UP | HOME

Automation

Modern “finance” is famous for piling up abstractions on to of each other to the point that no one understands the actual dynamics and this, unironically, keeps the whole thing “safe” since at some point no one touches it anymore - it becomes a solidified cathedral of abstract bullshit, until it crashes down, like that Celsius or FTX .

The finanical theory is general relies haavily on a fundamentally wrong assumption, which underlies the abstract probability theory too, that obverved events (values of a so-called “random variable”) are independent from each other, and therefore some inferred properties, like the “law” of large numbers hold.

Once the independence assumption (and the abstract concept of mathematical randomness) are removed, everything collapses back to uncertainty, which is exactly right, provided that the factors which will create further events are not there yet or did not have enough weight.

This is why event the less wrong, frequency-based probabilities does not apply, because the system under observation are not “closed” and fully observable in principle. Past performance does not imply anything certain, only reinforces common beliefs.

What then about all these PhD-level books? Just abstract sectarian bullshit.

Don’t you remember books about Freudian “psychology” with an ego, super-ego and shieeet? Well, just like the whole field went from Freudian abstract bullshit, via Milgram, to Evolutionary Psychology, finance went from its own abstract bullshit (assumptions of rationality, independence, efficiency) to what we call “behavioral finance” where huge crowds of idiots are swayed by greed, fear, mass hysteria and clever memes in an environment full of “bad actors” of any imaginable kind.

What we need, however, are simplest but correct mathematical modes which formally define what we are talking about. Every single word we use, at least in theory, should have precise mathematical meaning, and this is not that hard as long as we remain grounded into reality and did not invent abstract bullshit.

Where the “real stuff” may come from? From the “wire” which links transacting (in-wire representations of values) entities and from transactions themselves (what goes where).

Thus we do not start from the abstract theories of modern finance, but from reverse-engineering of how particular trading platforms. Binance did excellent job of making its API and underlying basic math publicly available, which was one of the reasons of its enormous success - they made the basic things clear and transparent.

So, unless we are market-makers we have to focus on on-the-wire transactions and how they affect our accounts. From this we could infer precise definitions of vague concepts.

Reactive, not proactive

It is tempting to belive that one could come up with a system that correctly forecasts the future and frontrun everyone. However, there is no such system has ever been made by Evolution itself (and there are fundamental philosophical considerations why this is impossible in principle). What Evolution managed to produce are reactive systems, which have certain “architectural designs” in common, namely - sensors and feedback loops.

To have a systems which does your reactions to the market conditions is enough for everything, and just like in biology, the system should be a hierarchy of layered interfaces, made out of simple pure functions.

Why no one did this so far? Because they always begin from abstract bullshit which constitute the current consensus, and from the memes of current mass hysteria, instead of universals and fundamental principles.

And yes, the actual code must not much bigger that underlying simplest mathematics. The pure functional code should even be an executable specification of itself.

Very few people have arrived at these decisions because, again, the whole industry is an organized “religion” of imperative, object-oriented crap, no one fully understands (yes, this is the same as modern finance). They start from C++ because it is “industry standard” and ended up with piles upon piles of low-level verbosity no one is able to comprehend.

I really envy these Cornell graduates and PhDs who has been taught the fundamental principles and proper languages (since MIT dropped Scheme and slided into the imperative OOP hell, Cornell becamethe number one classic school) and how they just applied their classic principle-guided knowledge to automate trading.

The usual problem is that disentanglement, reformulation, simplification and formal verification takes time, and time is the most scarcest resource after attention span and ability to focus.

Author: <schiptsov@gmail.com>

Email: lngnmn2@yahoo.com

Created: 2023-08-08 Tue 18:42

Emacs 29.1.50 (Org mode 9.7-pre)