A Tale of Two Skulls: Part 2

Sorry this post was so long in coming, I’ve been really busy with my PhD!

The Great Dying may sound like an exaggeration of a name, but if anything it is an understatement. The extinction that occurred at the end of the Permian, ushering in the Triassic, was the most catastrophic since animals appeared on the planet—96% of all marine species, and 70% of all terrestrial vertebrate species, were wiped out1. It was the only mass extinction of insects, 83% of all genera were destroyed: to appreciate this number, a genus is a group of species, and a genus is only considered extinct if every single species within it is wiped out.

This period of mass volcanism, runaway-greenhouse effect, increasing aridity and anoxia ended the synapsid and amphibian dominance over land: with all the large land-dwelling amphibians driven to extinction, their predators—such as Dimetrodon—also died out. The only large land amphibians that remained were aquatic, filling the niches occupied by modern-day crocodiles2,3.

Life on Earth would recover, but it would take 30 million years4, and synapsids would not reclaim their former dominance. Instead, diapsids would claim the land, diversifying into two major groups: lepidosaursthe ancestors of modern lizards and snakes; and the formerly-mentioned archosauriformes. Meanwhile, ocean-dwelling diapsids saw the rise of Ichthyosaurs5, fish-like lizards that ruled the seas.

Shonisaurus popularis, an Icthyosaur from the late Triassic, by Nobu Tamura.

The ‘crocodile-lizard’ archosauriformes would diversify throughout the Triassic, with many forms, from small lizard-like animals to metres-long crocodile-like monsters. By the mid-Triassic, one group of archosauriformes would start to stand out: the archosaurs6.

Archosaurs are best recognised by their descendants, as they would proceed to dominate the world for the next two hundred million years. Earliest to break off were the ancestors of modern crocodylomorphs7—crocodiles and their nearest relatives.

Ornithosuchus woodwardi, a late Triassic archosaur related to crocodiles, by Nobu Tamura.

The remaining archosaurs had, by the late Triassic, diverged into two further groups: the pterosaurs8, flying reptiles that ruled the skies; and the dinosaurs9, the most famous animals of the era.

At the end of the Triassic, two hundred million years ago, these groups had become well-established, and it was just in time: the Triassic closed with another mass extinction, the imaginatively-named Triassic-Jurassic Extinction Event10. This was the final nail in the coffin for any remaining giant amphibians, while marine life was devestated, with the extinction of over a third of all genera—including an entire class11. The cause of this extinction is unknown, but climate change marked with mass volcanism are two of the leading candidates. Whatever the cause, the Jurassic had begun.

A Conodont, the marine class extirpated in the end-Triassic extinction event, by Nobu Tamura.
A Tale of Two Skulls: Part 2

Gender, Sex, and Social Constructs

Gender is a social construct. That’s not to say it’s neither real nor important, especially in a society obsessed with policing gender — money is a social construct, and in a capitalist society its impacts are both very real and very important.

Further, being told what our gender is and means can cause dissonance and distress, and I don’t believe it’s based on conflict with some hidden ‘true’ gender. The way the term is used now is a clunky conflation of numerous factors, from gender-presentation to sex, and as these terms are more rigorously identified and separated, ‘gender’ has less and less meaning.

In a genderless society, trans and non-binary people would still exist, and for those that wished to transition the issue would be treated as an endocrine disorder, rather than a psychiatric one. Why is this? Because the idea of a binary and immutable/innate sex is also a social construct.

While separating ‘sex’ and ‘gender’ is a useful Trans 101 description, I think it reinforces a false dichotomy (the ‘body’ and ‘mind’ being separate entities), and devalues the self-identification of trans/non-binary people (‘your gender is female but your sex is male’ is the ‘politically correct’ way of saying ‘you think you’re a woman but you’re actually a man’).

Beyond this, the way ‘sex’ is used is completely lacking in nuance, and fails to reflect the diversity of humanity. Sex conflates a large number of variables into one title — gonads, chromosomes, genital, hormones, secondary sex characteristics, possibly brain structure — while failing to recognise that these are all distinct and meaningful concepts in and of themselves.

For example, if I am getting an abdominal x-ray, my doctor needs to know where my gonads are, they don’t need to know whether I’m ‘male’ or ‘female’; if I’m growing a distressing amount of hair, my endocrinologist needs to know my testosterone levels are, they don’t need to know whether I’m ‘male’ or ‘female’. Ultimately, ‘sex’ is a categorisation used to enforce social norms far more often than it is to convey useful information about a person.

Gender and sex are social constructs, and upon deconstruction may turn out to be completely useless (and it’s my opinion that they are); however, in a world where both of these constructs exist, they are still real, they are still important, and the identification of individuals on both counts needs to be respected.

Gender, Sex, and Social Constructs

‘Trans’ doesn’t mean ‘Transition’

I’ve recently seen quite a few people use the argument that ‘trans is short for transitioning’, which is used to deny the existence of non-transitioning trans people, and to argue for the legitimacy of sex as a binary — as if etymology of words has the power to alter the reality of what they describe. Still, it’s sufficiently pernicious and irritating that I’ve decided to write this blog post.

Trans as a word is derived from transgender and transexual, which themselves were derived from the German word Transsexualismus (meaning transexual), coined by German psychiatrist Magnus Hirschfeld in 1923. The trans- part of this word is the Latin prefix meaning ‘on the other side’.

Further, the claim ignores the fact that the meaning of trans in its current usage has absolutely nothing to do with the verb ‘to transition’, it’s just a spurious connexion based on vague etymological relations (transition is derived from the Latin transitio, meaning ‘to cross over’, and isn’t an example of ‘trans-‘ as a prefix).

So if you intend to argue about the nuances of sex and gender, the fact that two words look similar does not mean you are free to ignore the vast body of the scientific literature — which includes biology, psychology, and sociology — and ultimately the lived experience of people whose experience differs from your own.

‘Trans’ doesn’t mean ‘Transition’

What is 1 kg of coal?

Bituminous coal, courtesy of Wikipedia

This is a kilogram of coal – okay, so I have no idea how much it actually weighs, but let’s work under this assumption – and its primary use is to be burnt for electricity. What is in a kilo of coal [1]? Primarily, it is made of:

  • 860 g of carbon;
  • 50 g of hydrogen;
  • 70 g of oxygen; and
  • 10 g of sulphur.

Let’s imagine that we have a device that can burn this fully: no energy wasted into forming ash, no incomplete combustion creating carbon monoxide, etc. What would the products be?

  • 3 150 g of carbon dioxide;
  • 900 g of water;
  • 20 g of sulphur dioxide; and
  • 40 MJ of heat.

This estimate is about double the amount of energy usually produced [2], largely due to omitting ash formation, moisture content, and other factors that impede complete combustion. Still, maintaining this generous assumption, and taking into account the 40% efficiency of coal plants [3], and 6% loss from transmission [4], we find that one kilogram of coal can produce about 15 MJ of electricity, about enough to boil 25 kettles of water.

However, there are other trace elements in coal that can be of concern: this kilogram of coal also contains 1 mg of uranium, 1 mg of arsenic, 3 mg of thorium, 5.8 mg of lead, 98 mg of fluorine, 320 mg of chlorine, and 21 μg of mercury [5]. This means that, over an entire day, the average coal power-plant (burning 1.3 kt of coal [6]) will release 520 MBq of uranium & thorium, 1.3 kg of arsenic, 7.5 kg of lead, and 27 g of mercury.

It should be noted that this isn’t an exceptional amount of radiation, but over time it does build up: the area around a coal plant is generally significantly more radioactive than around a nuclear plant, and can be significantly enriched in toxic metals.

Real life, however, is more complicated that this ideal: as noted before, coal only produces half the calculated energy (about 22 MJ of heat per kilogram), many toxic elements remain in the coal ash (making its disposal an issue), and large amounts of contaminants leach from coal stockpiles into the water.

Unlike what the Australian government keeps propounding, coal is not a harmless black rock, and it isn’t good for humanity.

What is 1 kg of coal?

Pokemon Go and Individual Values

In Pokémon Go, each Pokémon has three values that determine how well it will battle: its moveset, its level, and its individual value (IV).

The moveset are the attacks that your Pokémon can use — things like mudshot and fireblast and hydropump — and can’t be changed. Every type of Pokémon has a moveset that is optimal (i.e. gives the highest possible damage per second of attacks available to you), such as Water Gun/Hydropump for the Vaporeon, and this will strongly affect how much damage you do.

The Pokémon’s level is marked by the arc in the background, which stretches from the very bottom left (Pokémon with 10 CP), to the bottom right (Pokémon that have the same level as their trainer does). When you click ‘power up’, you are increasing the Pokémon’s level, which will increase its HP and CP in accordance with its IV.

IV is the only piece of information about your Pokémon that is entirely hidden from you in the game, although it can be calculated using level, HP, CP and the information that your gym leader provides you, through external calculators. IV is broken down into three categories: Attack, Defence and Stamina, each of which take an integer value from 0 to 15, with the total IV simply being the sum of all these values (often presented as a percentage of the maximum value). So, what is a good IV for a Pokémon to have?

Thankfully, this is the same as rolling three sixteen-sided dice from 0 to 15, and we can use familiar maths to calculate both the average roll of m n-sided dice, m\cdot\left(\frac{n+1}{2}\right), and the standard deviation, \sqrt{m\cdot\left(\frac{n^2-1}{12}\right)} . This gives an average IV of 22.5 (50%) and standard deviation of 8 (18%). This means that two-thirds of all Pokémon have an IV between 14.5 and 30.5 (32% and 68%).

To illustrate this graphically, let’s consider every single possible roll of these three dice, written in code here:

for x in 0:15
  for y in 0:15
    for z in 0:15

This results in the bell curve shown in Figure 1, because there is only one way to generate 0 (that is, all dice roll zero), but there are many ways to generate a roll of 22 (192 ways, in fact).

Figure 1: Total IV plotted against the number of possible ways to achieve that IV. This shows a very strong weighting towards the centre.

While it is useful to know how rare your Pokémon is, deciding on whether to keep it should be based on how likely it is to have at least its IV; otherwise, it’s easy to see that 0 and 45 are equally rare. The probability of a random Pokémon having as good an IV or better as your current Pokémon is given by the curve shown in Figure 2.

Figure 2: The probability of finding a Pokemon with equal or greater IV than your current.

This suggests that Pokémon with 80% or 90% IV are incredibly rare, which matches up with what the Gym Leaders say: the top-tier evaluation of a Pokémon places it into this high (80%+ category). Let’s have a look at some particularly interesting IV:

  • 23 (49%) — this is the median IV, or the 50th percentile: half of all Pokémon are equal to and better than it, while half are worse than it;
  • 27 (60%) — this is the tier occupied by the top third of Pokémon;
  • 30 (66%) — the top fifth of all Pokémon have an IV of thirty or higher, meaning 80% of all Pokémon have a lower IV than 30;
  • 34 (76%) — your Pokémon is in the top 10% of all Pokémon;
  • 37 (82%) — at this point, your Pokémon is considered to be top-tier by the Gym Leaders when you select ‘Appraise’, and they are in the 94th percentile;
  • 41 (91%) — this is finally over the 90% IV mark, and represents the top 1% of all Pokémon. Anything with this IV almost has the best base stats of its kind;
  • 45 (100%) — the ultimate Pokémon. Each Pokémon you find has a 0.02% (two in ten thousand) chance of having this IV, making them incredibly rare.

To illustrate how rare these values are, let’s look at Gyarados: you need to catch a hundred Magikarp to make a Gyarados, so what is the highest Magikarp you’re quite sure you can catch (>90% chance)?

For your first Gyarados, you would be expecting an IV of at least 38 (84%). Your second Gyarados should be even higher, with an IV of 40 (88%), but after this, your probability starts falling off sharply. Your third Gyarados should have at least 41 IV (91%), but you can’t be sure of getting a 42 IV Gyarados until your fifth, or a 43 IV one until your tenth! At that point, the chance of you having a 45 IV (100%) Gyarados is 20%, so if you’re lucky you may have a truly vicious sea monster!

For some anecdotal evidence, I have caught 4327 Pokémon, meaning that I should expect to have found about one 100% Pokémon by now. And I have!

A Zubat.



Pokemon Go and Individual Values

Misrepresentation Pt 2 — The Senate

I didn’t analyse the Senate results in my last post, because counting had barely begun. While results still haven’t been finalised (and won’t be for a while), enough has been done now that it’s qualitatively useful to examine them. There are 76 seats in the senate, twelve for each state and two for each territory, of which 72 have been provisionally assigned and four are still in doubt. Excluding the latter, the seats are being allocated as follows:

  • Coalition: 30 Seats (45% of assigned seats)
    • 35% of popular vote
  • Labor: 27 Seats (38% of assigned seats)
    • 30% of popular vote
  • The Greens: 7 Seats (10% of assigned seats)
    • 9% of popular vote
  • One Nation: 3 Seats (4% of assigned seats)
    • 4% of popular vote
  • Nick Xenophon Team: 3 Seats (4% of assigned seats)
    • 3% of popular vote
  • Justice Party: 1 Seat (1% of assigned seats)
    • 2% of popular vote
  • Jacqui Lambie Network: 1 Seat (1% of assigned seats)
    • 0.5% popular vote

This gives a total misrepresentation of 33.8% (remembering that 0% is a senate that perfectly represents the popular vote, and 100% is equivalent to seats being assigned at random), with the usual suspects—Coalition and Labor— gaining the most benefit.

Parties gaining a benefit from this misrepresentation are Labor (7.63%), Coalition (6.3%), Greens (1.2%), NXT and JLN (0.9% each). The other parties that won seats lost on misrepresentation, with One Nation ending up 0.1% short, and the Justice Party losing 0.5%.

The biggest losers, however, were the Liberal Democrats, who gained 2.1% of the popular vote (enough for a seat and a half), but didn’t gain any seats, with the Animal Justice Party, Christian Democratic Party, Family First, and Shooters, Fishers and Farmers parties each missing out on more than 1% of the vote (from 80% of a seat for the AJP to an entire sear for SFF).

These results will gradually change over the coming months, but as a provisional analysis it does offer some qualitatively useful information:

  • The Senate is more representative than the House of Representatives (33.8% vs 46.5% misrepresentation);
  • The system disproportionately favours the two major parties (with a total of 14% misrepresentation, four tenths of all misrepresentation);
  • Minor parties with support from many states lose a significant proportion of their vote’s value: eleven parties had more votes than JLN, but since JLN supporters were almost entirely localised in Tasmania, they managed to get a seat.

As an interesting side note before the end, I decided to see the votes-to-seat ratio of each state (and territory): Tasmanians have the best representation, with only 43 thousand people per seat, while NSW has the worst, with 635 thousand people per seat (meaning a Tasmanian has about fifteen times the voting power in the Senate as a New South Welsh person).

Ultimately, the Senate was designed to serve the interests of the colonies before they became states: it was feared that the population of the large states would leave the smaller states without a voice, and no one wants to completely cede a say in their government. With that, it was decided that all states should have the same number of votes regardless of population, to enable the smaller states to stand up to the populous ones if their interests were not being served, and the Senate was seated according to state.

Of course, this is no longer the case: partisan politics now controls the country, and Senators represent their party rather than state, but this relic of federation still constitutes a significant proportion of the misrepresentation.

The misrepresentation—in both the House of Representatives and the Senate—is deliberately exploited by politicians to maintain the status quo, and since the four largest parties (Labor, Liberal, Nationals, LNP) benefit from it, and control two thirds of the vote, it will never change because it will never be in their interest to change it. Political parties don’t want better democracy, they want more power: the Coalition and Labor aren’t really fighting each other, they’re fighting the minor parties so they can maintain their duopoly in perpetuity.

Misrepresentation Pt 2 — The Senate


Misrepresentation is used in politics as a measure of how well (or badly) a political body represents its constituents.  The simplest form is the difference between the percentage of votes a candidate or party receives, and the total amount of representation they have in parliament e.g. Party A receives 20% of the vote, but only 10 out of 100 seats available in parliament, there is a misrepresentation of 10%. The sum over all candidates/parties is the total misrepresentation.

Total misrepresentation can range from 0% (the parliament exactly matches the popular vote) to 200% (the result is entirely in favour of a party who received zero votes), with 100% being the result you’d expect from assigning seats at random.

In the recent election in Australia, the total misrepresentation for the House of Representatives (which forms the government) was 46.5%; that is to say the total disparity between seats won in parliament and the popular vote is extremely significant. The largest misrepresentation comes from the Coalition (Australia’s conservative party), who won 15.2% more seats than they would have if the votes were proportional. The only other significant winner from this system was the Labor party (ostensibly left, but is doing its best to pursue the Coalition rightwards), with a misrepresentation of 11% in their favour.

Nothing more clearly demonstrates how rigged this system is in favour of the duopoly than the only other party with a large misrepresentation: The Greens. With 9.2% misrepresentation against them, instead of receiving 9.2% of seats (14), they received 0.7% (1).

A misrepresentation of 46.5% indicates that our electoral system is deeply flawed, even if we accept its populist nature. Deep changes need to be made, but there’s no political will to make them, as the winners of this system have no interest of losing their control of the political apparatus, and no one funding them wants to risk competition from grassroots movements.