NexusFi: Find Your Edge


Home Menu

 





How to break into HFT


Discussion in Traders Hideout

Updated
      Top Posters
    1. looks_one artemiso with 1 posts (4 thanks)
    2. looks_two Raider with 1 posts (0 thanks)
    3. looks_3 tpredictor with 1 posts (2 thanks)
    4. looks_4 rleplae with 1 posts (2 thanks)
    1. trending_up 3,023 views
    2. thumb_up 8 thanks given
    3. group 5 followers
    1. forum 3 posts
    2. attach_file 0 attachments




 
Search this Thread

How to break into HFT

  #1 (permalink)
 Raider 
Canada
 
Experience: Intermediate
Platform: NinjaTrader & Jigsaw
Broker: Rithmic
Trading: ZB UB
Posts: 143 since Dec 2016

Not sure if this is a good place for this thread but:

How can I break into High frequency trading? Im 18 now and will be going to nyu stern next year studying finance. A few things first: do you think height frequency trading is going to be relevant in say, 4-10 years? Or do you think this market will be saturated? I'm very interested in building an algorithm while at Stern, but I'm not good at programming or computer science yet. On the other hand, I'm very good at math and calculus. What sort of courses should I be taking? How difficult is it to make an algorithm? Also why don't more people opt for hf algos instead of discretionary algos? For a great hft do you need to get a phd or? Also, what makes the difference between an HFT that Goldman Sachs is using and one that some small boutique is using?

Anyways I'm interested in creating a high frequency algorithm down the line, and I should have the resources (stern & access to MBA + grad school work) but I feel behind the curve in the sense that many people my age can code very well already.

Thanks

Started this thread Reply With Quote

Can you help answer these questions
from other members on NexusFi?
Ninja Trader 8 Drawing Tools Wanted
NinjaTrader
NexusFi Journal Challenge - May 2024
Feedback and Announcements
The space time continuum and the dynamics of a financial …
Emini and Emicro Index
ZombieSqueeze
Platforms and Indicators
Looking for an MQL4 MetaTrader programmer/coder
The Elite Circle
 

  #2 (permalink)
 
rleplae's Avatar
 rleplae 
Gits (Hooglede) Belgium
Legendary Market Wizard
 
Experience: Master
Platform: NinjaTrader, Proprietary,
Broker: Ninjabrokerage/IQfeed + Synthetic datafeed
Trading: 6A, 6B, 6C, 6E, 6J, 6S, ES, NQ, YM, AEX, CL, NG, ZB, ZN, ZC, ZS, GC
Posts: 3,003 since Sep 2013
Thanks Given: 2,442
Thanks Received: 5,862

Very interesting question.

But first of all, i would like to see you break down what you understand with "HFT trading" ?

For me :
  • 1. This can mean, very fast trading, based on a edge based on speed. For example somebody buys 10.000 shares of IBM, as IBM is trading on multiple places in parallel, the best possible fill for that order would mean to buy every time at the best bid on all those places. This is supposed to happen unless you explicitly mentioned the place where to buy. What this means is that some people with very fast systems, see the first hit on the book and will grab the share in the other places in front of your nose and sell them a penny higher, based on algorithms.
  • 2. Some parties are liquidity providers and their role is to 'tighten' the spread, between bid and ask. Based on algorithms and the parallel market places.
  • 3. etc...

My goal was to kind of show you, there is different things under the term.


In HFT trading your edge will be based on infrastructure and algorithm.
In my opinion the infrastructure part is a game that you can never win, only the very biggest guys are in that space. Speed is the meters of cable between your server and the server of the exchange. The modules that you used to convert from FIX to other protocols and vice versa, etc.. Your total speed is the sum of all your components, network, servers, modules, etc... You quickly get a feel where i'm going at, it is a sum game of many disciplines.

If you are a brilliant mathematician and you can develop an algorithm that exploits an edge, trust me, 1/2 of a second is not the issue. Of course you want your code to be efficient and resilient but at the end you buy and sell, as fast as the profit is realized, but this is not a millisecond game.

The legislator (i should dive in my library to find it back) defined a HFT as a system that generates more than X orders per second.

For me the edge and the future proof-ness and viability of the last category of systems is much more guaranteed and promising.

Very quickly you will end up in the need for some skills to 'model' and test. While coding may be done by another person, you will need basic skills to model the algo. Python, R, Mathlab, etc... some machine learning tools, will be part of your toolkit... I don't see how you would otherwise be able to 'test' your idea.

Follow me on Twitter Visit my NexusFi Trade Journal Reply With Quote
The following 2 users say Thank You to rleplae for this post:
  #3 (permalink)
 artemiso 
New York, NY
 
Experience: Beginner
Platform: Vanguard 401k
Broker: Yahoo Finance
Trading: Mutual funds
Posts: 1,152 since Jul 2012
Thanks Given: 784
Thanks Received: 2,685



kentonpalmer7 View Post
Not sure if this is a good place for this thread but:

How can I break into High frequency trading? Im 18 now and will be going to nyu stern next year studying finance. A few things first: do you think height frequency trading is going to be relevant in say, 4-10 years? Or do you think this market will be saturated? I'm very interested in building an algorithm while at Stern, but I'm not good at programming or computer science yet. On the other hand, I'm very good at math and calculus. What sort of courses should I be taking? How difficult is it to make an algorithm? Also why don't more people opt for hf algos instead of discretionary algos? For a great hft do you need to get a phd or? Also, what makes the difference between an HFT that Goldman Sachs is using and one that some small boutique is using?

Anyways I'm interested in creating a high frequency algorithm down the line, and I should have the resources (stern & access to MBA + grad school work) but I feel behind the curve in the sense that many people my age can code very well already.

Thanks

A colleague and I were hypothesizing the other day that we wished we were born 5 years earlier so we would be making a killing at a time when everyone was still quite unsophisticated. But we quickly realized if we were born just 5 years earlier, there was a good chance we would've been electrical engineers or professors today because we chose to take courses in what wasn't considered exciting at that time. We didn't set out to join this field, it felt very much that the field chose us.

My general advice to undergraduates is to pick a good mix of courses that excite them and give them interdisciplinary depth rather than to optimize strictly for 1-2 career choices. For two reasons:

1. Finance moves fast as an industry. Nowadays if you want to make a lot of money in the financial industry, you become a lawyer. 5 years ago, it was something else. It's hard to predict whether something will be the big exciting thing 4-10 years from now, because whoever can do so will make a lot of money, and money surely doesn't come by that easily. You do however want to have a diverse mix of skills and knowledge so that (a) there's the off chance that you happen to be among the 1/100 early innovators that happened to pick the right thing to go into, or (b) you have the flexibility to be one of those 5/100 that ends up as an early adopter of the right thing to go into.

2. People change their minds very quickly at your age and there's very limited options for someone whose education and experience is tailor fit to this field. I've seen Putnam Fellows and IMO medalists who were prolific and passionate researchers until their junior year, yet ended up in finance. I've also seen people take 4 internships in finance and switch out to a tech firm by the time they graduated. I'd say more than half of the people I knew through university ended up in a different field than we would've imagined the first day we met, and of that group that changed their career plans, nearly 100% of them are happier now than they would've been in their original fields.

If you'd like to go into this field though, I'd recommend about 2 courses that give you experience with systems programming, be it a C course, an operating systems course, or anything, and a course on algorithms. This gives you context on what your colleagues are doing even if you end up as a trader or quant rather than a developer, and also helps you get a foot in the door at a prestigious firm.

For mathematical background, besides your probability and stats pre-reqs, any applied courses that depend on those pre-reqs will be useful. Ultimately it's unlikely that there's any course that will teach you directly applicable skills and you just need wide exposure to applied math in several domains so that you can pick new skills up quickly.

Very few finance courses will be useful, but I recommend at least 1 course so that you know you have enough passion or tolerance for the field.

As for differences between firms: The most latency sensitive part of Goldman is focused on the trade scheduling problem. On the other hand, small firms are generally more focused on extracting signals.

Reply With Quote
The following 4 users say Thank You to artemiso for this post:
  #4 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011

You absolutely need to learn to program. I would recommend Python and C# for starters. However, data analysis would probably be just as important, and learning environments like IPython Notebook, R, and Matlab would be useful and manipulating datasets. I think one problem is the way you are structuring your question. You want to break into HFT but you've never did any HFT. So, why are you interested in HFT would be my question? What's the allure to that? While not strictly HFT per se, you could start developing strategies now in NinjaTrader or Tradestation (Easylanguage), and I would recommend that.

Typically, one might imagine that strategies have a certain complexity level. A world class discretionary trader probably uses a very complex strategy. A complex strategy means it won't be able to copied easily. It can't be verified easily, either. One might imagine as trading strategies become simpler and more verifiable that another source of edge must come in and that's where the technology advantage comes in. Because most traders don't have access to that technology advantage, it is not an option to run the simplest strategies.

One issue is that you talk about advanced mathematics and they can be useful for something such as understanding how the backpropagation algorithm works. However, most people aren't good at math and don't think mathematically and that might be a hint that understanding the markets is not going to be a mathematical enterprise.

You probably want to decide if you want to be more applied or theoretical. If you are interested in theoretical work then you could start reading papers. NAAIM has some strategies published and publishes on a regular basis and that would be one place to start. It is middle of the road with both applied and theoretical concepts. You might look for and subscribe to other industry publications, i.e. trader mags like Active Trader.

Yes, industry reports that HFT profits are down. I think you should structure your career for a broader array of possibilities. I'd recommend to look for activities that can yield synergistic benefits across a myriad of possible future states.

You might want to look into the definitions for "buy side" and "sell side" for basic insight into firms.

Reply With Quote





Last Updated on January 8, 2017


© 2024 NexusFi™, s.a., All Rights Reserved.
Av Ricardo J. Alfaro, Century Tower, Panama City, Panama, Ph: +507 833-9432 (Panama and Intl), +1 888-312-3001 (USA and Canada)
All information is for educational use only and is not investment advice. There is a substantial risk of loss in trading commodity futures, stocks, options and foreign exchange products. Past performance is not indicative of future results.
About Us - Contact Us - Site Rules, Acceptable Use, and Terms and Conditions - Privacy Policy - Downloads - Top
no new posts