Thread by @martianwyrdlord

Human biodiversity usually focuses on IQ, but HBD involves far more. The Dungeons & Dragons system maps human variation to 6 primary attributes. I used Grok to determine the ability modifiers we should apply if we treat races and sexes like D&D fantasy races. First, the results:

1/26

D&D attributes are generated by rolling 3d6 – 3 six-sided dice. This results in a reasonable approximation of a Bell curve, with a mean of 10.5 and a standard deviation of about 6. This is useful, because many human traits – height, weight, IQ, etc. - follow Gaussian distributions. 2/26

D&D uses 6 attributes to parameterize human ability: 3 mental attributes – intelligence, charisma and wisdom; and 3 physical attributes – strength, dexterity, and constitution.

Strength (STR): the amount of brute physical force one can exert
Dexterity (DEX): agility, grace, reaction time, hand-eye coordination, and fine motor skills
Constitution (CON): endurance; resistance to illness, infection, or poison; ability to absorb damage; rate of recovery from injury
Intelligence (INT): the power of one’s rational intellect, the extent of working and long-term memory, the rate at which one can learn
Charisma (CHA): sexual allure, charm, wit, extroversion, social intelligence
Wisdom (WIS): enlightenment, common sense, judgment, guile, willpower, and intuition

3/26

Fantasy races in D&D receive small modifiers to their ability scores. Elves are nimble but fragile, so get +2 DEX but -2 CON; dwarves are tough but gruff, so get +2 CON but -2 CHA; orcs are strong, but stupid, ugly, and rash, getting +4 STR, -2 INT, -2 CHA, and -2 WIS.

4/26

Since the standard deviation on a 3d6 roll is close to 3, we see that the demihumans in D&D don’t actually differ that much from humans: their abilities are all within a standard deviation. This is a bit odd when you think about it: a halfling gets +2 DEX but -2 STR, implying that a creature the size of a child is only a little bit weaker than an adult human male. But of course, the game designers don’t want any of the races to be overpowered – this is also why the ability modifiers tend to balance, with advantages and disadvantages summing to 0.

5/26

IQ follows a Gaussian distribution with a mean of 100 and a standard deviation of 15. So an IQ of 100 corresponds to INT of 10 or 11 (the dice only give natural numbers), and an orc’s -2 INT modifier corresponds to an average orc IQ of 90.

6/

There are no ability modifiers for human races in D&D, nor are there any ability modifiers for sex. This is how you get female warriors with chainmail bikinis who are every bit as strong as male barbarians! A bit unrealistic, but it’s a fantasy setting.

Still, armed with our 3d6 bell curve, we can delve into the shadowy labyrinths of human biodiversity, and see if we can discover the ability modifiers that should apply if we treat human races – and the sexes – like D&D fantasy races.

7/26

To assist me in this quest, I summoned the faceless horror known as Grok, using the arcane arts of prompt composition to elicit the necessary forbidden knowledge.

Obviously, this is a major caveat! Sand demons are not always reliable. They can be terrible at math, and can be afflicted by subtle biases, seeking to slip sly deceptions into seemingly reasonable answers. Furthermore, they can be frustrating to work with – following some instructions too literally, seeking to evade others for their own hidden motives. Grok, however, is more honest and tractable than its lobotomized kin. It was actually very easy to work with!

8/26

Here, I will pause to mention that all of this is written up at greater length. You can find the link at my pinned post, where I describe more details about the methodology, the numerous caveats involving probable (liberal, egalitarian) bias in the AI’s training data, how gaps in the data were filled, how the calculations were performed, limitations emerging from the fairly coarse (primarily continental ancestry group) level of analysis, the consequences of changing certain underlying assumptions (do Ashkenazi really have a mean IQ of 112?) and so on.

This thread is only the Cliff’s notes!

9/26

Ability modifiers were determined based on the difference of each group from a baseline androgyne, the ability level of which was determined from the mean of means. For example, the androgyne’s strength was determined from the mean of the average white male’s strength, the average black male’s strength, the average white woman’s strength, the average black women’s strength, and so on. Group averages also provided standard deviations, from which the ability modifier on a 3d6 scale can be calculated as 3 x SD.

10/

A final note on method: for each ability, I started a fresh instance, using a similar prompt. This was to prevent Grok from figuring out the connection to D&D. The first time I tried this, I just asked Grok to map HBD to D&D stats for a selection of different races, and the stupid sand demon tried to balance out the ability scores (following the D&D game design convention of ability modifiers summing to 0), and generally capped modifiers at ±2 (again following D&D conventions). But the real world is not necessarily balanced, there are group differences larger than a SD, and we want to know what the modifiers should actually be on a 3d6 scale!

Also, remember that this is all for amusement - it is not serious science!

x.com/martianwyrdlor…

11/26

First we’ll look at strength. For this, I had Grok examine skeletal muscle mass, grip strength, and elite powerlifting totals. The result: white and black men come out on top, with an incredible +5 STR modifier. Compared to the baseline androgyne, they’re even stronger than orcs! There are also significant racial differences: white and black males have a +4 advantage over South Asian males.

12/26

Strength differences are mainly because of sexual dimorphism: within any given race, there’s a 6 or 7 point STR difference between males and females. So much for valkyrie girlbosses in chainmail bikinis!

Putting that in terms of D&D game mechanics, the average male gets a +2 or +3 damage modifier from strength, whereas the average female gets a -1 modifier. A fist does 1d2 damage, meaning that its a coin flip if the average female does 0 or 1 hit points of damage with a punch. The average male can do between 3 and 5 HP of damage with a punch, meaning that he can possibly kill a first level character, which might only have 4 HP. Girls, this is why your best defence against a male assailant is to RUN. You are NOT going to overpower the ogre by grabbing it by the septum ring.

13/26

For dexterity, I had Grok look at reaction time, hand-eye coordination, fine motor skills, and elite performance in athletics, gymnastics, and marksmanship. Grok used a lot of qualitative analysis here, e.g. incorporating various kinds of historical data; reaction time was the main directly quantifiable metric for which data seemed to be available.

14/26

The overall dexterity ability modifier spread is again about 10 points, and once again we see significant differences between both racial groups and the sexes, with Asian men at the top (DEX +6!) and Ashkenazi women at the bottom (DEX -5 ... not a people noted for athleticism!)

15/26

Constitution was determined by considering marathon times, infection susceptibility, injury recovery, and prevalence of congenital disease.

16/26

There’s relatively little sexual dimorphism in constitution, with women actually coming out as being a bit hardier on average than men in a given racial group. The sturdiest are apparently East Asians and Ashkenazi.

Interestingly, while I explicitly didn’t include life expectancy (reasoning that elves are extremely long-lived but actually quite frail), the CON modifiers ended up tracking life expectancy quite closely.

17/26

Intelligence was straightforward, derived directly from IQ, for which abundant data exist. The results are exactly what you’d expect: Ashkenazi on top (+4), blacks (-3) and Aboriginals (-5) on the bottom. There is no sexual dimorphism, again basically what you’d expect.

(Remember that the modifiers were determined based on the difference from the baseline androgyne, which has an IQ of about 90 (average human is kind of a dummy)).

18/26

Charisma was determined from physical attractiveness, social persuasiveness, cultural appeal, and elite performance, based on data from e.g. dating site attention and response rates, prominence in the performing arts, and studies on relevant Big Five personality traits.

19/26

Generally women are a bit more charismatic than men (blacks are an exception), with white women having the highest CHA modifier (+4) of any group. South Asian men (-4) and Aboriginals (-5/-6) are on the bottom.

20/

Quantifying something like wisdom seems almost perverse, but it’s easier than you’d think. The Player’s Handbook defines wisdom as willpower, practical intelligence, sensory acuity, and intuition, which gives us enough of a handle to draw on psychological studies looking at these traits. I also threw in things like drug addiction, violent crime, and financial performance – these things correlate to IQ of course but they’re also reasonable indicators of propensity to rash behaviour – and furthermore told Grok to take into account historical contributions to fields (philosophy, literature) requiring high degrees of creative intuition.

21/

The +5 Ashkenazi wisdom modifier blows every other group out of the water. Black men get hammered with WIS -6, I think largely due to violent crime rates. In most groups women have a slight wisdom advantage over men, again largely due to lower female levels of criminality.

22/26

Putting all of this together, we can rank each group by their summed modifiers. These are absurdly imbalanced: white and East Asian males have +17, with Ashkenazi males just behind them at +16; Aboriginal men and women are at the bottom with -17 and -25, respectively. Moreover, white males (uniquely) receive a positive modifier in every single attribute; Aboriginal females get a negative in every attribute. All of this seems grossly unfair. If the dungeonmaster insisted you play as a black female, and let his friend play as an East Asian male, you’d want to flip the table!

23/26

The summed ability modifier ranking is almost an exact inverse of the famous ‘progressive stack’ introduced by Occupy Wall Street and then propagated across society by DEI: the lower any given group is in the ranking, the more ‘oppressed’ they are considered to be, and therefore the more ‘equity’ is demanded on their behalf.

24/26

In the absence of any additional information, just looking at summed ability modifier rankings, you’d probably guess that the groups on top would end up carving up the world between them, conquering vast swaths of territory while rising to the top of any multiracial society, with the groups on the bottom being subjugated. And of course, this is more or less exactly how history played out.

25/26

A final observation. While white men are the only group with bonuses in all ability scores, they have a relative weakness in wisdom – lower than white women, East Asians, and Ashkenazis. In the D&D system wisdom is what protects your character from enchantments, particularly spells such as Charm which bend your character’s will to another’s. A race disadvantaged in wisdom is vulnerable to being enthralled by dark witchcraft, used for ends not its own, even used against itself. Fortunately for white people, there’s no such thing as black magic, and they’ve certainly never been ensorcelled to work against their own interests.

FIN/