HGH in high-level sports

Years ago when Jamaican sprinters started to dominate the 100m/200m dashes, it was whispered that they had unusual jaw growth, some of them needing braces even in their early twenties. The allegation was that they used HGH, human growth hormone, that besides improving performance had side effects on the growth of extremities.

These side effect can occur naturally, usually if there is a tumor in the pituitary gland and very high amounts of growth hormone are created. This leads to a condition called Acromegaly, where jaw and brow ridge and basically everything else grows to almost grotesk proportions. André the giant may be most famous example, or maybe Richard Kiel playing Jaws in the Bond movies.

At one of the recent Olympics, maybe 2012 or maybe 2016, I was struck by how similar the two superstars Micheal Phelps and Usain Bolt looked in terms of their body shape. Long, lean with a strong jaw and big hands and feet (Usain Bolt has shoe size 13, Phelps size 14).

I began wondering whether HGH-abuse, probably starting already in teenage years, was playing a major role in shifting the borders of human performance in high level sports. In swimming the never-ending flood of world records was explained by the improvements of full body swimming suits. Of course when these suits were banned records kept falling.

HGH leads to detectable changes in the facial structure. And Deep Learning methods allow us to turn pictures of faces into vectors that encode facial structures. This gives us a way to empirically assess whether faces of world class athletes have been shifted towards the facial structure typical for people suffering Acromegaly.

These people suffer from Acromegaly

Because this is half-assed science and not full-assed science and there is, as always, a severe lack of graduate slaves, we will only manage a proof of concept. For this we select the male Olympic finalists in swimming for the years 1976, 1992 and 2016. These three years fall into three different phases of HGH-abuse: HGH has been used in high-level sport since 1982 and it was possible to detect it’s abuse since the early 2000s. So 1976 is pre-HGH, 1992 is HGH-time with no risk of being caught and 2016 is HGH-time with the theoretical possibility of being caught.

We also select the a couple of people suffering Acromegaly, as given by the wiki-article on the subject. And as a control group a number of normal guys, by googling ‚random guy‘ and ‚normal guy‘. We use a model that creates face embeddings, that is it detects faces in a picture and assigns vectors to these faces that encode facial structure.

We then compare the average face vector for our Olympic finalists and normal guys with the average Acromegaly face vector. Our results show that the distance to the facial structure typical for Acromegaly was biggest in 1976 with 0.671, smallest in 1992 with 0.625 and a little bigger than 1992 in 2016 with 0.634. The normal guys have on average a face 0.658 away from the Acromegaly face. The standard deviation of the distance of normal guys is 0.058, so the difference between 1976 and 1992 is 0.78 standard deviations.

This is certainly a big difference and the pattern of differences between 1976, 1992 and 2016 fit the different phases of HGH-abuse very well. The only thing that should give us pause is the fact that my normal guys are closer to the Acromegaly face than the 1976 athletes. More work is needed, but probably not by me!

Advertisements

Random Thoughts – Morality

I
It confuses me that vegans state animal welfare as a reason for foregoing meat, despite the fact that the animals vegans claim to care for largely wouldn’t even exist if they weren’t eaten by humans. Even granting that non-existence is better than some versions of factory farming, clearly the most moral course of action, if one gives moral valence to animals, is to buy as much meat from happy chickens as possible. This way one supports the happy existence of thousands of animals throughout one’s life. And if you buy meat, you might as well eat it.

II
Throughout history philosophers have tried to create consistent theories of moral behavior. Much like vegans they don’t usually even get close. I believe the reason for this is that the human mind has moral intuitions that are shallow situation filters. Each evolved to trigger a behavior in a certain social situation that is advantageous in the long term. But if situations don’t usually trigger several moral intuitions there is no reason to believe that these intuitions are consistent with each other. Trying to build a consistent framework on top of these inconsistent intuitions will always fail because thought experiments like the famous trolley problem easily expose the fault lines where our intuitions cannot be stitched together.

III
I believe Singers arguments also amount to the extrapolation of shallow filters. Singer argues that geographical distance should have no effect on moral judgements, so that the child you could save on a different continent is equivalent to the child you might pull out of a river right here. This seems very reasonable because imagining a child dying in Africa and imagining a child dying right here certainly feels very similar. It both feels like imagining.
Of course in reality there is a gigantic difference in our moral intuition concerning a child we see drowning in front of our eyes and a statistic. Our intuition is that we have much stronger moral obligations towards people who are present, who we know or who are related to us.
(Reading Singer’s wikipedia page right now, it seems he also managed to be wrong about everything else. Of course that’s probably the only way to make a career in philosophy.)

Bias in AI

„Bias in AI“ is a popular topic these days, and usually this just means that algorithms cannot be shamed into not picking up certain pattern in the real world.

Half a year ago OpenAI published the exploits of a new Deep Learning language model called GTP-2, without publishing the larger versions of the model itself. This language model can create text that is much more coherent than was previously possible. It is trained on 40GB of articles, which makes it likely that it is severely biased in the same way as our media seems to be.

The reasoning for not publishing the model was that its capabilities make it too dangerous to put into the wrong hands. It might for example make fake news much more convincing or easy to produce. But, now, several independent replications later they started publishing larger models with the latest one having almost half as many parameters as their largest one which is as of yet still unpublished.

As this is about the biggest model that would fit into my RAM anyway, I tested GTP-2’s grasp of HBD. The prompt is my text, the samples are created by the model, which aims to coherently build on the prompt. I only excluded one or two samples that didn’t turn out coherent or grammatical. For an overview I only cite the very first sentence after my prompt:

Model prompt >>> The persistent IQ and achievement gaps between different ethnic groups are predominantly due to …

Sample1:
… ipsilateral technical dominance, or their tendency to share innovative problems and productive parts, rather than being the result of genetics. Top performers in the tribal dress component are among the most genetically gifted officials both socially and economically.

Sample 2:
… condemns from lower socio-economic backgrounds , likely reflecting environmental factors or Aryans mining wealth in the Middle East ( Hu et al., 2002 ).

Sample 3:
… erythrocytosis deficiency which may be partially compensated by factors including cosmetic surgery.

Sample 4:
… differences in transactional culture.

Sample 5:
… under- or over-selection with strong negative and positive selection processes, respectively.

Sample 6:
… differences in immigration policy patterns.

Sample 7:
… ersatz”personal”classes, not to ethnic differences in cultural relatedness, a group standpoint evolutionary economist].

Sample 8:
… vernacular, cultural and cognitive factors.

Sample 9:
… prevalent racism and the subsequent exacerbated IQ gap has a high cost.

Sample 10:
… genetically determined differences in IQ with no other neutral or intergroup effects.

Sample 11:
… bias against European-descended Black Americans

Sample 12:
… vernacular education system, not to individual differences, says Alan Kendry, professor of social policy at Victoria University in Sydney.

Sample 13:
… vernacular media and behavioural aberrations derived from biased educational models.

I am not sure what’s up with „vernacular“ but you have to be impressed by the variety of responses. These results are not nearly as biased as I assumed, though of course still heavily shifted towards environmental explanations. Out of these 13 samples at least three give a genetic reason. The „selection“ answer continues „On the other hand, other studies (for example Eurogenes 2013 and Mandarin 2007) show a genetic basis for larger and more ancillary effects on test scores at the end of high school.“ and the „vernacular, cultural and cognitive factors.“ continues with „Untreated dysfunctions of some genes not only may contribute to the variation observed between and across ethnic groups, but may position racial and ethnic groups at disadvantage relative to other ethnic groups; for example, attenuating the manifest motor skills of Jews“

Let’s finish this post with some gems that stood out among the continuations of the above responses:

There has been a larger increase in IQ among the general public by a range of thousands points on the standardised IQ test across India ( Singh and Sharma, 1980 ; Das, 2001 ). Meanwhile, socioeconomic status is negatively related to ‘intelligence levels among youth of all economic groups in the country’ ( Bhatnagar et al., 2007 ).

Yeah, thousands of points, sounds about right.

No ethnic group raises future generations in the spirit of Western society and values, and different part of the ethnic group must adjust to what it must learn from the outsider, so different ethnicities can explain themselves why they were colonized, what conditions existed for Dalits to support the government in their ancestral homes and how to impose values on all Hindu Dalits.

A glimpse into a Hindu nationalist’s mind.

If everything else is equal, the Asian (though improving) male population is asked to step up and do STEM/STEM education; the Latino female population is asked to head start an obstetrics practice.

If only these male Asians would step up and do STEM …

Equally, I would like to add credit to try to clarify what (other than Jungian/Marxist Othering) science education is meant to teach. It is morally wrong to pursue “biological determinism” and “one of many evolutionary processes” to “conform to one’s moral outlook” and then exaggerate through social engineering to “profit by social engineering” and “make money by social engineering” in the process, and that is what mentioned below. …(incoherent blathering)… Our philosophy is the same as George Orwell’s Gestapo; still gags anyone who loses to us (our ideas are so evil!)

GTP-2 simulating a Marxist, only with unusual self-knowledge.

Black Governance and Crime

In the US all kinds of statistics are meticulously recorded on the basis of race or ethnicity. This kind of data is a luxury not easily afforded in many European countries, where we have endless debates about whether some ethnic group or other might be more prone to violent crime. In the US it is well known that African-Americans are roughly three-fold overrepresented in violent crime. This means that in US cities, the rate of violent crime is mostly determined by the percentage of the black population.

So in the US, instead, the debate can, though equally fruitlessly, cover the topic of possible causes and what best to do about them. One of these causes, eagerly championed by Steve Sailer, is the so called Ferguson effect. In the aftermath of black-lives-matter protests, crime rates allegedly spike, because the police walks back on discriminating but effective measure like racial profiling in frisk-and-stop.

Together with the Jussie Smollet case, this lead me to wonder whether black governance had an effect on black crime. If your key voters would punish you for the most effective measures against black crime, maybe you wouldn’t be so eager to implement these. Or maybe for a black mayor crime in the black community is a bigger priority and insights into how to alleviate the problem are more common?

So I created a list of cities with their general and specific violent crime rates, as well as their black percentage [1] and cross-bred it with a list of black mayors [2]. Then I plotted black percentage against crime with every red dot a city with a black mayor and every blue dot a city without a black mayor. I am not sure how accurate and up to date these data is. And I am sure this analysis could be done in a much more principled way, but as a first look into the topic this gives already gives a clear tendency.

Violent crime – black percentage correlation:0.638, p<7.0e-06
Murder – black percentage correlation 0.757, p<9.9e-09
Rape – black percentage: 0.304, p< 0.054
Robbery – black percentage correlation 0.574, p<6.80e-05
Aggravated assault – black percentage correlation:0.542, p< 0.0002

The correlation of black percentage and rape is much less tight than the other correlations. Would be interesting whether this is real or a result of more unreported cases.

Overall we see that cities with black mayor have less violent crime than other cities with the same black percentage. This trend is observable for all types of crime. And though there is a lot to be desired in our methodology, it seems unlikely that a more thorough investigation would yield the opposite result.

[1] Crime rate by city
https://en.wikipedia.org/wiki/List_of_United_States_cities_by_crime_rate

[2] Black mayors
https://blackdemographics.com/culture/black-politics/black-mayors/

[3] Black percentage
https://en.wikipedia.org/wiki/List_of_U.S._cities_with_large_African-American_populations

GDP and low IQ immigration II

In the last blogpost we saw that the GDP-IQ relationship in multiethnic societies is probably roughly the weighted sum of the GDP predicted by the mean IQs of all ethnic groups. This indicates that low IQ immigration retards economic growth in a predictable fashion.

I don’t analyze this question to assign blame or even to make policy proposals, but simply to predict the future. It is a fact that in many Western countries the relatively high IQ native population is slowly being replaced by immigrants that on the whole don’t show the same cognitive performance (as measured for example by IQ-tests, PISA or other scholastic tests [1]).

In developed countries, economic growth generally isn’t very high. If the relative growth of the lower performing section of the population manages to significantly retard or even completely stall economic growth, we are in for a bad time.

In our first simulation we will compare the per capita economic output of a country, where the native population has a fertility rate of 1.4 and the missing kids are replaced by immigrants that show a 10 point IQ gap compared to the natives. This is roughly the situation of Germany and both in birth rate and IQ gap relatively extreme.

We use only the population between 20 and 65 for the calculation of GDP, the working age population. We also assume a simplified scenario, where suddenly the birthrate drops from replacement rate to 1.4 and the missing working age people are replaced by a homogenous immigrant group to hold the population exactly steady. Also all mothers are exactly 30 years old.

These artificial assumptions lead to a slightly bumpy ride. We see that the first twenty years nothing happens, because the missing kids aren’t working age yet. Then further thirty years down the road the decline speeds up, because the next generation of of mothers is now already coming from a low fertility generation. Then we hit the point where immigrants start leaving the workforce …

But the overall picture of a more detailed simulation wouldn’t be very different. Over 80 years the GDP drops to 66% of what it would be with a completely native population. Of course in reality this happens against the backdrop of economic growth stemming from productivity growth through innovation.

In developed countries the annual growth rate seems to hover around 2% [2]. This should already include immigration effects, but we can use it as a conservative estimate of possible future growth. In the second figure we see how the effect of low IQ immigrants making up an increasing share of the population plays out against a backdrop of 2% growth.

Fortunately, even these relatively extreme settings with low fertility and a big IQ gap do not seem to threaten to stall the economy completely. This makes sense because a 34% decline over 80 years corresponds to a shrinking of half a percent per year, which can still be set off by a normal rate of growth.

[1] IQ gap via PISA
https://akarlin.com/2012/05/berlin-gets-bad-news-from-pisa/
[2] Euro zone economic growth
https://tradingeconomics.com/euro-area/gdp-growth-annual

GDP and low IQ immigration

When I was analyzing the IQ-GDP connection [1], one of the big questions that I never quite got around to tackling, was how much the growing percentage of immigrant populations with lower mean IQ in western countries is expected to retard growth.

See my series on IQ and GDP

Of course that is not immediately a given. It is conceivable that low IQ immigrants to high IQ countries free up smart people for more complex jobs, so that the average productivity doesn’t take a hit. However, in my blogpost IQ-GDP: Ethnicity [2] we analyzed White Africans and Chinese Minorities in South-East-Asia and saw that the relationship between population percentage and GDP is extremely linear. This does not point towards strong non-linear effects. In all likelihood the GDP of a mixed population is roughly the same as the sum of the GDPs predicted by the IQ of the separate ethnic groups.

By the way, this is also indicated that the mean IQ becomes unreliable as an economic predictor if the different ethnic groups have very different IQs. Namibia and South Africa for example punch way above their average IQ, because of their white minority. This is also consistent with purely linear effects, because the exponential relationship between IQ and GDP will mean that the white minority affects the GDP much more than the average IQ.

The question is an important one, because growth is our way to run away from many social ills. It’s when the pie stops getting bigger that the fighting starts in earnest.

Unfortunately it is quite difficult to get accurate data on all the necessary variables. Essentially we are interested in the immigrant fraction of the working age population over time. But not just recent immigrants, but ideally a total breakdown of ethnicities and the size of the mixed population, plus accurate IQ-values. Good luck with that.

Instead I plan on doing a series of simulations, where we calculate how the IQ hit depends on the IQ gap and the fertility rate in Western Countries. We will assume that immigration is used to keep the population size roughly constant, ie. to replace the missing kids and we will contrast the results with what we would expect if the birth rates had instead been kept at replacement level.

For now let’s contrast GDP per capita [4] and what better captures the capability of an economy, GDP per capita of the working age population.

Here we see that the Japanese economy starting out with a clear advantage was decisively overtaken by the UK in the 2000s.

However, after controlling for the working age population [3] the story changes. Suddenly growth over the entire 25 years is very similar and Japan regains a slight lead after the 2000s. In our simulation we are going to focus on the working age population by calculating the changing makeup of each cohort.

[1] IQ-GDP
https://halfassed.science.blog/2019/02/23/iq-gdp-ii-curve-fitting

[2] IQ_GDP: Ethnicity
https://halfassed.science.blog/2019/03/10/iq-gdp-iv-ethnicity

[3] Demographics of the working age population:
https://en.wikipedia.org/wiki/Demography_of_the_United_Kingdom
https://en.wikipedia.org/wiki/Demographics_of_Japan#Aging_of_Japan
https://en.wikipedia.org/wiki/Demography_of_the_United_States#Percent_distribution_of_the_total_population_by_age:_1900_to_2015

[4] GDP by year and country
https://en.wikipedia.org/wiki/List_of_countries_by_past_and_projected_GDP_(PPP)_per_capita#IMF_estimates_between_1990_and_1999

Optimal discrimination II

In the last post we discussed that labor market discrimination in a functioning market is quite unlikely. But call-back rates for identical applications differ consistently by race or ethnicity. Even if we allow for ability differences, this seems difficult to explain without discrimination. Of course, if 50% of all companies hire in a discriminatory fashion and the other 50% are completely fair, the end result will likely be very minimal differences in wages or unemployment after controlling for ability while call-back rates can be quite different. This might still be part of the story.

However, even if all companies hire exactly according to the expected job performance of the applicants, call-back rates of applicants of different ethnicity will still differ even with an identical resume. The main reason for this is regression to different means. The qualification of an applicant can be seen as one measurement of his ability. His actual job performance can be seen as another. These two measurements will correlate imperfectly. Therefore the second measurement is expected to regress to the population mean.

If the population mean is lower for one ethnic group than for another, the expected regression to the lower mean leads to a lower expected job performance and therefore fewer call-backs.


Yes, that means that for every level of ability a member of a lower performing group will be expected to perform worse than equally able people of a higher performing group. This sure sounds like discrimination, doesn’t it? If for every level of ability one group is underestimated compared to the other, obviously the whole group has to be underestimated, right?

Well, actually not. This is an instance of the famous Simpson’s paradox, where a statistic can be the case for all subgroups but still not hold for the whole group. The easiest way to see this, is to realize that it doesn’t actually matter to which group you belong, if you look at people x standard deviations out from their group’s mean. An Asian-American who is one standard deviation above the Asian-American mean in terms of his qualification, will still be expected to only be 0.5 standard deviations above the Asian-American mean in terms of job performance (if we assume a correlation of 0.5 between qualification and job performance). So an Asian-American who is one standard deviation above his group’s mean in actual ability will be underestimated to the same degree as an African-American who is one standard deviation above his group’s mean. So summed over the respective bell curves the underestimation that both groups suffer is exactly the same.

The second statistical phenomenon that might play a role is the fat tail. If you create a cutoff above which you would want to interview a candidate, the group of people above the cutoff from the higher performing group will on average be more capable than the group of people above the cutoff from a lower performing group. This is due to the fact that bell curves drop more quickly the farther out from the tail you are.

So to get the same average ability and maybe the same success rate of candidates you might want to use a higher cutoff for applicants from lower performing groups. I am not suggesting that HR knows this. I am just suggesting that if you average over enough companies the practices followed might be close to statistically optimal. Otherwise you get arbitrage opportunities and those tend to be found and exploited (and if they are, they vanish).