Talk:Breeding

Congenital traits inheritance table

I failed to find any evidence for the table at Breeding#Congenital traits, either in the game files or in developer posts/replies. Is someone else able to verify this information? WeirdMatter (talk) 21:43, 6 October 2020 (UTC)

In the absence of replies, I have removed the table. WeirdMatter (talk) 00:03, 18 October 2020 (UTC)

Pure-blooded effect on inbreeding

NOTE: There were inaccuracies in this table, mainly due to oversight (inactive traits were not recorded) and other errors in the code regarding trait indexes. The original table has been redacted, with a newer, accurate one posted as a comment.

To figure out how exactly the pure-blooded trait affected inbreeding, especially when both parents have the trait, I decided to run some tests.

The game version used is 1.1.3. For each test, I modified the game's history files to edit the characters to fit the requirements. The characters in question are Heinrich IV of the Holy Roman Empire and his full sister Jutta (1066 start date). Neither character has inheritable traits. For the control test, I removed both of Jutta's parents, making her a non-relative of Heinrich.

I used a modified form of the "Start Pregnancy" interaction, with the num_of_children set to 1,000. At game start, without unpausing, I used the interaction on Jutta as Heinrich. I then used speed 5 until the pregnancy was finished (22 June 1067). On the same day, without unpausing, I saved the game. Then, I used a Python script to gather relevant data for all 1,000 children. If requested, I can share this script for those who wish to see it.

I performed the above test four times:

1. Control: non-relatives, no pure blooded trait
2. 0x pure blooded: full siblings, no pure blooded trait
3. 1x pure blooded: full siblings, 1 pure blooded trait between the parents (I personally used Jutta, I doubt that it would change if Heinrich had it instead)
4. 2x pure blooded: full siblings, both parents have the pure blooded trait

Here are the results. The table shows how many times each trait showed up in each test. Through brief observation, I noticed that some children had multiple negative traits. Note that in all tests, all children had the "twin" trait.

(table removed due to inaccuracy; accurate table posted below)

Observations made from the data:

1. Anomalous prevalence of the Albino trait. In all groups, it was the highest naturally-occurring trait. What is most interesting is the second test, where it occurred in approximately 1 in 5 births, and is ~42% more common than the second most-common trait (Barren/Sterile). Explanations for this may be some form of alphabetical priority, or Albino has a larger weight for selection as an inbreeding trait. See below for up-to-date observation.
2. The absence of the following traits: Fecund, Wheezing, Spindly, Bleeder, Possessed, Lunatic and Melancholic. A larger sample size may be required to rule them out as not naturally occurring via birth. Note that in preliminary tests, I observed 1 child with the Bleeder trait (sample size = 24, both parents with pure-blooded). See below for up-to-date observation.
3. The markedly increased prevalence of these traits with inbreeding: Albino, Barren/Sterile, Clubfooted, Giant, Scaly, Impotent, Hunchbacked, Dwarf. It can be surely concluded that these traits may be picked as an inbreeding trait. See below for up-to-date observation.
4. The lack of the Inbred trait. This is supported by developer comments in the game file regarding the enables_inbred value, which suggests that the Inbred trait may only be gained if a character has such a trait (e.g. Dwarf, Scaly). More investigation is needed on this point.
5. The effect of having one pure-blooded parent. Someone with a background in statistics may be able to offer a more technical explanation, but I can see that there is an approximate 50% reduction in inbreeding traits.
6. The powerful effect of having two pure-blooded parents. The amount of traits compares with the control group, suggesting that the two −50% Inbreeding chance modifiers stack to −100%. Note that this only impacts additional traits from inbreeding, and not naturally-occurring negative traits.

After performing these tests, I noticed that the characters' father (Heinrich III) had the Stuttering trait. However, given the lack of significant difference between the control group and the other tests, I don't believe that this has had a massive effect on the final observation. WeirdMatter (talk) 23:33, 25 October 2020 (UTC)

I have edited the above due to inaccuracies caused by the data gathering script. Additionally, this table was auto-generated rather than filled by hand, so any other minor changes may be expected. Only the script has changed, not the source. Here is the new, accurate table:

Trait	Frequency (#)
	Control	0×	1×	2×
Giant	3	106	57	1
Slow	4	6	3	1
Handsome / Pretty	1	5	4	0
Melancholic (inactive)	1	6	3	0
Genius	0	2	0	0
Wheezing	6	93	62	1
Frail	1	1	0	0
Stuttering	3	7	7	1
Hideous	4	2	0	0
Scaly	1	101	66	4
Pure-blooded	0	7	161	770
Dwarf	5	95	53	3
Intelligent	1	3	1	1
Homely	1	2	2	1
Albino	7	6	5	3
Comely	4	4	6	3
Hale	3	4	7	2
Delicate	6	4	4	3
Robust	2	3	1	1
Lisping	4	5	5	1
Feeble	3	1	0	1
Hunchbacked	2	99	57	3
Barren / Sterile	6	128	70	3
Imbecile	0	2	0	1
Ugly	2	2	2	0
Bleeder	3	98	57	1
Stupid	8	2	1	0
Quick	4	6	5	1
Possessed (inactive)	6	2	5	2
Beautiful	1	3	1	0
Spindly	4	100	58	0
Lunatic (inactive)	7	5	5	0
Divine	1	0	1	1
Club-footed	4	115	63	3

Updated observations:

1. There is no one trait that is prevalent in all groups. The findings appear to line up with expectations, based on the natural prevalence (birth) of traits from the game files.
2. The absence of the Fecund trait. More testing is required; however, one hypothesis is that it does not generate among naturally-born kids since there is no birth value assigned to it in the game files. The Fecund trait does not appear randomly at birth (barring legacies).
3. The dramatic increase of the following traits with inbreeding: Barren / Sterile, Scaly, Club-footed, Wheezing, Spindly, Giant, Hunchbacked, Bleeder and Dwarf. Correlates with the enables_inbred property. A causative link can be established with further testing. There appears to be no real weighting behind the trait selected (i.e., no trait is preferred over the over), apart from potentially Barren / Sterile (more investigation required).
4. No change.
5. No change.
6. No change.
7. (New observation) The traits Lunatic, Melancholic and Possessed are given to children as inactive traits. WeirdMatter (talk) 16:12, 27 October 2020 (UTC)

Natural prevalence of Fecund

Using a similar method to before, I tested whether the Fecund trait naturally appears for newborn children. Parents are unrelated, sample size is 30,000, version is 1.1.3. Note that there were no legacies unlocked, including the level 4 blood legacy that gives the chance for Fecund to appear. The expected value is based on the birth value of the trait, and does not account for groups, order, etc.

Trait	Frequency (#)		Prevalence (%, 2 d.p.)
	Expected	Observed	Expected	Observed
Giant	150	137	0.5	0.46
Slow	150	163	0.5	0.54
Handsome / Pretty	75	76	0.25	0.25
Melancholic (inactive)	150	145	0.5	0.48
Genius	15	9	0.05	0.03
Wheezing	150	160	0.5	0.53
Frail	75	76	0.25	0.25
Stuttering	150	153	0.5	0.51
Hideous	45	40	0.15	0.13
Scaly	150	153	0.5	0.51
Dwarf	150	127	0.5	0.42
Intelligent	75	73	0.25	0.24
Herculean / Amazonian	45	54	0.15	0.18
Homely	150	158	0.5	0.53
Albino	150	133	0.5	0.44
Comely	150	129	0.5	0.43
Hale	150	159	0.5	0.53
Hunchbacked	150	142	0.5	0.47
Robust	75	67	0.25	0.22
Lisping	150	161	0.5	0.54
Feeble	45	42	0.15	0.14
Delicate	150	144	0.5	0.48
Barren / Sterile	150	133	0.5	0.44
Imbecile	15	10	0.05	0.03
Ugly	75	76	0.25	0.25
Bleeder	150	136	0.5	0.45
Stupid	75	75	0.25	0.25
Quick	150	142	0.5	0.47
Possessed (inactive)	150	138	0.5	0.46
Beautiful	45	35	0.15	0.12
Spindly	150	143	0.5	0.48
Lunatic (inactive)	150	144	0.5	0.48
Club-footed	150	148	0.5	0.49
Fecund	?	0	?	0

As seen, all other congenital traits that are expected to appear do so. Barring the level 4 blood legacy, Fecund is not randomly given to newborn children. WeirdMatter (talk) 23:30, 27 October 2020 (UTC)

Inbreeding Chance Level

00_defines.txt only says that numbers of common ancestors dividing levels are 15 17 20 22, so it is unclear how levels are divided, 0~15 / 16~17 / 18~20 / 21~22 / 23~30 or 0~14 / 15~16 / 17~19 / 20~21 / 22~30.

To find out which one is right, I changed the Inbreeding Chance Factor to 0/7/7/7/7, which can provide a chance over 100% to 2nd level but 0% to 1st level. I tested to Marzoban Wahsudan of Gilan in 867 with his half-sibling sister Zerya, because their faith have Unrestricted Marriage. They have a lot of missing ancestors(there are only 3 ancestors, which are father, grandfather, great-grandfather), and it is also uncertain how to count the number with missing ancestors, but they are half-siblings, so their ideal number is 15.

First test was with 15 17 20 22. Their baby had no bad trait. Second test was with 3 17 20 22. Their baby had a bad trait. Third test was with 14 17 20 22. Their baby had a bad trait.

So their number is 15 and Inbreeding Chance Levels are divided to 0~15 / 16~17 / 18~20 / 21~22 / 23~30, or their number is 14 and levels are divided to 0~14 / 15~16 / 17~19 / 20~21 / 22~30, "And" parents with missing ancestors have a special rule which reduce the number.

I tested another couple who are parent-child, and got the same results.