/sci/ - Science & Math

>>14822050
econophysics is bullshit

>>14822050
>Econophysics

>>14822050
you gotta get good at normal investing before you toy around with fancy mathematical models

>>14822570
economists want to get their hands on the dark matter trick.

>>14822050
>>14822552
learn AI instead

also, I'd be reading research papers, not books

>>14822645
They already do. There are no Economic models which agree with reality on a nontrivial basis.

>>14822660
i already took andrew ngs course on coursera on ML and some of the NLP specialization


its just regression and classification lol

>>14822050
Unless you have a job lined up you're going to fail. Study something more rewarding than the financial markets.

> dude what if stock markets worked like particles XD

>>14824580
Immunity dog just killed you.

>>14824580
Errrm actually, Feymann-Kac establishes a relationship between stochastic differential equations and path integrals. So it's not so far fetched.

Consider diesel and network lag parities especially on the bosonic level.

>>14822711
Austrians do

>>14825423
Mention one.

>>14825431
Rothbard

>>14825447
>Rothbard
Oy vey goyim, make sure to atomize your society more. You'll be more "free" this way.

>>14824445
So I should get a job at a firm then work on my own trading algorithms?

>>14825447
I said a model. Rothbard was an Economist, not a model. More specifically, I am asking for a complicated math model on Positive Economics, like the ones studied in Econophysics, that agrees with reality on a nontrivial basis. By nontrivial, I mean a some use of the model that would have been difficult to realise without the construction of model. Protip: there are no such models.

>>14824580
They basically do. Financial markets act like particles with known exact measurements but unknown in-between paths.

Anyway, the best texts for this are first to gain a thorough understanding of stochastic Calculus. Fundamentally, you need to understand how these systems theoretically behave. You need to understand both the diffusion and jump process.

>>14825696
>Get a job at a firm
It's not that easy but yes that would help. Even then, it would probably be fundamentally different than what you'd do by yourself. Stochastic control is a very big deal and many of these firms have enough control (i.e. influence) on markets that they effectively guide them. Where as a little guy, like you, has little control on the market to the point that it behaves much more randomly from your perspective.

Fundamentally markets are a stochastic mess with lots of controls that don't necessarily even out. It's why Black-Scholes didn't solve the market.

I'm not saying don't study it, but you're likely not going to make much money from some advanced model of the market because you really don't know what the players are doing. Likely the best model of the market for simple trading is just the naive Brownian walk and then machine learning algorithm/AI scouring the internet/Twitter for sentiment and looking at other closely related stocks to indicate whether a buy or sell is likely to be better.

>>14825915
cool

thanks for the high iq advice anon

i appreciate it

>>14825796
>I am asking for a complicated math model
Then no.

>there are no such models.
This is one of the core tenets of the austrian school btw

>>14825417
Based statistical mechaniclets ITT btfo

>>14824580
You mean like taking an ensemble of possible basis states with some probability? Yeah imagine that lol.

>>14826057
No, it's not. All economists are a bunch of faggots who are only right about common sense things.

>>14824314
>andrew ngs course on coursera on ML
yuck

>its just regression and classification
although that's a large part of it, there's a lot more to AI/ML than just that

>>14825915
>thorough understanding of stochastic Calculus
>Black-Scholes
>if you truly understand stochastic calculus you'll realize it's also bullshit

If you're gonna be able to pull this off, you'd need a deep understanding of the problem and the tools you have at your disposal. You should quit OP, you're not gonna figure this out

>>14827302
if you truly understand stochastic calculus you'll realize it's also bullshit
that wasn't mean to be quoted

>>14827310
meant, the typos man.......

Not gonna tell you how to Jew the market. That's not what Christ gave me these gifts for.

Aren't you happy with the performance of your Jewish Aladdin Program? Bankrupt Larry Fink and Blackrock first and then we can talk

>>14827310
Is stochastic calculus related to infinitesimal calculus?

>>14827302
what resources do you recommend for ML?

>>14824580