations
  • Europa Universalis IV: Conquest of Paradise
  • Europa Universalis IV
  • Divine Wind
  • Europa Universalis III Complete
  • Crusader Kings II: The Old Gods
  • Crusader Kings II: Legacy of Rome
  • Crusader Kings II: Charlemagne
As I said
- The bad people will die out in the end as AI does not seek for bride for them so tediously, thus resulting in a 'natural selection', which as a general rule is faster if there are more people.
If the natural order of things were there in CK it wouldn't be a problem, otherwise you can only put some makeshift fix for it, like AI marrying their worst people first (which would be entirely stupid) or generating only bad courtiers (bye bye cool courtiers) or another even less fun thing than the very fast natural selection that we have.
And then, as I said earlier, the natural order of things wouldn't be fun either, so basically, we should be happy with things as they are.

EDIT: correction - we should be happy with things as they are except for the balance algorithm which only creates more disbalance (by resonating the difference and creating 'loops').
 

unmerged(6777)

Field Marshal
Dec 10, 2001
12.470
5
binTravkin said:
EDIT: correction - we should be happy with things as they are except for the balance algorithm which only creates more disbalance (by resonating the difference and creating 'loops').
That's fair. I'm not so convinced that the algorithm is all that out of whack, but if the concensus is that there is an excessive disparity somewhere then I am prepared to revisit it. I do not think Johan will export the variables to a master file, though, so that's a pure code tweak and can't be done very often (or Johan will get testy at continually revising his formula).

Is there a concensus that there's a disparity or is the approximate ratio of males to females usually about right?
 

Veldmaarschalk

Cool Cat
151 Badges
Apr 20, 2003
30.115
1.840
  • 200k Club
  • 500k Club
I personally have never expirienced any trouble with the male/female ratio, or with their stats.

So I don't really see a need for tweaking, but I am also not involved in breeding programs within my games.

EDIT
But if other people feel that it needs tweaking, go ahead. :)
 

binTravkin

Annoying Latvian
29 Badges
Aug 18, 2004
3.243
19
  • Europa Universalis III Complete
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Death or Dishonor
  • Hearts of Iron IV: Together for Victory
  • Crusader Kings II: Reapers Due
  • Hearts of Iron IV: Cadet
  • Crusader Kings II: Horse Lords
  • Crusader Kings II: Way of Life
  • 500k Club
  • Victoria 2: A House Divided
  • Victoria 2
  • Semper Fi
  • Europa Universalis IV: Res Publica
  • Europa Universalis III Complete
  • Crusader Kings II
  • Heir to the Throne
  • Hearts of Iron III: Their Finest Hour
  • Hearts of Iron III
  • For the Motherland
  • Europa Universalis IV: Call to arms event
  • Europa Universalis IV: Wealth of Nations
  • Europa Universalis IV: Conquest of Paradise
  • Europa Universalis IV
  • Divine Wind
  • Europa Universalis III Complete
  • Crusader Kings II: The Old Gods
  • Crusader Kings II: Legacy of Rome
  • Crusader Kings II: Charlemagne
It's actually not a big problem unless you go to extremes killing many (hundreds of) people of the same gender as I and Woz the Early mentioned or if you get 14 daughters and 1 son (which happens time by time, but still rare enough).
I mean, there are bigger issues than this for you to fix.

But once, you're at it, it would be very nice if there was a way to script the education popups NOT to appear and instead use a special moddable script which takes some values (by default - all the things that affect outcome of education, but moddable so one could say mod in condition to give higher priority on bishops if there's no diocese bishop) into account and decides what will the education be..:)

Or make techadvances suspendable to the log (instead of popping and pausing the game)..
 

unmerged(2456)

Pure Evil Genius
Mar 29, 2001
11.211
0
www.hero6.com
binTravkin said:
No that isn't.
People are not generated by some preset 1/1/1/1 or other pattern.
People are generated with percentage likelinesses, e.g.:

1/1/1/1/1 = 0.1%
2/2/2/2/2 = 5%
3/3/3/3/3 = 10%
4/4/4/4/4 = 18%
5/5/5/5/5 = 25%
6/6/6/6/6 = 18%
7/7/7/7/7 = 10%
8/8/8/8/8 = 5%[

OR a very similar scheme, were there is slightly better possibility of average people generated than overly bad or overly good.
In such a scheme if you generate 10 people, there is 10 times less chance to get someone really good than if you generate 100 people.

Not really. There is not a tendancy upward. If anything your FE forgetting 0/0/0/0/0/0 stat person which would in that instance shift things down somewhat.

But even so, they are weighted to be in the middle and if your on the low end, getting a low end stat is more common because it still computes negaitve values, but afterowrs rounds them to 0.

The reason why stats continue to climb is because people breed good courtiers with good courtiers, not continous breeding of bad ones.
binTravkin said:
Furthermore consider following:
- overall quality of people (education related) is improved by good advisors. If you have 10 people and a 7 diplomacy 9 stewarship silver tongue steward, it gives your children slightly less chance to become midas toucheds than if you have 100 people and you can find a 7+ diplomacy 11 stewardship midas touched steward. It's not just in 0.95 * 0.95 (two bonus modifiers), but it's in the relative percentage of midas touched chance versus all other level and type (diplomacy, intrigue) chances.
Education is not a signifigant modifier. Only kings are likely to have all best education and uber stat people for everyone until we see tons of bloated courts, and then some dukes. Even so its not a major modifier.

Education and other traits are not figured into the forumula, except for bastards, for determining children (unless that was recently changed).
binTravkin said:
- AI tends to search 'best for best', e.g. it will moslty marry some good character to his good advisor resulting in more good people due to parent education bonus and randomised combination of the good parent stats.
Exactly, but that's all the more reason to limit the ai if it already starts searching best-for-best.
binTravkin said:
- The bad people will die out in the end as AI does not seek for bride for them so tediously, thus resulting in a 'natural selection', which as a general rule is faster if there are more people.
Not really. Once they run out of their best people, they go on to their medicore and the poor. Only the truly horrid, or clerics, the ai doesn't try to marry off (unless they get too old as well).
binTravkin said:
Obviously you don't have much grip on the reality.
Are you pretending the current child birth rates, people death rates, ages and other things about health and recreation are realistic?
If they were, we'd be soooo owercrowded in our 21st century..
No. I wasn't pretending. These are not the same because these are represnting a minority class, nobiles, not the majority, peasants. I did not ask for them being realitic in their numbers, just to not continue to grow expontnetially every generation, even during the plagues.
 

binTravkin

Annoying Latvian
29 Badges
Aug 18, 2004
3.243
19
  • Europa Universalis III Complete
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Expansion Pass
  • Hearts of Iron IV: Death or Dishonor
  • Hearts of Iron IV: Together for Victory
  • Crusader Kings II: Reapers Due
  • Hearts of Iron IV: Cadet
  • Crusader Kings II: Horse Lords
  • Crusader Kings II: Way of Life
  • 500k Club
  • Victoria 2: A House Divided
  • Victoria 2
  • Semper Fi
  • Europa Universalis IV: Res Publica
  • Europa Universalis III Complete
  • Crusader Kings II
  • Heir to the Throne
  • Hearts of Iron III: Their Finest Hour
  • Hearts of Iron III
  • For the Motherland
  • Europa Universalis IV: Call to arms event
  • Europa Universalis IV: Wealth of Nations
  • Europa Universalis IV: Conquest of Paradise
  • Europa Universalis IV
  • Divine Wind
  • Europa Universalis III Complete
  • Crusader Kings II: The Old Gods
  • Crusader Kings II: Legacy of Rome
  • Crusader Kings II: Charlemagne
Not really. There is not a tendancy upward. If anything your FE forgetting 0/0/0/0/0/0 stat person which would in that instance shift things down somewhat.
Have you played up until 1453?
In your games there is no tendency upward?
Well in mines there is slight and I guess Im not the only one.
I am forgetting that person because of very simple and logic reason you agreed yourself to:
That person would never get married if such is ever created, while a 10/10/10/10 will almost always get married (the only cases being fertility killing illnesses or residing in a court of religious enemy).
For that reason the mentioned low level person will not change anything in overall picture.

But even so, they are weighted to be in the middle and if your on the low end, getting a low end stat is more common because it still computes negaitve values, but afterowrs rounds them to 0.
Where do you get this assumption from?
If it is true there should be plenty of people with at least one stat being zero (as rounded up from negatives), but there are almost none.
To take this as serious argument I would require MrT or someone else who knows something about the generator, to approve it, it seems plain unbelievable to me that the generator gives any negatives, not even talking about the irrationality of such concept from the programmatic aspect.

The reason why stats continue to climb is because people breed good courtiers with good courtiers, not continous breeding of bad ones.
Well, this contradicts your previous statement
There is not a tendancy upward.
So I wonder what the point was, but I agree (that AI breeds good with good and is slightly less willing to marry the bad ones, which seems pretty logical and smart).

Education is not a signifigant modifier. Only kings are likely to have all best education and uber stat people for everyone until we see tons of bloated courts, and then some dukes. Even so its not a major modifier.
It is.
Mother midas touched + father midas touched + steward midas touched (the mother herself) with at least 7 diplo and 10+ stewardship = 0.9 * 0.9 * 0.9 * 0.9 = 0.6561*Midas Touch MTTH (!!!)
Now, this is only absolute value.
To get the real impression you need to compare it to other type and level changes to get the proper picture:
Code:
	BASE MTTH	S
---------------------------------
L1	35		44.375625
L2	30		28.5
L3	40		29.322225
L4	50		32.805
---------------------------------
We see that the MTTH L2 (which is the trap for most educations) versus L4 ratio changes by approx 50%.
The column basically says that if the basic chance to get an L4 education is a bit less than 20%*, then with proper advisors, it's 25%** and it's only the advisorial bonus.

More calculations show that buildings add up to 10% for the L4 bonus and traits and stats add another 10% (this is variable as noone can predict how many positive or negative traits a character can get, but generally 1-3 positive or negative is the average).
Those are all the factors:
- advisors and parents 5%
- buildings 10%
- character stats 10% +/- 3%

So, I can agree that traits and buildings are a major factor, yet advisorial and parental bonuses don't seem much minor too with half the influence and being one of only 3 factors.

And there's a slight tendency for AI to use 'proper' persons for the positions - it will use a intrigue 10 theologician for spymaster if no flamboyant schemer+ with the same or better intrigue is around, but once there is, it sometimes even puts someone with a bit less ability (advisoral skill) than someone other because he/she has the proper education for position.
In big courts AI will always find 'proper' people thus aiding himself.

Only kings have big courts?
Do I need to ship you same saves?
It is of course less likely that counts have courts as big as king ones, but the difference is really insignificant and only becomes less as the game progresses, as there are all kinds of migrations, especially if someone conquers lands of his religious enemy - the border areas are usually filled with refugees who eventually convert to the proper faith and start to breed.
Another factor, not of little significance - kings give titles away to their childs, so do dukes, but counts rarely have anything to give, thus all those bred relatives have nowhere to go.

Education and other traits are not figured into the forumula, except for bastards, for determining children (unless that was recently changed).
No they aren't and I didn't say so, BUT they are taken into the formula which calculates the education of those children.

One small example - in my current Byzantine game A.D. 1019
I am not in position to attack the humble Georgia while having 12 king titles and 50+dukes as wassals myself.
The Bagratunis are known for their high stewardship, but now the king is a Midas touched monster with 18 stewardship and his steward is his wife with 13 stewardship.
2 of his bigger provinces can beat all my vassals in Anatolia and have bigger income than Byzantion.
This is a challenging AI.

Exactly, but that's all the more reason to limit the ai if it already starts searching best-for-best.
We don't need limiting the search best-for-best. It's natural, but we could limit the reproduction rates a bit more and add to the death rates.

Not really. Once they run out of their best people, they go on to their medicore and the poor. Only the truly horrid, or clerics, the ai doesn't try to marry off (unless they get too old as well).
Yes, once.
Means the good people are married first.
Means the skill pool goes up as first married and first to reproduce are those above average (and the truly good included whereas the truly horrid are almost never married).
Means in the end the average level of people will be better than in the beginning.
Natural selection. The more people, the faster.

No. I wasn't pretending. These are not the same because these are represnting a minority class, nobiles, not the majority, peasants. I did not ask for them being realitic in their numbers, just to not continue to grow expontnetially every generation, even during the plagues.
Even for nobility they are unrealistic.

I'd say the best way to do what you want is to:
- reduce birth rates by some 10-20%
- raise death rates by approx the same
- add a major ilness boost (x2 - x4) in sick provinces (there are numerous sicknesses appearing before Bubonic plague, yet I've never seen my courtiers getting sick with them yet)

I guess the heightened labour death rate was implemented in the latest betas. IMO it's not of any help as the brides are found anew and the show goes on.
 

unmerged(2456)

Pure Evil Genius
Mar 29, 2001
11.211
0
www.hero6.com
binTravkin said:
Where do you get this assumption from?
If it is true there should be plenty of people with at least one stat being zero (as rounded up from negatives), but there are almost none.
To take this as serious argument I would require MrT or someone else who knows something about the generator, to approve it, it seems plain unbelievable to me that the generator gives any negatives, not even talking about the irrationality of such concept from the programmatic aspect.
Well i have routinely seen many courtiers spawn, as children, with 0 on one of the main stats. Also i have seen them even below both of the parents modifed and unmofied stats. This is in-game evidence. Try playing a bit differently and you might actually see these.
binTravkin said:
Well, this contradicts your previous statement
No. I'm saying that even with a means average, 2 5/5/5/5 courtiers won't produce overall as good as two 10/10/10/10 courtiers.
binTravkin said:
It is.
Mother midas touched + father midas touched + steward midas touched (the mother herself) with at least 7 diplo and 10+ stewardship = 0.9 * 0.9 * 0.9 * 0.9 = 0.6561*Midas Touch MTTH (!!!)
Now, this is only absolute value.
To get the real impression you need to compare it to other type and level changes to get the proper picture:
Code:
	BASE MTTH	S
---------------------------------
L1	35		44.375625
L2	30		28.5
L3	40		29.322225
L4	50		32.805
---------------------------------
We see that the MTTH L2 (which is the trap for most educations) versus L4 ratio changes by approx 50%.
The column basically says that if the basic chance to get an L4 education is a bit less than 20%*, then with proper advisors, it's 25%** and it's only the advisorial bonus.

More calculations show that buildings add up to 10% for the L4 bonus and traits and stats add another 10% (this is variable as noone can predict how many positive or negative traits a character can get, but generally 1-3 positive or negative is the average).
Those are all the factors:
- advisors and parents 5%
- buildings 10%
- character stats 10% +/- 3%

So, I can agree that traits and buildings are a major factor, yet advisorial and parental bonuses don't seem much minor too with half the influence and being one of only 3 factors.

And there's a slight tendency for AI to use 'proper' persons for the positions - it will use a intrigue 10 theologician for spymaster if no flamboyant schemer+ with the same or better intrigue is around, but once there is, it sometimes even puts someone with a bit less ability (advisoral skill) than someone other because he/she has the proper education for position.
In big courts AI will always find 'proper' people thus aiding himself.
Which should in practicality and game-balance wise only be fore kings and powerful dukes. A backwater count having enough courtiers to fill such a large court should be the exception, not the rule it is now a couple generations in. Its why the respawn rates for counts, dukes and kings are 3/5/7.
binTravkin said:
Only kings have big courts?
Do I need to ship you same saves?
Ok. Yes, eventually almost everyone will have big courts, unless they are brand new. That is unhistorical. It should be as i said above, that only kings and extremely powerful dukes have big courts, but because the ai goes and marries almost anyone it can, that isn't the case and rank no longer means much for courtsize that it should.
binTravkin said:
One small example - in my current Byzantine game A.D. 1019
I am not in position to attack the humble Georgia while having 12 king titles and 50+dukes as wassals myself.
The Bagratunis are known for their high stewardship, but now the king is a Midas touched monster with 18 stewardship and his steward is his wife with 13 stewardship.
2 of his bigger provinces can beat all my vassals in Anatolia and have bigger income than Byzantion.
This is a challenging AI.
The ai could do that without marrying anyone not tied down as well. And that's not really much different than the ai being able to stay at 0 gold as well. But unlike the latter, the AI now that is has ways of seeking out best brides similar to humans with bridefinder programs, they could do 95%+ as good, perhaps moreso since the AI wouldn't be constantly spending its time looking for brides for medicore people once its best are married off.
binTravkin said:
We don't need limiting the search best-for-best. It's natural, but we could limit the reproduction rates a bit more and add to the death rates.
There are already a lot of events. Adding more death events to there would be like robbing peter to pay paul. I agree that deathrates for children are too low though, even on VH. As for the limit to repdocution, that might help, but doesn't go to the root cause, the ai marries off almost anyone. If that were taken care of, it would signifigantly curb the population and reduce cpu load more than both of the other two.
binTravkin said:
Yes, once.
Means the good people are married first.
Means the skill pool goes up as first married and first to reproduce are those above average (and the truly good included whereas the truly horrid are almost never married).
Means in the end the average level of people will be better than in the beginning.
Natural selection. The more people, the faster.
Only to an extent. The ai doesn't base who to marry at all and who not to on relative stats, but absolute. Even when most everyone has 15/15/15/15, they'll still marry a 7/7/7/7 if they run out of good people.
binTravkin said:
Even for nobility they are unrealistic.

I'd say the best way to do what you want is to:
- reduce birth rates by some 10-20%
- raise death rates by approx the same
- add a major ilness boost (x2 - x4) in sick provinces (there are numerous sicknesses appearing before Bubonic plague, yet I've never seen my courtiers getting sick with them yet)

I guess the heightened labour death rate was implemented in the latest betas. IMO it's not of any help as the brides are found anew and the show goes on.
And stop the ai from marrying everyone under the sun. Atleast some restiction on who it can marry. That would curb that more that all those listed combined.

As for sickness, those weren't meant to represent deadly sicknesses mostly, but just moderate illnesses beyond the sniffles.
 

unmerged(47151)

Colonel
Aug 4, 2005
1.019
0
You mean go to another court? Those are almost all bloated as well. a handful might find an opening, but 95% will continue to take up valuable cpu resources.

Only if such a court is available, basicly surplus courtiers should go to rulers of the same culture first and then rulers of the same religion and then rulers of a different denomination (Catholic vs Orthadox) and only then to rulers that are of different religion.

If their are no available courts to go to that haven't already been filled, then the courtier should go into alternative employment, that is another job to what they are doing at the moment, maybe the church, army, civil service, espionage corps, trading, etc according to their education and abilities.
 
May 31, 2004
532
0
Erm...returning dangerously to topic, if Johan has time to give the formula a tweak then I for one would appreciate it.

I agree with Mr. T up to a point - the formula isn't wildly out, it does work fairly well most of the time, but as soon as its out of kilter once then it can take decades to sort itself out. If it ever does at all.

Somehow it seems to respond really REALLY badly to what it considers to be 'extreme' variations. Some kind of cap to prevent the birth imbalance automatically going higher than, say, 1:3 would be nice. That should still be more than enough to help correct the imbalance within a generation (maybe two), yet hopefully prevent extreme numbers of men or women from being born all at once.

Plus, that'd be (AFAIK) a relatively simple and painless code change for Johan. Far better than trying to rethink the entire formula.