|Jan8-13, 07:31 PM||#1|
6 months after dropping out of college - my tips
It's been six months since I dropped out of college to bootstrap my financial research startup. I had an old thread about this and received a lot of advice against it, then the thread got deleted. Some had come forward to PM me asking how I've been doing since, and I figured it would be useful to start a thread with advice on dropping out of college. This sounds bad - and I don't ever encourage someone to drop out of college - but realistically, I think it's better to share my experience rather than sweep it under the rug.
Some context for those who never saw the old thread: I took a leave of absence at the end of my sophomore year to start a trading firm. Before that, I was doing decently at my academics. I had several research accolades at the national level, represented my college at the Putnam and MCM competitions, and did well in a host of graduate classes. And I didn't leave because I was particularly disillusioned with physics. Quite the contrary, I love physics. I regularly visited conferences and still do.
However, I felt strongly against paying for my college education any further. I felt that I was quite foolishly paying for single courses that cost more than my entire high school education even after scholarship money. At some classes, I felt I was truly just paying for the certification rather than paying for the instruction. The learning opportunities out of college were just as abundant, and much more affordable. I felt in today's age where you can access quality course material either on professors' personal websites or online learning aggregators (like Coursera and edX), and when a private college education can set you several years in debt, I'll be part of the first wave of dropouts who form the norm rather than the exception. You can disagree with me. I weighed the opportunity cost of rushing to finish my degree one year early against taking the leave of absence, and decided for the latter.
I've become increasingly tight-lipped about what I do on a daily basis ("financial research"), to the point that I deleted my Facebook and LinkedIn accounts. But I will talk a bit. During this time, I had the good fortune to tackle problems at the frontiers of several disciplines. I sold most of my physics textbooks, but still refer to my mathematics textbooks frequently.
The NASDAQ ITCH feed alone takes up 7~8 GB of space per day. That's a lot of experimental data to play with. When the data set grows to such a magnitude, you have to deal with several nontrivial problems on the infrastructural side. You'd find ways to push the limitations of the software/hardware that you are working in. Examples: The automatic GC locking up your program for 5~10 ms every few minutes without your control. Handling events every 1, 10, 100 us. Reducing thread synchronization overhead. Bypassing the linux kernel. Estimating timespans and sychronizing clocks down to microsecond precision. Some of the problems persist in 10^3~10^5 dimensions, which requires a lot of computational firepower, if not programming ingenuity, to work around. I lead a brilliant team (we have more top-tier degrees split among us than number of people) and sometimes wonder what they're doing around with me.
I usually get roundabout questions regarding my financial stability. The furthest I'll go is that I pay about $3,500 of rent every month, excluding utilities and monthly expenses. And I'm fully independent from my parents now. I think this is acceptable for the first year out of college.
One of the first things I did was to set up a background app that monitored what I did every day. I generally work 90 hours per week, at least 60% of that time on an IDE/vim and churning out work. The remaining time reading references, doing research, and organizing the business end.
I absent myself from talking about my work inasmuch as possible, but the people on my team often receive snide remarks about how they left a scientific field for a profit-motivated field. The remarks fall under two categories: the regulatory kind, and the activist kind. Regarding regulation, I don't think that it should involve judging the ethical value of traders' orders, but rather ensure transparent markets. Regarding activists, I feel that there are many institutions we should dismantle before we can even get to the financial industry on a moral scale. Besides that, I genuinely believe that the work we do cannot generate any profit if not that we provide some kind of service, because it is extremely cost-prohibitive to set up the infrastructure.
1. If you feel like taking a leave of absence, try take a semester or two to cool off your head and do things differently to "prepare" yourself for the leave of absence. You might change your mind or strengthen your resolve at the end of it. I already had the notion but decided to give college a try for another semester. During that time, I took some classes just because I felt they would prepare me for the work after I dropped out - and it was a great decision. I even feel now that I should have waited one more semester before I left, because a lot of the time was spent learning things that I didn't know so I could bootstrap things myself - accounting, more esoteric fields of mathematics, programming in other languages etc.
2. STAY IN TOUCH WITH THE PROFESSORS WHO HAVE TAUGHT YOU! As many of them! I've come across several problems, in the least expected areas, where I wish I could just email one of them without feeling awkward.
3. Having a supportive family, significant other and group of friends makes the largest difference. It's a *huge* mental challenge to have a very successful partner and having him/her explain to his/her parents that you are planning to finish your degree at some indefinite time. If you think this is a problem and you have a choice to continue college, ask yourself if you can really bear to lose your girlfriend/boyfriend because of your career choice alone. If the answer is yes, then you can still do it. These people are the ones who will keep a check on you, instill a sense of discipline in you to do meaningful things despite the seemingly infinite time on your hands.
4. For most of the things you can drop out to do, nowadays, having programming skills is almost a must. I've come across several other dropouts and read many articles on this since then, and come to believe that this is true not just for myself. Knowing how to program today, I feel, is like knowing how to use a word processor 10~20 years ago - meaning, you have no excuse not to. I'm not speaking about what is necessary for doing a financial startup, but speaking in general - whether you want to do a mobile app startup (I'm against these), become a web designer, become an IT consultant, become a photographer, most of the careers that you *could* manage to do with only high school credentials seem to benefit from some kind of programming skill.
5. The most important personal skill that you need is self-discipline. This sounds like a trivial statement, but there was one occasion when I wasted a whole week playing a MMORPG (yes) and felt terrible for it afterwards - I got out of it because of point (3), I had a reason to work hard and my significant other helped to threaten me whenever I strayed off course.
6. Having been there, I always thought that majors like accounting, economics, management science, psychology, philosophy, history were for softies. I felt economists were silly guys whose only unit root of interest was 1. I even felt that chemistry and biology were even less deserving of being natural sciences. This sounds rude but I'm being absolutely honest with what I used to think - I was definitely not alone in this mindset and I still come across people with this mindset these days - through conferences or other networking events. I always thought that so long as you were good at physics and mathematics, you had this magical toolbox and could easily learn what others were doing in lesser time, so I would defer things involving them or even try to avoid them altogether. But if you are dropping out of physics/mathematics in your second or third year of college, trust me, the chances are that your immediate ability to earn money is just as good as the U.S. is at reducing its public deficit. I can't think of anything which you could do that entails even 49% of your time purely doing physics/mathematics. Not even if you are doing a startup that involves a technical invention. Maybe tutoring high school kids. If you want to avoid non-science activities: go to college and maybe grad school. So if you really think these non-science activities are less worthy of your time, you will be very unhappy with the unsheltered life outside of college.
7. Estimate how much money you have, or your parents are willing to support you with, and honestly ask how long it can last you while you don't have money coming in. Multiply the duration by 3~4, and ask yourself if you can last that long without turning in a profit.
8. Set high ambitions for yourself, if not higher than the ones you had while enrolled in college.
|Jan23-13, 06:04 AM||#2|
Thanks for making that thread. It's good to hear from you again. I always thought that you requested to delete the older thread, but I digress.
1) Are you against mobile apps, or startups dedicated to them?
2) What kind of academic background (i.e, math/physics/cs/other subjects) did you have when you started the start up, and what about now? I'd understand if you can't be specific about what you know now. With the nature of your work being what it is, that's understandable.
Essentially, what, in your opinion, are the prerequisites for one to get his hands dirty with the finance stuff?
3) If you can afford a house/an apartment at that price, then I'd say you're doing pretty well. Probably better than many college grads. Good work!
|Jan23-13, 09:39 AM||#3|
1) I am against startups dedicated to mobile and web apps. Firstly, because it encourages mediocrity and a different kind of gambling. It used to be that your first grand milestone as a startup is when you have to file your first tax return. Most mobile and web apps view their first milestone as when they hit the submit button on the Apple app store or when you get seed funding. It's easy to pick up LAMP, Android SDK, Xcode development stacks nowadays, so the mobile app startup culture has really encouraged people with limited skills and work experience to dig for gold, when what they're really doing is analogous to launching a pizzeria and hope to become the next Pizza Hut or Domino's. The largest percentage of failed startups are in this area, for good reason.
Secondly, because I think that there are so many areas in this world and even our own societies that are worthy of our attention besides your internet experience. So long as there is a city without affordable housing, public transportation that doesn't arrive on schedule, power grids that are inefficient, a person picking up the aluminium can in the trash after you, a decline in population and growing mean age in a suburban region, there is so much more to be done for the world than to develop a mobile app.
2) Most of my classes were standard college mathematics:
- Multivariable calculus
- Real analysis (Loomis and Sternberg) up to bits of measure theory (Kolmogorov, Taylor)
- Complex variables (Brown and Churchill)
- Calculus of variations (Gelfand and Fomin)
- Abstract algebra up to rings and fields (Artin, Gallian, Dummit-Foote)
- Numerical analysis (Burden)
- PDEs (Strauss, Evans)
- Differential forms (Flanders)
- Statistics (Casella)
- Topology (god that was horrible)
- Nonlinear optimization (forgot the name)
- Differential geometry (one of the Spivak volumes and Kreyszig)
- Statistical physics
- Fluid mechanics (my research area)
- Classical mechanics (Morin, Taylor and G-P-S)
I'll be honest, I think it's a lot of work and I didn't have the talent to become an expert at all of them: if you gave me a problem set in any of the above now, I'd still struggle with it. One of the things I did early on was to acknowledge the things I was bad at, and then choose to steer the business in the direction of what I was comfortable with. I'm very good with what I'd call a "mathematical dependency" thing, where I read something and it's in a subject I know absolutely nothing about - I'll work backwards all the way to the fundamentals. Because of this, whenever I get thrown with a problem where I *must* solve it using a tool I'm unfamiliar with (e.g. some esoteric area of signal processing, some programming language you've never used), I just get my hands dirty and pick it up.
I still refer to my textbooks, so I've kept some of them:
As for the things I had to learn from scratch after leaving university, econometrics, convex optimization, further time series analysis and state space models are among them. But the alpha in them has been mined out, so these are just prerequisites. I learned a *lot* of programming. Algorithms and compiler theory are especially important, but that's the bare minimum. The bottleneck nowadays are in memory access and optimizations based on hardware architecture, OR parallel computing. Chances are though, depending on the type of work you do, you only need to be an expert in one of the two.
I should explain what I mean by the alpha being mined out: One of the things I've figured out is that there's a difference between risk and uncertainty. (Just as there's a difference between a quant and an actuary.) The former entails randomness whose model parameters you are uncertain; the latter entails randomness whose models you are uncertain. People don't pay a lot to hedge their risk, but few know how to pay the correct amounts to deal with uncertainty. The difference between the valuations by these two groups of people is the amount of inefficiency (alpha) you can capture and pocket. That's why I'm not afraid of people who have 20-30 years of experience wiring low-latency server infrastructure and IT backend, because I've found that these people do such work because they are only interested in solving problems of the known kind.
So the topics that are really useful are the ones taught in "fringe classes", you know, the ones you see on a graduate catalog with names that you just know will attract a small class enrollment.
3) Thank you. I still have a long way to go, and I don't know if I'll blow up today, or tomorrow.
Of course, take my words with a pinch of salt, I'm certainly not the most-qualified person to refer to on these matters.
|Jan23-13, 11:01 AM||#4|
6 months after dropping out of college - my tips
1) Cool. I also don't like the way in which the social lives of people have spread onto the internet. More places than I'd like feel like the Misc section of the Bodybuilding forums, or 4chan. Some might say that if I don't like it, I might as well go elsewhere, and that's what I do. I just don't like it, that's all. Maybe it's because of my perpetual romanticism of a decade I missed (the 90s).
2) Good god! I could be wrong, but do these courses not amount to a BS degree in math? Or did you just "look like a math major" when you were really a physics major who was curious about math? :-)
Econometrics. Did you have to learn micro/macro as well? While at college, did you not take any liberal arts courses then? I read the OP, but I figured I'd know for sure if I asked.
"I'm very good with what I'd call a "mathematical dependency" thing, where I read something and it's in a subject I know absolutely nothing about - I'll work backwards all the way to the fundamentals. Because of this, whenever I get thrown with a problem where I *must* solve it using a tool I'm unfamiliar with (e.g. some esoteric area of signal processing, some programming language you've never used), I just get my hands dirty and pick it up."
This sounds like fun. Do you have any issues with concentration, by any chance?
|Jan23-13, 12:21 PM||#5|
1) Lol, I'm young enough to know what the misc section of the Bodybuilding forum and 4chan are.
2) Yes, I was a physics major through-and-through. IIRC, my TA for multivariable calculus in freshman year was also my classmate in differential geometry, and later, the graduate statistics course. That was awkward. My high school had a pretty large library budget, so there were graduate-level textbooks in a cozy corner of the physics section that no one used. I assumed the workload must have been much tougher 20~30 years back, because that was the last cluster of loan dates that I saw on them. I spent my time browsing through them and knew early on that I was not going to be a good physicist unless I worked on my math, so I bought the Courant/Spivak/Artin books in high school and started to work on it.
The diff geometry class was actually extremely challenging for my first semester, but the best lesson I learned from it was an exam trick (that I don't encourage): Professors are very aware of the type of problems you can solve in 1-3 hours. Since some of the proofs at the grad level are near-impossible to conjure out of your own ability in 3 hours, the types of questions that they can grill you on are pretty limited. I assume this trick doesn't extend to defending your PhD thesis, though.
I took classes in micro and macro, and admittedly wasn't very good at it. I did linguistic philosophy and a course in writing - and was good with those, somehow. But I didn't have the strength of heart to finish all of the arts prerequisites, and that formed part of the reason that I dropped out. I argue that these classes are very important though. The interesting things to do are cross-disciplinary, e.g. I believe tax planning is an optimization problem, and I wish I knew enough accounting so I can apply what I know onto it.
3) Indeed! I have a poor attention span. But I've acquired various productivity tricks to make up for it.
|Jan23-13, 05:39 PM||#6|
You know, throughout college I always believed I would have learned more both efficiently and faster on my own, however points 3 (mostly) and 7 made this an unrealistic goal. One think I noticed you didn't add to your list which I think also encompasses point 3 is positive mental well being. Looking back, this would have changed my life dramatically.
I suppose I am in a similar position to what you were in your sophomore year - I left my first semester of graduate school when my quality of life when down hill drastically. As it stands I am in limbo, deciding what to do. A job would be great so I could have my own money, then work on what I want. The though of spending 8 hours a day 5 days a week barely challenged makes me sick to my stomach. Grad school would be nice because I really love aerodynamics and fluid mechanics and, of course, learning. I am worried though that I will not be able to sustain myself on a crappy stipend - to me it is not worth it unless I get my dream project...
Ehh. its quite stressful.
Overall I think you really know what you are talking about, nice post. It's good to see people sharing their experiences. If it was done more often I believe people in general would be in a better place.
|Jan25-13, 02:59 PM||#7|
I don't know what to suggest, but hope you've found something to do. I think fluid dynamics is amazing. I worked on microfluidics and two-phase flows before leaving college, and the applications are tremendous (e.g. the new smartphone technology that pops actual keys out of your screen when activated, treating stroke less invasively). I love F1 because it shows how with enough budget, the sky is the limit with research in things as seemingly simple as diffuser design and every year they have to revise the rules drastically or the cars are only going to get that much faster. I found it a little misfortunate that the field doesn't have a fixed place in any department; sometimes it's in an aerospace engineering program, other times it's in a mechanical engineering program, an engineering physics program, or an engineering science program. It's more unfortunate that people who brand things "rocket science" don't actually know what it is. This seems somehow to diminish its value on the surface.
|Jan26-13, 01:42 PM||#8|
I think I remember your thread a while back. Why did it get removed?
I'd like to ask some questions:
1. How did you get into the trading game? Did you just happen to pick up a book and start teaching yourself programming or something?
2. I come from a poor family, so I feel really selfish if I go into academia when I could be helping my mother out. As such I've also decided to go into finance. I have a week work experience lined up at an investment bank, but I'm not really sure what to make of it. I'd rather be doing what you do, and start up something, rather than work for someone else. Do you have any experience with internships?
3. As you said you didn't get any help from your parents, how did you manage to work it out? Did you save during your undergrad years by living at home or something?
|Jan26-13, 08:12 PM||#9|
1. Mods figured it was turning into a blog.
I attended a mathematics conference in early freshman year and one of the lectures was about "statistical arbitrage". After that first exposure, I did some reading and started to do the typical "simulated portfolio" thing - and I outperformed everyone on a consistent basis - I thought I was going to be rich! Turns out the real thing was going to be much, much harder than I expected, and much, much different from playing with simulated portfolios.
I was a little fortunate, as I soon made a friend who was on a sabbatical from his hedge fund (he had a successful streak, returning about 53% per year and returned all of his clients' money to go on his sabbatical). I was good (enough) at math and he was good at discretionary trading, so we complemented each other. I learned a lot from him. He was not afraid to actually open large positions right in front of me just to show the mechanics and gave me a few tries occasionally, funded a small account for me at first, and let me take the profits while he sustained the losses. It was just to get me acquainted with the "psychology" of trading. He stayed over at my place and we'd be up till 8 AM every morning actually trading the European session and FX (which is 24/7) - and I'd get too tired to attend most of my classes. The feeling that you actually skipped class and could make enough in the matter of seconds to let you live comfortably for the rest of the day was indeed extremely alluring. I wouldn't deny that. I forgot all about statistical arbitrage and just did this brainless discretionary trading thing.
But we later had to part ways and I learned that it really takes more than experience, psychological strength and discipline to do it on your own capital. And without adequate funding, making enough to feed yourself each day is a daytrader's (gambler's) delusion. I went around a full circle and went back to simulated portfolios, except this time with a very different approach, bent on learning how to do it properly.
Still, that I actually had a mentor who was willing to help was a huge influence. I'm estimating 90% of the people fail simply because they've never found a mentor who genuinely had their interests in mind.
2. I don't, but I'm a little informed about internships because of the team I'm working with. A week (an externship?) is a good step in the door as far as I know. Since it's already lined up for you, I'd just say make the best of it. As for your future internships, it seems that the general consensus is to find one that really focuses on their training program for your first real internship (full summer). I can't possibly know them all - guess WSO/Wilmott are better places for some color on the quality of specific internship programs. There are many internships at bulge bracket banks that just make their interns do coffee runs.
I wouldn't do it for the pay, as most of them involve long hours and are located in cities where you'd be poor while residing there anyway. SV pays well and has a culture of making its interns actually do things. Companies like Dropbox have a LOT of extra capital at the moment. Being able to work with big data is also an advantage if you want to go into finance later. Worth thinking about if money is a primary objective.
If I had the choice to work at an investment bank or prop firm with minimal effort selling myself, I would. However, because of the experience and knowledge that I accrued in a very short amount of time (again, I was lucky), the opportunity cost of spending time to compete for a competitive internship weighed in favor of starting out on my own.
If you've never actually felt like puking for something non-medical - but simply because you lost ridiculous amounts of money in a volatile environment; or if you don't know every order type at the back of your hand, then it's a bad idea to consider trading your own money as an option for the next 1~2 years. Because then you're just speculating.
3. I had financial aid but had a small amount of money set aside for my personal expenditure during college. I worked up from that.
|Mar2-13, 02:47 PM||#10|
do you have a blog by any chance?
|Mar3-13, 08:50 AM||#11|
|Mar5-13, 06:46 PM||#12|
Anyone who has tried trading for more than a few months of time will soon realize that this is a very competitive field and that you cannot carve a profit unless (1) you provide liquidity or (2) you produce more informative prices. To get to either (1) or (2), you have to produce some kind of service for other market participants, either through economies of scale (making a market), being faster (developing better technology), having better models (employing more brilliant people) or so on.
There is plenty of utility that I'm providing my employees, clients and other market participants - these include retirees, farmers, real estate developers, policy makers, universities, whom I hope can go on to generate other kinds of public consumption that raises the GDP. However, I'm not about to become a retiree myself. There are many activities and institutions that we need to get rid off from this society before we even get to hedge funds; I'd start with alcohol, tobacco and marijuana producers, lotteries, weapons manufacturers.
If money was everything that I'm concerned with, I'd join a mobile/web app startup:
(1) It's much easier to convince and get a VC to fund you with $2 million for some web app than to go through the legal hoops and loops to get a measly $2 million in AUM (which by the way, is only about 1/5 of the amount of AUM you need to pay off 1 experienced employee).
(2) I can hack the Linux kernel, develop scalable apps that write/query 16 million data entries in a second, program in a variety of languages (object-oriented, concurrent/functional etc.), answer quite a few probability/statistics brainteasers, all of which are more appropriate skills for passing an interview with a SV startup and developing the next Instagram than working on econometric models at 3 AM in the morning.
|Mar5-13, 09:49 PM||#13|
Had to interject here:
|Mar6-13, 01:15 AM||#14|
Blog Entries: 2
I think what xk was really hinting at is that your story seems a bit to good to be true. I personally have no vested interest in finding out if it is or not. Nevertheless, I do feel the need to say, that I do find it hard to believe that someone within a years time frame could learn enough to do well in financial research. Who knows, maybe you did learn Girsanov theorem or the Levy Process within that year, despite the fact none of the books on your shelves would indicate that you have books that would prepare you for the prereqs for such topics.
I did financial research before my current job, and what I read from you is a lot of surface information easily gathered from reading secondhand accounts. I guess my real purpose for writing here is just to ask, Why would you bother to come back here just to tell is how well you're doing?
|Mar6-13, 03:39 AM||#15|
@MarneMath: Fair enough. I think one reason for your disconnect in understanding the motivations for my thread is that I actually had an older thread on this whereby I made it clear that my hopes were much bleaker, and I received mixed advice and varying levels of good wishes and skepticism. The thread got deleted - I just wanted to come back to follow-up and share my advice.
The goal of my thread is not to suggest finance as a viable career path, nor to exhibit what I know or do not know, nor to list all of the books/journal articles that I've read or are in my possession, so I don't see why you were expecting me to post these in detail. There's reason for me to be obscure about what I've read or studied, or crop off the bottom half of my bookshelf for that matter, and this is the nature of the business. If I said that I spent 6 months intensively studying Geller-Kapernikov embedding AND run an investment management firm, then Google is going to get 1000 hits on Geller-Kapernikov even though I made that name up.
However, since you've brought up the topic of what realms of knowledge are useful: I want to make it clear that sell-side valuation is very different, and most of the traditional MFE material is not directly useful in what I'm doing. I'm familiar with the Girsanov theorem and Lévy processes, but I've had to apply neither in my day-to-day activities. On the same thread, I am very unfamiliar with replicating exotic options or swaps, and I've had to use neither; I might know a thing or two about replicating a deferred annuity bond from TIPS, but that's about it.
|Similar Threads for: 6 months after dropping out of college - my tips|
|Tips and advice for a young college student majoring in (astro)physics?||Academic Guidance||2|
|Tips before college||Academic Guidance||0|
|Dropping out of college to pursue career||Career Guidance||21|
|The teacher is picking favorites.. I'm seriously thinking of dropping out of college||General Discussion||58|
|Dropping voltage without dropping current||Electrical Engineering||2|