I love William Gosset aka Student
I love his most brilliant paper : The probable error of a mean. This paper revolutionized experimental science : it tells randomness from correlation with so few samples required it is a piece of cake to do science.Basically, the paper states that there are random variations on controlled axis, like when you brew a beer (Student was QA for guiness) regarding the quality of grain, the quality of the yeast, the temperature of the water, the room .... and that if a mean tries to sabotage your QA in Guiness through a simple test we are gonna detect you.
Ok, the first time I read the paper, it was my feeling that basically it is an applied mathematics HOWTO detect non randomness.
Easy to read, nice examples, and at the end a table which modus operandi is given on how to estimate your condidence given your numbers of parameters and how much samples you have. The kind of math physicists expect from mathematicians : given with a notice that is simple to follow : THANK YOU Gosset !
This paper let you read for each input parameters, each measures a matrix of « is random/maybe correlated ». If your measure looks like a galton board it's random, if you have more than one peak, ALERT !
Multimodal is the way we call these camels' backs that appears with non randomness.
The beauty of Student is that thanks to his mathematical studies we can now cross X input times Y parameters in N samples and have X times Y tests made.
It's a massive parallelisation of screening for results.
Student's Test are probably massively used by all the controlers of casinos and « opérateurs de jeux de hasard » around the world to check if by any chance randomness is biased.
This paper is one of the many breakage that followed Darwin paper, when science WAS political. Queen's and King's saw randomness as a perfectly suitable mechanism for selection as an attack on their genome.
Kings et Queens love species deterministicsly sorted, hierarchized, that shows the human at the top of the tree, and the nobility at the top of the human through cross breeding.
It's probably Darwin that inspired physicist to have more interest in telling randomness from non randomness, because they probably came out with the first exploitable data.
You see after Mendel, and before, the catholic church had a habit of noting details about their wine and cultures (what, when they flower, how much they produce). The catholic church was bureaucratic before bureaucracy was invented.
These are the same records that states that grapes are flowering sooner, producing more sugar and mass thus being correlated to the parameter « HOTTER ».
With Student methodology, it is really hard to be climatoscepticists. It is also good for detecting bullshits, that's the derived work by Fischer called the p-values used as the kind of CRC-32 of what to expect in randomness if we try to measure by ourselves. So much information encoded in such a small data is what real intelligence is compared to artifical ones.
I hate Student aka William Gosset
Having a stand on the usefulness of randomness in selection of the species was risky, especially in England. The bitch Victoria was colonizing the world in the name of the superiority of her race, and had a hobby of applied racism with dogs, cats and betting on her horses « pur sang » the week end. The free mason, thought of themselves as (pseudo) scientifics too, applying a certain physics to a certain society. The church too had a word in science since they were basically the best places in the world providing the teaching of it.
Randomness in science came as a mawashigeri in the face of some biggots having the pretention to know everything by determining the root causes of everything.
Let's say that when for making explosives you need good quality insurance and thus taboo or not taboo, you teach the math of probabilty and statistic.
Having a stand on randomness in 1900 was not apolotical, it was tainted by the pseudo-science of justifying the racism (cf this motherfucker of Cuvier). Basically Darwin was seen as an anti-slavery/anti-colonialist/anti-aristocrat with or without his consent or willingness.
So let's say that Victoria loved her point of view of race superiority so she found a scientific paper scientific paper on the topic : « eugenics review » that was applied science for selecting better species.
Don't let be fooled by the absence of already tabooed experimentation on humans, because 1) Victora was an assumed racist, 2) eugenism was coming with another « science » called bio-socialisme, a society where the proven most optimal laws of Nature are applied for the « better good ».
If Darwin is right, the royalist inspired bio-socialisme dies in tremendous agony.
In this context of a certain fear on the mad men beginning to play with the genome of corns (maïze), wheat in a contexte of famine, Student was asked :
« hey, what about bio-diversity, if Darwin is right about natural selection, are we not taking risks reducting the variations pool given by genetic variability ? »
To wich he answered :
To sum up: Winter has in this experiment succeeded, by continuous mass selection, in producing two races of maize, one of which has more than twice, and the other less than one-third, the normal oil content.No worries, eugnism is risk less because the loss of biodiversity is compensated by mother Nature (How do you measure ?) Hence agricultural eugenism (that still prevails as we speak) is good because without a risk.
...
And so we reach the conception of species patiently accumulating a store of genes, of no value under existing conditions and for the most part neutralized by other genes of opposite sign. When, conditions changethe species finds in this store genes which give rise to just the variation which will enable it to adapt itself to the change.
...
It follows that the change appears to have produced the variation which it has merely selected from among those potentially present. ‘Thus we can reconcile the view held, amongst other people, by the late Walter Heape, that the environ- ment produces the required variation, with the older Darwinian selection of random variations, to which it appears at first sight to be diametrically opposed.
And I love william again
Remember, the part I have hidden ?
When, however, conditions change, unless too suddenly or drastically, the species finds in this store genes which give rise to just the variation which will enable it to adapt itself to the change.There is a stark reminder by the eugenist William Gosset that lack of biodiversity is threatening us right now.
When this guy is the corner stone of the justification for human selection of species of agricultural interest it may be a good idea to listen ot him.
Human selections of agricultural species DO NOT WORK when the changes are too drastic according to him. And the actual climate changes can be considered as drastic for plants.
If we trust him, we should check with experimental solid evidences that our selective model beats random selections in case of drastic changes.
