Stoikov?
About Order Book Models
- Thread starter bbt80
- Start date
-----------------------------------------------------------------------------------------------------
Originally, I've created this thread in the "Journals" forum (because it is supposed to be one).
But somehow we are here in Automated Trading and of course people are right in criticizing my approach. Talk is cheap. Show me the money. I agree. May we move back, please?
-----------------------------------------------------------------------------------------------------------
Back on track:
A lot of traders survive solely by exploiting opportunities they find by mining historical data and finding the so called patterns. There are several books out there teaching heuristics to filter out noise and keep the signal. I think it works. Specially if you add position sizing, risk management, realistic expectations and patience.
You can call them descriptive statistical models. Technically they present a challenge: a lot of free parameters. The retail trader have a solution for this problem: trading platforms with optimization tools (Genetic Algorithms, Particle Swarm Optimization etc.). Believe me, you can be profitable this way. There's a lot more I could talk about it. But everything you need to know is in those books, and all you have to do is to come up with a process of observe -> idea! -> test -> optimize -> test -> reject -> observe -> ...
It's hard, but not impossible.
I'll move on and left this behind. Remember the "a lot of free parameters" above? Goldilocks would say the porridge is too hot.
Originally, I've created this thread in the "Journals" forum (because it is supposed to be one).
But somehow we are here in Automated Trading and of course people are right in criticizing my approach. Talk is cheap. Show me the money. I agree. May we move back, please?
-----------------------------------------------------------------------------------------------------------
Back on track:
A lot of traders survive solely by exploiting opportunities they find by mining historical data and finding the so called patterns. There are several books out there teaching heuristics to filter out noise and keep the signal. I think it works. Specially if you add position sizing, risk management, realistic expectations and patience.
You can call them descriptive statistical models. Technically they present a challenge: a lot of free parameters. The retail trader have a solution for this problem: trading platforms with optimization tools (Genetic Algorithms, Particle Swarm Optimization etc.). Believe me, you can be profitable this way. There's a lot more I could talk about it. But everything you need to know is in those books, and all you have to do is to come up with a process of observe -> idea! -> test -> optimize -> test -> reject -> observe -> ...
It's hard, but not impossible.
I'll move on and left this behind. Remember the "a lot of free parameters" above? Goldilocks would say the porridge is too hot.
Welcome to the dumbshits club in this thread.
Dude, you can’t google? All you need is the keyword to unleash a never ending chain of information.
That's not the point. If it were more than 3 letters....then maybe.
Example: Why don't you goggle this: "ABC", then try "123" ?
For the original poster not to provide even 1 link ? How f*cking lazy is that ?
If you add context, you will get the right result - “AQR + HFT”. The reality is that the most knowledgeable contributors have the least to gain from posting here, so whenever and whatever they post, I give a big mental thumb up! Would it be nice if they just laid it out in layman’s terms? Sure, but beggars can’t be choosers, imho.
You are not portraying yourself as a seasoned systematic trader, unlike @syswizard or @stochastix (though it's also possible that they are trolling), so you can get away with not knowing who the key market participants are.
AQR (https://www.aqr.com/) is a large systematic asset manager, probably the largest of its kind. They kinda suck at trading and their founder is a total douche, but they are big and well known.
Unfortunately, he died this summer
I am still heartbrokenThe ultimate examples are all of the flash crashes we have experienced in the recent past. The idea is that if they find themselves in an environment where they might have to carry inventory beyond their horizon, they will stop providing liquidity. If you have access to order book data, you can see how the order book activity changes when you get big events like these with some imbalance remaining. For example, if you find some smaller stock hitting an air pocket, most HFT MMs who got hit on the first gap will cancel and wait to submit new orders until they see some bids.
Several guys here (e.g. @ValeryN ) are doing automated swing trading and they are (knowingly or intuitively) acting as second-level liquidity providers.
Dude, fuck that shit, how hard is it to type intelligibly, don't you use your fingers? fuck all that task switching and hailing lord google every 5 seconds. and since i do not worship wealth i do not know the name of every capital company there is... they are all boring and unremarkable
But back on point, you will find Volterra integral operators worthy of study I am sure. https://en.wikipedia.org/wiki/Volterra_operator when dealing with point processes