HFT

Are you jewish?

>>62310844
I am not! Though good question. English ancestry and 100% white American.

Are your bosses jewish?

Do they talk about genociding white people in earshot of you?

>>62310838
Some years ago I read an article that some system builders were making overclocked, watercooled servers for HFT people, since the business is about "gotta go fast" instead of "make moar coars" like the rest of the server world.

Is that true? Are they still doing that?

>>62310838
Why an AMA on 4chan?

>>62310851
Hmm I work with some jews but I don't think any of my bosses are jewish. Most people are Chinese actually. Quant firms are full of nerds, it's a bit different from the demographics of Goldman Sachs.

What operating system do HFT servers use?

Is it true that HFT servers are literally in the same building as the exchange so they can do trades even faster?

What qualifications do you need to work in Wall Street? How much $$$ do you make?

>>62310854
That sounds silly, it seems to me that analyzing stocks naturally lends it self to parallelization.

>>62310838
What does your hardware topology look like? I've heard HFT is done almost entirely on FPGAs.

>>62310838
Are you guys hiring?

>>62310854
I'll be honest I don't work on the systems/devops stuff so I'm a bit removed from the live trading computers. But I will say, while speed is important, the raw game of "gotta go fast" was a race to the bottom from the get go and we've reached it. Profitable strategies can no longer rely on simply getting information/responding first.

>>62310857
lol idk I'm a loser faggot I guess

>>62310865
We use an in-house unix-like OS

Yes it's called "colocation"

>>62310868
You need to be very good with math and programming. And what I make is different every year based on profit share. Base pay is $300k but bonuses can be anywhere from $200k to $1M depending on the year

Also most people went to a top school though there are some exceptions to that

>>62310876
>That sounds silly, it seems to me that analyzing stocks naturally lends it self to parallelization.
If I remember right they said it was so they could get orders sent in a few milliseconds before the next guy and front-run him.

>>62310898
>the raw game of "gotta go fast" was a race to the bottom from the get go and we've reached it
weren't there a few places that spent piles of money having their own dedicated lines leased or installed just to shave off a few milliseconds of latency? But yeah makes sense, the speed of light will shut that down at some point

>>62310876
parallelization is too slow lmao

Yes FPGAs are used but like I said Im a quant dev. I dont work on the live trading systems I work on the strategy development so Im not as knowledgable about that stuff sorry

>>62310886
Always but we reject almost 100% of applicants because most people arent able to pass our interviews

>>62310899
> a few milliseconds
nigga we operate on the nanosecond scale.

>>62310911
Yup! First there were new underground data lines then there were microwave towers built. But that game kinda slowed after 2014 since it became basically impossible to improve without prohibitive costs. But ya billions have been spent just to shave milliseconds and nanoseconds

>>62310914
>parallelization is too slow
What? You have a single FPGA do all your trading?

Also, what do you mean by strategy development? You make high level rules for trading that peons have to actually implement?

>>62310914
>Always but we reject almost 100% of applicants because most people arent able to pass our interviews
I had an interview at one of Apple/Amazon/Facebook/Google/Microsoft a while ago. I nailed every interview and solved all the problems right on the spot, would that be good enough? (I declined the offer because I didn't want to relocate)

>ama
Get out reddit.

>>62310933
No we use multiple machines!

So there are two sides to every quant firm. You have strats and implementation. Basically I use math and Python (with some machine learning libs) to identify signals in market data. It's all about research and backtesting

>>62310940
>(I declined the offer because I didn't want to relocate)
Why apply in the first place?

>>62310940
LMAO no. You think shitty leetcode problems are difficult to solve? Nah senpai. The questions that big tech companies ask can be solved by anyone with a CS degree if they take even a week to study CTCI on their own. You need to have a very good understanding of probability/stats as well as a good abstract reasoning ability. There are more engineers that currently work at facebook than work at ANY top HFT firm. This is a very small and elite industry.

>>62310961
> Nah senpai

Ah it's been a while since I've been on here. Looks like "Nah fām" get's replaced with "Nah senpai"

>>62310961
Then give an example of just how difficult your interviews are?

>>62310997
I suppose applicants have to solve a millennium problem to get a job there.

>>62310997
Jane Street is not an HFT firm, but they're one of the larger quant firms and there is a good list of examples on their glassdoor

https://www.glassdoor.com/Interview/Jane-Street-Interview-Questions-E255549.htm

Which programming language is most commonly used in your field?

>>62310914
I'll apply. Fuck your rejection rate.
>>62310961
Give me what to study up on. I already have an algo trader which has been netting me some cash over the past year. What sort of stuff do I need to do to move to the super 1337 industry instead of just my hobby project.

>>62311011
> Angle between minute and hour hand at different times.
kek
>Given a matrix of chars from the set {"u", "l", "d", "r", "x"}, where the matrix is guaranteed to have one "x," and each of the other chars represent a direction up, down, left, and right, and a point on the matrix, write a function that returns true if "x" can be reached by following the specified path.
Pretty easy still

>>62310838
What's the job like for the infrastructure/system developers? How much is the pay, what language do they use, how competitive are those positions, what qualifications/degree do they usually have, working hours, etc.?

>>62311033
Python for strats
C++ for live trading implementation

>>62311034
Well first off congrats on making money from a side project! The best "path" in if you don't have a degree from Harvard, MIT, or Stanford is probably to work at Google or Facebook as a data scientist or software engineer. Do well there and then you should be able to get an interview in HFT

>>62311041
Then apply!

>>62310914
>parallelization is too slow lmao
the same bullshit larper as last time

>>62311047
So for systems work I think being an expert on linux/unix-like is a must. So I guess that means they use a lot of bash?

I dont really ask my coworkers what theyre payed but I mean I doubt anyone is pulling down less than $250k total comp even for systems roles. But those are still pretty competitive, though less so than the profit-center roles

Top degree or experience at a prestigious tech company will help you get noticed

does anyone know the vhdl code to make an or gate using only nand gates, i have two inputs A &B

>>62311054
Last time? What?

And ya totally this is all bullshit. Everything I said is 100% a lie. Literally nothing is real and it's all just some fantasy. You're SO smart :)

>>62311041
The matrix question seems weirdly easy, unless I'm misunderstanding it.
Like, just follow the four possible paths from the starting point, mark any visited nodes and quit a path early if you encounter an already visited node.

>>62311118
It's expecting an optimal dynamic programming solution, not a brute force solution like you're likely thinking

>>62311133
Ah that make sense, sorry for being a codemonkey brainlet

>>62311075
its due in 5 hours help a noob out

I am soon to be a PhD in physics (computatuonal physics) guy. What are my chances of getting such job?

>>62311160
Awesome! You're just the kind of person these companies look for. Where are you studying?

Tell me more about the machine learning please. What do you use as input - chart data only or do you analyze news articles too? What libraries? Which algorithm do you use for training?

>>62311171
Thanks for hope man. I am doing some protein folding algorithms and shit. We do that in linux so I am familiar with quite some things you've said earlier like python bash etc.

That being said I do have zero knowledge of finance though. What are my chances?

>>62311192
Ooo good question! Unfortunately I can't get specific about what we use, that starts to go into proprietary information. BUT we obviously make use of all relevant data. The most important is the full history of price/trade data across all exchanges we trade on. We use this data for backtesting. But we of course make use of anything that might give us signal ;)

>>62311210
>But we of course make use of anything that might give us signal ;)
How about the cosmic microwave background radiation?

>>62311194
Finance knowledge is irrelevant. It's easy to find assholes from NYU Stern and UPenn Wharton that worked as business analysts for Goldman Sachs. That knowledge is very easy to learn and we teach it to you! The more important things we look for are strong understanding of math (focusing mostly on stats) and a strong research background. Are you published?

>>62311236
We stick to things that have strong predictive power

>>62310940
>Always but we reject almost 100% of applicants because most people arent able to pass our interviews
If I solved the Riemann hypothesis would that qualify me?

>>62310838
What kind of programming skills do you need for your position?
Any suggestions on what kind of programs I can make as a side project to show potential employers?
Realistically I don't think I can get a job like yours now but it'd be nice to have something related to it on my GitHub.

>>62310838
I'm finishing off a PhD in statistics. How do I get maximum pay?

We fuck around in R with the odd bit of C++ to speed shit either inline or with a poorly written compiled blob. I'm doing KDD and Kaggle for Python but that doesn't seem useful unless you are doing high level strategy at Citadel (where they like to hire Physics and look at feature sets you'd never have thought of on your own).

I thought about maybe directing my research to time series because 1-2 professors in my department had done work in "finance" before. It turned out that their notions were 15 years old and the time scales were orders of magnitude too slow - think LTCM era. The stochastic process theoretician kept talking about Black-Scholes like it wasn't 35 years old and the times series woman literally asked me if I could improve her script for scraping quotes from Yahoo Finance.

That shit ended quick. However, learn about TCP vs. UDP and colocation and the SSD/HDD tradeoffs (with no CS background). It taught me that my professors should never be trusted to know about anything other than publishing papers. I did call IEX (from Flash Boys) a few times twice and and they were super helpful and said that I'd need $10,000 to address my research outline.

How the fuck can I pivot into a market facing job given that I don't know how optimize compiled languages, don't have access to processional tools like KDB (only know about from job postings) and am super over-qualified to fuck around at a Craiglist-tier prop shop?

I'm dreading taking a job that involves 3 hour meetings but a comfy salary to fret over failing a central bank stress test (which won't happen because the executives were all roommates in college).

>>62311253
If you published a peer reviewed paper about it, word would definitely get out. I'm sure you would have no problem getting an interview

>>62310838
What algorithms do you use for data mining and on what key elements?

>>62311239
Misread your initial post. I am in krautland. I've published a few papers. And to be honest I joined molecular dynamics field because I was hyped from biotech firms but seeing the scope is few these days, I am trying to jump ship to non physics field.

File: beetlejuice reaction.jpg (54KB, 381x327px) Image search: [Google]

54KB, 381x327px

>>62310838
Hey, I'm a legit new student in a Quant program in NYC.

What advice could you give regarding coding for high-frequency trading? Is there something you wish you were taught in school, or anything else the industry is doing that isn't obvious until one enters it? Thanks in advance!

I've been fucking around with an algo trading bot I wrote myself in Python, but I'm finding it hard to structure the bot in such a way that multiple strategies can be tested out and evaluated on my tick data.

Any good resources on this kind of thing you can recommend? Ebook recommendations? Do I just need to learn how to program better?

>>62311210
>across all exchanges
So your concept is arbitrage trading?

>>62311261
>What kind of programming skills do you need for your position?
Read the thread

> Any suggestions on what kind of programs I can make as a side project to show potential employers?
If you're just looking to get noticed, I would say it's unlikely a simple side project would improve your resume enough to make the difference

> Realistically I don't think I can get a job like yours now but it'd be nice to have something related to it on my GitHub.
Then do it! It's fine to have passion projects

>>62311273
> I'm finishing off a PhD in statistics. How do I get maximum pay?
Well you're on the right track! If you're asking about absolute max, RenTec pays the most and also lets you invest in their Medallion fund. But I dont work there and it's unlike you or I ever will lol

> and the times series woman literally asked me if I could improve her script for scraping quotes from Yahoo Finance.
I actually laughed out loud

> with no CS background
As far as the resume goes, this isn't a huge deal. But you will need to fill the gaps with independent study

> given that I don't know how optimize compiled languages
Dont worry about that. You are looking to become a quant. Quants only care about the math not the implementation optimization.

I think you should seek out a mentor. Someone in industry that can help you navigate the process. You have the right background, but your professors arent going to be the ones to help you get paid.

Where are you studying, btw?

>>62311302
A quant program? Do you mean an MFE?

If youre studying to be a quant, focus on that. Dont worry about the specifics of HFT. The people that write C++ go a very different route from what youre pursuing

>>62311331
What are you struggling with exactly? I dont know what you mean by bot "structure"

>>62311334
That's certainly a big part of it. Statistical arb is a great way to find mismatched prices to turn a profit! But that's not the only thing we do.

>>62311369
Well, I was under the impression MFE and Quant. Finance were two different names for the same thing.

So, if C++ is only really useful for HFT, is Python + R used for everything else? Any specific tips on what packages/etc you use in the industry? Really anything you know would be useful, even if by whetting my utter beginner's curiosity.

>>62311487
C++ is used for the live trading

Python is what we use but R is great too! We develop all libs in house to fit our needs. I know that seems silly but it's kind of the way of the beast

If youre looking for actionable advice, get a data science internship at Google or Facebook. That combined with your MFE will hopefully get you an interview

>>62311338
>RenTec pays the most
I have a friend from Caltech that works there now, for what it's worth he really likes the environment, and obviously the pay is crazy good.

I interned at EWT back in undergrad, they went on to merge with Madison Tyler and then Virtu, but I went on a different career path. What do you consider the top shops now days? RenTec being the defacto #1.

Any opinions on HRT, GTSX, TGS in Irvine/Princeton or Narang's new shop Mana? Thanks for doing the AMA btw, nice to have a good thread on here every once in a while.

>>62311563
> I interned at EWT back in undergrad
Neat never heard of this firm before

Oh dude this just got weird. I wish there were PMs because I'd love to chat about specific firms but then we start to get into personally identifiable information about me and my experience

>>62311520
That is food for thought. Thanks!

>>62311563
The only thing I feel comfortable saying is I have not and likely never will (as much as I want to) work for RenTec

I dont know anything about maths, programming or finance. Could I still get a job at your firm.
I am over 50 too.

>>62311636
Can you? Yes. Will you? No.

Algo anon back again.
>>62311050
I don't work as a data scientist, but I work with a bunch. I mostly am coding in Python and Rust. They code mostly in Python and R. No degree unfortunately.
>>62311133
That question sounds a lot like A-path strangely enough. There is probably some trick I'm not aware of.
>>62311210
Mine is pretty simple. I take the full data set of the asset I'm trading. From there I look at a few technical indicators to try and guess the moving average and relative strength of the stock based on much more recent data. Because I work at a slower pace I still enter and exit trades manually because I don't trust it to not explode at any moment. It hasn't yet, but better safe than sorry. I just found I was wasting too much time trading and always looking at the same data to make trades so I automated it. I don't use it to find new assets, but move in and out of positions I trust to move in a generally strong upward direction.

>>62311338
Thanks for the advice. I bought a suit with money I don't have at Macy's today for a CCAR interview on next week. I want the job badly but just because it's 100,000USD+ in NYC and I've been at 20,000USD for the past few years.

As my grammar suggests: I'm drinking hard liquor (and eating ramen) while slumming it in Brooklyn. Job applications are highly divergent with quant jobs and data science jobs having very different pre-reqs. I've heard that some firms only hire grad students from certain disciplines so the managers can benchmark applicants to their own life experience - physicists are most widely loved so that they may work in unrelated commercial domains(?). The CCAR jobs seem like too much brain power aimed at too narrow a problem since nobody wants to risk pissing off regulators.

As an aside I got into the MFE at Cornell and Carnegie Mellon but during campus visits the students were terrified about not getting internships and having huge debt loads. I went with the PhD instead that year. It cost me no cash but a lot more of my faith in humanity. :)

>>62311192
>do you analyze news articles too

Remember when someone got into some military twitter and reported explosions at the white house? There was a millisecond long Great Depression and stocks sunk through the floor.

>>62311613
EWT was pretty cool back in the day, this was early-mid 2000's and it was still a very tight knit shop and they were printing money. They, along with TGS where the main quant firms out here in LA so they recruited heavily from Caltech, normally Wall St didn't pay much attention to us.

>we start to get into personally identifiable information about me and my experience
I know what you mean man, its a small world out there, especially in the quant/HFT space.

>>62311626
No shame in that anon, they set the bar high. With that being said, I know most quant folks want to chalk everything up to meritocracy, and while the quant world is far closer to a pure meritocracy than Wall St will ever be, you can still open some good doors if you have a good rep and know the right people. No joke, but I met Nat Simons through a mutual friend on a fishing trip and he offered me an interview at Meritage, obviously this isn't RenTec but still, shows how much random chance can play a role.

Are you west coast of east coast based, if you don't mind divulging. Any idea how well Edgesteam is doing? I know they merged with Merfin in Walnut Creek a while ago.

>>62310838
Make me rich with one sentence

Are you a furry in your spare time?

What are the different types of quant jobs and the requirements for each?

>>62310838
I don't know enough about taxes to think of questions about them but tell us interesting things about how your company pays taxes

Is a gap on my resume before starting college a bad thing? I wasted three years before finally registering for a math major. Now I'm doing good but three years seems like a lot.
Is it a death sentence to not be from an elite university? I'm European, so things might be different here than in the US, but while my uni is good, it's not Cambridge/Oxford/ENS good.

Can I make money with a cryptocoin trading bot?

Do "classic" traders even exist anymore or is it just quants now?

>>62310898
>Yes it's called "colocation"
how big are the servers
are you considering amd

How much influence does (((Wall Street))) have on the us government

>>62311273
Try Morgan-Stanley. You'd fit in.

>>62314306
Are there different company cultures in Wall Street?

Awesome thread.

So did OP leave?

On the off chance I'm from hobotown midwest and got a mostly worthless finance 5 years ago.

now I spend all my free time studying math, economics, commodities, microstructure and designing strategies.

Is there any way I could get an interview without high-profile stuff on my resume? I'm too old to intern at google/ I don't know why they'd hire me as a data guy over some young CS grad.

I'm not even that interested in the money, I just wanna be part of 'the center'. I always figured you guys were the center, that's why I went to school for finance. But to my dismay, my backwater highschool and uni taught me that trading was still mostly happening on the floor. When I asked about market making or HFT it became clear that none of these guys ever made it out of the state. Working in the project finance dept of a mid-cap corp is as high-speed as they get.

>>62317939
>trading was still mostly happening on the floor
Aren't traders getting replaced?
Either way, I thought the skill set required was mostly the same. Quants are, of course, required to know a lot more math, but that's it

>>62317971
there's still open outcry in some of the CME pits, but I visited a while ago and it was mostly a ghost town. The guide said it was a busy day.

I'm just bummed that I went to a shit highschool and subsequently a shit college that didn't afford much opportunity or even paint a picture of how the financial world works that isn't caught in 1993. Pivot tables are the height of technological sophistication around here.

Since I realized how fucked I was in my junior year I've been self-teaching math and data based on curriculum & textbooks listed in data/math/finance programs at top schools, but my resume is totally fucking bare of anything impressive. Purchasing manager for a downtrodden US manufacturer? Big deal. Want to talk about trends in commodity markets and what you think price action on corn futures indicates? "Fuck you, think you're smarter than everyone, too good for your hometown, huh?"

I'll probably move to a city soon, but again, idk how to catch the attention of an employer with my background. In person I'm not afraid to do a little social engineering to physically get into an office, but in my experience, financial firms don't see that as gumption but as arrogance. But it seems like the only other option is to send my resume in just for it to get auto-filtered by an HR bot when I don't have required schools or job experience.

Guess I'll stick it out as a hobbyist for now.

>>62318286
Well that fucking sucks anon.
Isn't it possible to make a decent living as a hobbyist, though? As long as you have good knowledge and some starting capital you should be fine.

>>62318333
I make some fun money, but growing up where I did, I've only held 5 digit jobs, so I could only afford to use 5k as starting capital. I see returns, but it's not like I can live on it. Also, since I have to work to support myself, there's a limit on the success of my strategies because I can only dedicate so much time to watching the market, backtesting, programming, etc.

a pianist who only gets to learn and play on the weekends may have lots of potential, but won't ever be carnegie hall worthy unless he gets the opportunity to work on it full time. Time really is a limiting factor, moreso even than capital or retail fees.

>>62310838
I'm really interested in high frequency trading, algo trading, quant, etc, and have been studying it for awhile. But only out of curiosity, not actually trying to trade my own money because I doubt I could beat 'em. I want to learn more about this because it's very interesting. How to break into this if you don't have a degree and the right pedigree? It seems to me most of the job postings say they require not only a degree but it has to be from a top 10.

>>62318333
I'm heading toward my 30's now, so it kinda seems that if I ever want a family I'm gonna have to give up even more time and just set aside the dream. I wish there was another way.

If I have kids though, bet your ass they'll be exposed to more of life than "highschool doesn't matter, you're gonna work in this town forever, might as well blow off you classes and drown yourself in booze and weed on the weekends. Philosophy? What's that?"

>>62318610
not OP, but go to your local university (if there is one, otherwise call and ask if you can buy a guest library access) and use their system to look up scholarly journals on microstructure, market making, speculation, arbitrage, market connectivity, etc.

also, major exchanges list firms that hold a seat. go through the list, figure out a firm that you think does the kind of trading you're interested in and shoot off an email like "hey, I'm a highschooler/freshman college student and want to work toward a career there. what do I study?"

>>62318693
I already probably have hundreds of pdfs, papers, etc (arrr matey) about those topics market microstructure, orderflow, market making, etc. I also have a lot of textbooks like Larry Harris, David Aronson, Adam H. Grimes, Ernest Chan, Barry Johnson, Perry Kaufman, Rishi Narang, Kevin Davey, etc and many many more. But I'm not a student I have no degree and I'm over half way through my 20s so too old to even pretend to be a freshman or HS student.

>>62318876
well we're in a similar boat it seems, check out my posts:
>>62318286
>>62318592
>>62318645

I think the answer is: put some money into a fund and start running algo strategies on TD or Etrade until you can afford to do so in a more sophisticated manner on IB. That's what I did.

Past that, there's not much you can do but send resumes out with examples of your strategies+returns to companies and see what response you get. Or maybe move to a city and try to network with those people. I don't know, the only time I got to meet one was on a trip to NYC, I overheard to Virtu workers having a conversation.

Most people don't get to live their dreams, especially when their dreams are such a small target as ours. No reason to stop trying though. Best of luck, friend.

File: 1486460004710.png (198KB, 999x834px) Image search: [Google]

198KB, 999x834px

>>62319314
Eavesdropping in a city with a lot of trading firms (Chicago, NYC) is an interesting idea. So how does that work out, butting in on other people's conversation talking about trading?

In regards to trying it on your own, there are many reasons why you will fail. But every once in awhile I see some stranger on the internet allege that they have been able to.
https://www.reddit.com/r/algotrading/comments/28teiu/how_to_build_a_trading_platform/

File: 1501861884863.png (201KB, 952x888px) Image search: [Google]

201KB, 952x888px

>>62319493
Here is another briefly about colocating you server
https://www.reddit.com/r/algotrading/comments/5p3jja/anyone_trading_on_level_2_data_and_collocated/

>>62319314
And in regards to ideas about getting in in a non-standard path,
https://warosu.org/g/thread/61549484
>I work on Wall Street as "quant" for a high frequency trading firm. AMA
>I was in school for CS but dropped out. I instead became a REALLY good software engineer and then transitioned to an HFT firm that way. Once I was in I could prove my quant skills and then I was able to make an internal transfer, even though I didn't have a degree.
>CS dropout got in an HFT firm You got in as a quand dev right? How hard was it to make the jump from developer to trader? Do you mean to say trading firms don't have degree requirements? By "really good software engineer" you mean C++ expert?
>Yes. They'll hire software engineers without degrees if they are C++ experts. Then, once youre in the company, you get to work with everyone else. You become close to the quants, show them your skill and ask about a role transfer. You'll still have to reinterview, but you not having a degree won't stop them anymore because they already trust that you're skilled.
>If you think you can make it in this industry without a strong math background, you'll be stuck on the software side forever.
>[phd] It's "mandatory" in that it's the easiest and most common path. You'll have to overcome obstacles along the way if you don't have a PhD.
>I was actually programming the platform itself, not a quant dev role. Pure software. And it was hard, but once you work with those people every day, it's not too bad. You just gotta show them what you can do! Do you mean to say trading firms don't have degree requirements? It's a VERY rare case that someone can do what I did. I would say they require a degree, unless you can prove you didn't need it. By "really good software engineer" you mean C++ expert? That's part of it (important for the interview). But also because I became a senior under 3 years and had strong references about my talent when I applied to trading firms.

>>62319493
https://news.ycombinator.com/item?id=1517339
Ask HN: Is it feasible to do high-frequency trading as an individual

https://news.ycombinator.com/item?id=9686841
A high-frequency trading model using Interactive Brokers API in Python

When's the s&p crash coming?

>>62319493
>So how does that work out, butting in on other people's conversation talking about trading?
Haha, I didn't interrupt them. I was standing nearby and noticed two guys talking about order flow/finance etc. It was an experienced guy talking to an obvious new hire about what it's going to be like working there. Older guy gets off at a stop and I said I couldn't help overhearing that they were talking about market making, which was why I chose my degree. I asked the young guy what his educational background was, etc. etc.

Funny thing was that this PHD math guy was genuinely kinda nervous about "not knowing anything about finance". I reassured him that if he can handle PHD level discreet mathematics, he won't have any trouble picking up everything he needs to know about financial topics. Hell, CAPM and the market line theories don't go much past early algebra. Nice guy, probably thought I was a little off my rocker. I don't blame him.

>In regards to trying it on your own, there are many reasons why you will fail.
Of course. I didn't say I was making a ton of money, just that my wins tend to beat my losses. Of course I don't aim at HFT strategies. Most of what I do is algos that run over the course of a few hours to a week based on a stochastic implementation of ideas I get from reading or concerning upcoming news/economic releases/earnings reports etc. Like I said, I'm merely a doleful hobbyist.

Thanks for those links though, maybe I'll find some helpful stuff.

Seems you're just talking about the prestige and pay of your job. I don't see how this is pertinent. Why don't you go fuck off to biz or adv, rent seeker?

>>62310838
What kind of time series models do you use?

What's the dependent variable?

>>62310838
We all want a cut of the pie, but obviosuly HFT is prohibitive to the layman.

Is there any advice you can give to the "basement trader" that wants to make his own automated trader? I have lots of programming experience across various languages but I lack any financial knowledge. Should I be day trading and develop my own strategy naturally and just code that?

Maybe >>62311034 can share some wisdom too.

>>62310838
How many of your banker colleagues are true sociopaths? When reading the news there's always that one investment banker behaving like real Pat Bateman. How many hours you work on normal week? 50+? Do you use Solaris any longer or is everything just bsd/linux system and if so, what brand you are using?

>>62310838

I'm a soon to be phd in comp sci. Most of my work to this point has been in networks and systems, but I have a strong background in AI and ML including deep learning.

What kind of work would there be for someone like me in an HFT/quant firm?

File: opal derp reaction.jpg (21KB, 187x223px) Image search: [Google]

21KB, 187x223px

>>62320354
>What's the dependent variable?

I'm the kid that just got into a Quant program, no where close to working, but how can you ask such a question? A) it's not just one dep. var., and B) how can he give away his company's trading secrets?

thanks guys, this was a pretty interesting thread.

>>62325529
I dunno this guy didn't really say anything that some guy perusing some subreddits for a day couldn't. I'm always skeptical about these threads. He could just be some LARPer.

I'm a CS/math double major thinking of going into grad school for applied math after I graduate next year.

Is HFT an option for me? Sounds like something interesting to do for a few years.

>>62326494
I notice this is a difference between me and other anons, a lot of times anons can't get past the OP for whatever reason OP puts them off and disregards the rest of the posts in the thread, while many times I get more out of the rest of the anons than the OP.

How do you feel about the fact that HFT is unethical and should be illegal? You literally suck money out of the system without providing anything of value and also contribute to unpredictability and overcomplexity that leads to things like flash crashes

i never really understood how the ticker goes up and down on any stock.. like who the fuck controls that and how

>>62326914
It's a auction. The rabbit hole goes much deeper than that but that's the extremely condensed summary.

>>62310860
>Most people are Chinese actually.

Isn't this true of any top-tier software job?

>>62326901
More transactions increases the liquidity of the market, allowing individual investors to easily sell and buy their stocks. Also rapid transactions bring help keep each commodity's price anchored to it's actual value.

Finally, HFT would really only negatively impact day traders. HFT has very little (none basically) impact on buy and hold investors, which is what most ordinary individual investors are doing, and should be doing anyway.

How do I get into HFT with a PhD in ECE from a decent state school? Most of my experience is in aerospace computing but there is so much more money in financial shit and you get to live in NYC

>>62327260
>state school

RIP. Maybe you can get a job as a barrista at starbucks if you're lucky.

>>62327362
I already have an awesome job lined up and don't even really want to work in HFT. I just thought I'd get your perspective out of curiosity. Don't be a dick. My school has a better ECE department than most Ivy Leagues anyway. Enjoy your money cause your work is not respectable to a real engineer. Don't bother replying cause I'm leaving the thread

>>62322543
You'd have an easier time finding a job at an HFT firm I think