Xkcd 1132
Say no to statistics altogether. If we form a compact front, we can eradicate the disease of statistics from the face of the earth.
As motivation, I’ll explain why statistics is only good for stealing:
- Statistics is used to invest in the stock market, which is stealing by definition
- Statistics is the foundation of modern AI, which as of now is mostly used for stealing work and intellectual property
- There is no real statistical research, but every other paper is forced to have a little useless graph and a p-value made by some statistician, who steals fame from the real researchers who made the rest of the paper
- Statistics is at the core of the gambling industry, which preys and steals from the elderly and economically weak
- Every fucking formula for calculating probability needs to have a “mathematician’s” name even if it’s always sums and scaling that a toddler could come up with. Remembering those names steals neurons from students
- Etcetera
Oh, they played us for absolute fools!
Quality shitpost, but the naming thing is true of virtually everything in mathematics, with good reason, because otherwise you’d just be talking about “that slightly different combination of arbitrary letters by which we do something very similar to, but measurably distinct from, the use cases of the other three equations like it”.
See:
- Pythagorean theorem (geometry)
- Dijkstra’s Algorithm (graph theory)
- Fermat’s last theorem (number theory)
- Peano axioms (formal logic)
- For that matter, the word “Algorithm” comes from the Latinised name of the dude who invented algebra, and the word “algebra” is just an overly truncated version of the title of that dude’s book.
This is also doubly true in science, where there are 5000 different “laws” and “theorems” surrounding something like gas behaviour, so at some point, you have to differentiate them based on their history, rather than what they do. Hence “Charles’ law”, “Boyle’s law”, “Gay-Lussac’s law”, “Bernoulli’s principle”, the “navier-stokes theorem”, “rayleigh-benard convection”, etc…
I’ll admit that was a bit of a stretch. But I also think the naming thing is a problem. Especially in mathematics, even when it is not named after a person, you often have no clue about what it is from just the name (i.e. what do you think is a magma in mathematics?)
I believe that they contribute to understanding, because human minds are wired to engage with stories. If your chemistry teacher was worth their salt, they’d teach you Gay-Lussac’s law by telling you about how, when the hot air balloon was first invented, Gay-Lussac was seen as a mad young upstart by all of the older scientists for wanting to go up in one. Well, not only did he nearly die making measurements, he also showed that, at higher altitudes, there was lower pressure and lower temperature. Then, your chemistry teacher should pull out a spray-can of keyboard cleaner, invert it, spray the liquid into a beaker, and let everyone feel the adiabatic temperature depression from expansion (of course, most of the endothermicity is from the boiling of the liquid, but the point stands) they can explain that any compressed gas gets colder when you release it, whether the keyboard cleaner, spray paint, or the compressed coolant in the coils of your refrigerator. Lower pressure, lower temperature. Gay-Lussac’s law. Now, all of those students will, when they think about the relationship of pressure and temperature, remember Gay-Lussac in a hot air balloon, at low air pressure, and low temperature.
Now that I think about it, I think my teacher called it just “lussac’s law” because you cannot pronounce “Gay-Lussac” in front of a classroom of 14 year old boys. I guess you are right about the stories, but I’m not sure the name actually helps with that
Is this a Bayeskisser meme? 😛️
I told a guy that a wide variance in data essentially means that results were random then he proceeded to explain p values and I’m like “yeah I’m sure the random values values came from nature”.
Moral is p values kinda worth less than variance
Are you saying you can’t determine a difference in aggregate statistics by performing more trials if the variance is high?
You can the issue with A-B comparison is that is kinda expected for one group to be higher on average simple based on the data points you selected.
If the averages are kinda similar and variance are high in both group A and B then I’d say both groups are statistically the same even if the statistical values are different
So you have two groups of ten experiments, mean if group A is 100, mean of group B is 105, variance is 25 (for both groups). Obviously we are not confident that these groups differ.
Now suppose we repeat the experiment two billion times. The group A average is now 99, and the group B average is now 103. The variance is still 25. Are you still not confident that the groups are different?
I’d be curious what constitution a group a versus a group b and how you get 2 billion of them.
But I’ll interpret it as sperm on a race track since you can get 2 billion runs with one nut.
I have to say after billion trials the averages that you calculate did come from random sample, but it would be indicative average for that group since it can’t move far from that calculated average.
I’m visualizing the 2 billion points of both groups and seeing a bell curve with a lot of overlap. I guess they would be different, but overall very similar since the variance is pretty wide.
Right. But that’s what p-values quantify: given the number of trials and the observed variance and means, how likely is it that the two groups are drawn from the same distribution versus actually having different means?
So variance isn’t “more important” than p-values; high variance means that (by definition) your p-value is lower (less confident) than it otherwise would be.
It’s like 50% of Eliezer Yudkowsky’s personality too, if you need more motivation to hate it.
There’s literally a scene in HPMOR where Harry Potter, 11 years old, sits down and gives Draco Malfoy a lecture on Bayes’ theorem.
I maintain HPMOR is one of the greatest artistic works of our time.
Imagine writing 122 chapters of fanfic where you establish your boy genius can do literally anything with magic, and then asking the community to “solve the best way to kill Voldemort” like you think they’re literally all too stupid to see what you’re doing and it works because they are that stupid.
That’s an artist at work, the art in question just wasn’t literature.
and then asking the community to “solve the best way to kill Voldemort” like you think they’re literally all too stupid to see what you’re doing and it works because they are that stupid.
what is it they were doing?
He wrote himself into a corner with his villain being too powerful and well-prepared to defeat and wanted ideas on how to fix it, but he was incapable of admitting he screwed up. So instead he told the readers they had to “guess” his plan (read: come up with one for him), and even threatened to give the story a bad end if they didn’t. It was utterly transparent but his fanbase ate it up.
Be that as it may, I like a good power fantasy story with an intelligent and ethical hero. I like Doctor Stone and The Culture. I don’t like Overlord, because after the first season, which was quite fun, they moved the perspective away from our overpowered and quite ethical hero, Ainz. I like Kamen Rider Zi-O, which has an OP protagonist and plays with themes of good and evil, tyranny and justice, very artfully. And HPMOR, I read and enjoyed before I even knew of all of those. I liked how Harry reforms a Nazi and I like how he thinks Azkaban is shit.
HPMOR isn’t… good… but it’s exactly the specific kind of nonsense that most strongly appeals to My tastes, and it’s a damn sight better than the originals by most any measure.
Still, the best kid genius story remains A Series of Unfortunate Events.
Is the other 50% smugness and delusions of grandeur?
I noticed that Soviet textbooks were frequentist e.g. Khintchin, while modern textbook by Murphy is Bayesian. The frequentist approach makes more sense to me.
