Prop Trading vs. Software

Lebowski · Post by **Lebowski** » Sun Nov 01, 2015 5:29 am

Since posting this thread, I've been given an internship offer for software on a FX, Fixed Income, and Commodity trading desk at a U.S. bank. I will likely take it as it will allow me to stay in NYC and the pay is quite good as far as internships go. The big tech companies don't start recruiting til later (for the most part), so I'll likely roll with this. I may look at software on the West Coast next year if I dislike this job. Obviously, this is going to put a temporary halt on my trading on the side. I went to Chicago a couple weekends ago for a final round interview with an options market maker, but I was frankly pretty disenchanted by the experience. The pay isn't as good, the job security is terrible, and frankly I'm not convinced anyone there is making a significant amount of money besides maybe the founders. The company has 200 mouths to feed and, according to Google, their revenue was under 10 mil...the math on that is not pretty, obviously. Although it may be worrisome to TSWP (who, from what I've gathered, may have more professional experience than me, but is nonetheless attempting to start his own venture also [no offense intended, i wish you the best]), I don't view trading on the side as inherently a bad idea. I'm a pretty data-driven guy, I'm gonna look for situations where I have positive expectation and start extremely small. I don't see that as an inherently doomed approach, in fact, the tl;dr of my question really amounts to this: Suppose I'm a software developer making 100K+ a year. Let's say I start retail trading futures with maybe 50K after some data driven research. I retain my software job and have relatively good mobility. My losses are capped at 50K, and I still have a livable cash flow. My upside potential is extremely high as I don't have a boss, etc. I also likely have a lower probability of success than a prop trader (which is not that high based on what I've seen in my brief experience with prop trading). My question is, will I have a higher long term expected payout following the above approach than if I go the prop trading route and make 80K and "eat what I kill" for bonus in your opinions?

a路径积分 · Post by **a路径积分** » Sun Nov 01, 2015 5:42 am

> ...and if you're not expecting $150k+100% at age 22 (only Putnam Fellows can get that)

Do you mean 100% of base? Are you saying guaranteed bonus or performance bonus? We usually quote in % of PnL.

And no, this is an untrue statement. 150k is on the higher end but not unusual. Explanation: Top tier prop firms generally hire college graduates with 1-3 years' educational/experience edge over the typical IBDs. Since the job market is very efficient - it's usual to hear people turn down mid-tier shops like Akuna/Volant for GS - you expect top-tier prop firms to pay enough (risk-adjusted) to compare to GS on first to third years. This means 150k is run-of-the-mill.

I've seen very average/ordinary (by prop firm standards) people earn $150k base out of college. Example: If you've done solid undergrad research with one of the letters composing CLRS (i.e. exceptional resume but not Putnam-level), Teza, Citadel and KCG have a history of paying this amount.

I've also seen 2 orders more total comp by age 23-24 - no Putnam fellowship. Of course, these are extreme cases, granted some assistance from luck going into prop in the good years.

> I'll be able afford colo at Aurora, the MDP 3.0 protocol has vastly leveled the playing field in terms of data quality for retail/"semi retail" traders, and the TT API (if the exchange gateway is installed on the same server and you configure some stuff correctly) is as fast as a lot of DMA implementations.

> I think that's why you see all these tiny new prop companies springing up around Chicago...kids like me came in and spent five years writing software at Citadel, Jump, Tower, whatever...

From what I know, the number of prop firms that take off conditioned on 5 years of software experience at {Citadel, Jump, Tower} is practically zilch. I can name more firms that had a CTO at Tower + CTO at Getco in their starting team and still haven't taken off. The occasional guy who tries to monetize software skills + epchan/Q.SE literature is going to be fodder.

Lebowski · Post by **Lebowski** » Sun Nov 01, 2015 6:38 am

Chan's stuff is useless, some of it may work on equities, but he spends chapters on pairs trading stuff that most undergrads who know what PCA is could probably beat. I knew it was nonsense when I read the first few chapters of it before undergrad. The Akuna/Volant/etc. variant of prop firms was the type of aforementioned Chicago shop I was pretty unimpressed with. The pay wasn't on pace with the bank that offered me a job which you somehow divined the identity of. Akuna wasn't where I was, but my understanding from people I know who have interned there is there are worse places to be.

With regard to the individual HFTs being fodder, I know of at least three (two of whom members here) who have been pretty successful on CME as market makers as individuals. They're the exception rather than the rule, but I wouldn't say that it's hopeless, just exceedingly rare.

I'm not excited about being "fodder", but I think I'm better off being fodder as an individual with a steady job than as the employee of a prop shop where my job is on the line. I wouldn't try to go up against people with DMA with TT. Typical stacked book market making with 80-100 microsecond latency is just not gonna be viable when I'm 60+ microseconds behind the competition. I may attempt to do some market making in some capacity, but I definitely am not going to be trying to fight for queue position on ES against Teza, etc. That being said, I think you'd agree that the availability of cheap clearing, good data, and colo have increased greatly for retail traders. Again, I'm not saying that I expect to beat DMA traders at the speed game, but I'd hazard that I could get reasonably good at capturing the spread after a while, and if I don't, at least I'll have a day job that I enjoy if I go do software. I've read all your posts, @a___, so I know you do this professionally. May I bounce a couple ideas off you privately to assess the viability of them?

svisstack · Post by **svisstack** » Mon Nov 02, 2015 8:17 pm

>> Also, the "some automated trading on the side" bit scares me, it is not a good approach. If you want to be successful in this business you have to dedicate yourself to it, 24/7.

I wouldn't agree to that, have in mind that some people are more effective than others and value of that difference in effectiveness can be as high as 500% without raising eyebrows.

@Lebowski: I think your analysis about higher long term payout as software developer is correct. I'm on same path based on similar initial assumptions and it works good for me as far I am in, also it works better and better as time goes, so first steps are hard. Have in mind a lot of problems like lack of money on investment in long-term software and hardware infrastructure.

TSWP · Post by **TSWP** » Mon Nov 02, 2015 8:35 pm

My 2 cents: I speak from personal experience, in the last 9 years I have met A LOT of software programmers saying that they were going to make a bit of automated trading on the side. And in 100% of the cases their results have been long-term disappointing.

PS - when I say "24/7" I mean that it is your full time and only profession. Then the actual trading time may be 5 minutes a day, as long as one knows what he is doing.

svisstack · Post by **svisstack** » Mon Nov 02, 2015 8:47 pm

>> PS - when I say "24/7" I mean that it is your full time and only profession. Then the actual trading time may be 5 minutes a day, as long as one knows what he is doing.

I'm just trying to say that is possible that he commits 2 hours per day to this and be more effective than someone who is full time in this, bcs he will knows what he is doing, he certainly dont know at this point, I can conclude from existence of this thread. But I admire long term planning like this.

I think if someone has inspiration and motivation then it should try, otherwise if you are not trying then what the fuck are you doing exactly?

Also based on my personal experience I consider a lot of better to have 2 professions instead of committing 100% to one thing and not from diversification point of view.

But it depends on everyone personal view and its going more into offtopic imho.

a路径积分 · Post by **a路径积分** » Tue Nov 03, 2015 3:56 am

> I'm better off being fodder as an individual with a steady job than as the employee of a prop shop where my job is on the line.

I think you need to revisit your premises here and ask: Why would your job be any less on the line at Google vs Two Sigma? Or Twitter vs Akuna?

The median tenure at GOOG/AMZN/FB is around 1-2 years. Sounds just about as secure as starting out at SIG. Even a shop that's easier to get in, a la PEAK6, has 90+% profitable traders on a volatile year. A more stable location like MSFT is matched with Morgan Stanley.

My group rarely comes across people who were let go at their previous prop jobs. Instead, most often we come across people who leave because they have many doors open to them and are empowered with the skill/ability to do whatever they please. *Granted, this is a self-selecting sample since people who were fired probably often choose to leave the industry anyway.*

Lebowski · Post by **Lebowski** » Tue Nov 03, 2015 4:48 am

@a___Would you not think that a big part of that median tenure number is probably due to the West Coast culture of moving in and out of start ups?

That number you gave me about PEAK 6 is pretty reassuring actually. I thought the profitability and by extension the turnover would actually be much worse. I'm pretty much done with looking for an internship this year as I already have found some good opportunities, but I'll definitely reconsider prop trading next year. If I could get a prop job that was straight up HFT software development or research like what seems to be the case at the smaller reputable ones like Teza and Headlands, I'd take that in a heartbeat. Would it be fair to say that the day to day of a trader at an HFT shop like those is more research and hands on with market data than a lot of the other Chicago shops? Maybe this means trading isn't for me, but I really would only wanna be a prop trader if my job was more akin to data driven research than watching the screens or answering a bunch of phone calls from broker dealers. I really wanna be a "strategy developer" vs. a trader...that is if "strategy developer" or something akin to it is indeed a job that exists outside of my bedroom. I'd still be happy to watch my own algorithms throughout the day and participate in projects, etc that were for the good of the team though obviously. Maybe the job I'm describing doesn't exist or maybe it's for PhDs only. I really don't know...good thing I have a phorum full of experienced folks to ask. Although I sat with a number of traders on their desks over my trip to Chicago, I'm still a bit unclear on what they even do. They explained option market making and how they come up with quotes for the dealers, but what exactly is a trader at a Akuna/Volant/PEAK 6/Belvedere kind of shop actually doing throughout the day? Do they formulate opinions on input parameters to market making models? Just make sure nothing goes wrong? I really can't say I know at this point. Thanks for your help.

@svis Thanks for your encouraging reply. I think software would be pretty complimentary to trading, or at least the type of trading I would like to do.

@TSWP I respect that trading is a legitimate full time job and that it takes extreme dedication. Moreover, it takes a lot of dedication to learn. I've read a lot of books, but I'm not sure that suffices for the real thing. My internship this year is on an FX, Fixed Income, and Commodity trading desk at a large bank, so I hope to get some legitimate professional trading insight there while having an opportunity to improve as a software developer. I thought this would be a good balance of my interests that would leave a lot of doors open.

Rashomon · Post by **Rashomon** » Tue Nov 03, 2015 5:43 am

jslade, almost everyone must agree that picking a fun problem and killing it with a small, self-funded team, exiting for k•10^7 per founder within a few years is the spiritual and cultural ideal of the moment.

Also, having worked in VC I agree that founders should shy away, unless you meet an investor who knows your industry better than you. If you go to enthusiast angels it will be a distraction not worth the money. And if your business looks good on paper then all the finance type angels will pile on at once.

But. The management challenges to go from 1/2/3 to 10/20/30 emp are significant. And absent killer IP, how you going to exit at 100M without a significant sales team? Furthermore, if the founder is (as you seem to be) a technical problem-solver, that's going to be less fun than normal work.

Best response to this sub-thread would be someone who knows how to hire a good CEO and hire/manage a small team. (I mean venture/traditional business, not fund management)

a路径积分 · Post by **a路径积分** » Tue Nov 03, 2015 11:16 am

> Would you not think that a big part of that median tenure number is probably due to the West Coast culture of moving in and out of start ups?

And why would a career moving in and out of startups be any more/less stable than growing up as a PM in a prop shop?

I understand where you're coming from when you're saying that you would rather join a software company. I honestly think that your posts are a roundabout way of saying:

1) Software seems like a local optimum for you rather than a global optimum.

2) Columbia undergrad math is not a program with good 'admission' rates to the upper tier shops (we're talking about IMC instead of Akuna).

3) To a software company, you probably look like an Ivy League graduate with some blend of diverse software and math skills. You probably have some leadership qualities and a number of programming languages on your resume. You probably interned at bank where you are prouder to name the company than specify what team you're on. You also sound like you know or have experience working in a more entrepreneurial setting at a small software company. If I were to ask you to identify your skills as a subset of {math, CS, brand-name internships}, your honest response would be {math, CS}. Your average classmate's responses would be {{}, {math}, {CS}, {brand-name internships}}. This is the perfect fit for a data analyst position that pays better than to your average classmate. However you are not competitive in the pool of applicants to a firm like Jane St (as you stated in this thread), HRT or Two Sigma.

4) On the other hand, a top-tier prop firm recruiter just sees this in your resume:

> You probably have close to zero research experience - I mean zero to nearly zero citations on Google Scholar.

> You probably have no IMO/IPhO medal, ISEF or Putnam title.

> Your math/software skills are probably better described as 'broad' rather than 'deep'. This means:

>> Yang-Mills and Lie groups are out of your reach.

>> Your algorithms background is probably in some epsilon-ball of the basic sorting, data structures and graph traversal algos that they teach you in class.

>> You can compile a library, clone from github etc., but you don't know say the basics of tr/sort/sed/'' vs ""/{} off the top of your head - say, you don't know how to histogram a word corpus or emulate an associative array in bash.

> You haven't interned at a nontrivial subset of {GOOG, FB, DE Shaw, Jane St, KCG}.

This is where I get into career advice:

1) This is fine. Not everyone is born to fit Jane St and there is no indication that Jane St would lead you to a richer terminal point than Some West Coast, Inc.

2) For someone in your position: You either make a serious push to get better OR you take the best job opportunities available to you now (local optima) and hope to find some way to cross the activation energy to get into another region. There is no shortcut that NP can give you.

3) But please do yourself a favor and don't feed. Don't let some idiotic - and I rarely use this insult and I'm only reserving it for this instance because I'm talking about people who are ruining the life of someone who hasn't graduated from college - ISVs, bloggers or academic microstructure researchers convince you that solo automated trading is a viable thing you can do on the side. You're in for an astonishing defeat.

Let's say it works out the way you hope. Let's say you put a server in colo and finally make your first trade. Let's say you earn your first dollar irrespective of the long-term profitability. This is going to happen after several years. (And when that day happens promise me you're going to buy yourself a drink.)

Now, what did you spend those several years doing? Becoming very good at some obscure compiler hacks? If you're going to work full-time at GOOG/FB, the opportunity costs are huge. You could've been trying this automated trading thing on the side, or you could've been working on the next unicorn doing something seriously better for yourself and the world than scavenging a blog for ideas. Don't shortchange yourself.