What are the challenges of starting a financial company as a college dropout?

[meanrev]

I decided to keep a topic to talk about my life as a college dropout, starting last week, for my own introspection, as well as to yield some insights/deterrences for others.

Just some background: I was an above-average physics freshman, not the best but also not the worst. I was pretty ambitious in freshman year, and had 11 classes one semester, doing well in various senior/graduate classes. But the problem was already stirring at this stage - my education was rather unstructured, and I was partly rebellious, while my advisers never knew my name. I took classes in quantum chemistry, fluid mechanics, statistical mechanics, differential geometry etc., and I never cared about my graduation credits.

At the same time, I had some involvement in extracurriculars. I was selected to represent our college at the Putnam Competition twice, I attended the International Mathematics Contest in Bulgaria, and I had an AIP award and publications under my name by the end of my first year. I became president of our SPS by sophomore year.

Halfway through sophomore year, I realized that quite a few of my classes wouldn't help my graudation requirements - foolish, really - and my expected graduation date would be stalled if I had wanted my double degree. I turned around and churned at humanities courses, PE requirements and so on and hoped I would find some way around my graduation. A problem never solves with time; and I soon changed my mind and decided I should perhaps still graduate early, but give up my physics degree in favor of mathematics. But graduating with a physics degree has always been my high school dream.

At this point, I was just wrestling with the frequently-encountered college mistakes. Nothing unusual. But a series of events changed my views completely:

- My mother lost her first job, my father became terminally ill, and we got poorer.
- The APS petitions against budget sequestrations, the Bell Labs shutdown etc.

My research adviser, a young but ambitious guy himself, decided on his own to market our research to the military for $20,000, and I became very disillusioned with the purpose of our work. Simultaneously, I happened to be at New Orleans for the Joint Mathematics Meeting in 2011, where I first learned of "statistical arbitrage" at George Papanicolaou Gibbs lecture. The enigmatic story of astrophysicists Nunzio Tartaglia whatnot captured me, and I began an aggressive reading of quantitative finance and so on. I remember the part in his address where he joked that there aren't many seminal papers in quantitative finance nowadays, because "When you find out how it works, you just want to go out there and do it."

It turns out to be true. Coincidentally, I made a close friend while I was traveling inbetween my research stint. I went on and on about my views about mathematics and arbitrage - before I found out that he was a successful hedge fund manager. I learned a lot from him. He lived with me as I was doing my research, and at one stage, we were making more money trading at 2-3 AM than my research adviser was trying (futilely) to get funding from the military. He spent months writing proposals. We spent minutes getting there. The experience was surreal. I won't talk much about the ethical issues - there are people out there who think that what we do is immoral - but I'd say that monetary reward has never been my priority, even all through the financial difficulties from my family situation. Instead, I've always valued my autonomy most in my career, and it just happened that I was able to pursue my own interests more easily in this path. Besides, there are only a few ways to move capital between those who hold it and don't - and without our contributions to market efficiency, debt will be expensive and equity will be scarce. Capital markets are there for a farmer to protect his crops, for insurers to be insured themselves, and for airlines to be protected against fuel costs, and I don't see it immoral.

As (<- edit: typo) this went on, I finally decided that I had a good shot at this at one point. My girlfriend, who was working at Goldman Sachs' asset management division at that time, was the first one to convince me that I won't be happy working for a financial company nor in research. I eventually convinced my parents that I want to take a shot at this, or fail trying. I adjusted my classes so I would have time to focus on this, took classes that the MFE students were attending, dropped my research and slept less than anyone for the entire semester.

I had my reservations though - I felt that the networking opportunities of an internship were valuable, too, and that it would buy me more time to become confident and more prepared to set out on my own, but a chain of rejections made it clear to me that I would have as much chance trying on my own. So I made up my mind to start my financial company by the end of this year, beginning this summer.

I weighed my opportunity costs, asked around for advice, and finally decided that it was naive of me to choose a college that cost all of my parents' life savings and more. I take about $96.3m of trades per month (makes sense: leverage is huge), and this was going to be the biggest trade I would take: to pool all of my education costs into my startup.

I promised my mother that I would first see to it that she can retire at ease, and that I would one day go back to college to finish my degree, then a physics PhD. I fear success as much as failure - because if I succeed, I would probably have too much on my hands to go back. They agreed that I would struggle to find the money to finish my double degree unless I had concentrated on my work full-time.

I tried to be a regular college student for the last 2 weeks of my time there. I did homework, attended classes, got home at 9 PM, and would spend all that time until 2 AM (when the Eurex opens) reading on quantitative finance, trading, business, whatever I can find. I backtested for hundreds of hours, programmed several strategies, even if it required learning 3-4 different programming languages in short shrift.

Day 1: I went back to college to pack up the things that I had left behind at my friends' apartment and clear my mailbox. On my way, I made a transcript request - a copy might come to use - and picked up a leave of absence form. I spent the rest of the evening with my girlfriend, watching some films.

Day 2: The next day, I listed what I considered my talents, personality traits, marketable skills/completed projects, and started to think about how to find a way to penetrate a market that has incredibly high barriers to entry. I still haven't found the perfect area to start up in, but I have to pursue several things at the same time for my self-improvement. I made some substantial breakthroughs in my trading strategy, and hoped to trade the Friday open - though it wasn't the right setting. Then I wrote some 300-400 lines of code for a nonlinear optimization problem in portfolio selection - I decided to diversify into equities, as well - and filed for my equities trading account in the meantime.

I talked to one of my ex-roommates at Harvard, showed him some results, and he decided to throw in $30,000 for our fund. We decided that he would have involvement in the trading itself, so I would take a 10% performance fee - that's smaller than the industry standard. 10% isn't a lot - I would take home a pittance from his investment, but having an extra $30K under management is a good start. I figured that the best setup is to open a third brokerage account, colocate my algorithm to his apartment at London, while I booked a flight to Tokyo for a mix of reasons: I could more easily carry out discretionary trading without being up at 2 AM whatnot, and I needed a small break from everything. I picked up a book to learn Japanese.

I told my friend that I am now barely making enough to pay for my rental (Cambridge is an expensive place to live in) and day-to-day living without much growth in my operations - so I need to figure out a way to earn more, and fast. I plan to raise enough to hire an accountant and incorporate, hopefully in less than 6 months. I reiterate that I still haven't figured out exactly what I am doing, but it doesn't hurt to be doing some small trading activities while I am at it.

I wrote my first goal in large font: do NOT become a daytrader. I always view daytrading as a job that doesn't require much ingenuity, and that doesn't run well with my physics background.

 

[Devils]

Man you and I share the same interest in making money, but you seem to have gone completely off track. Go back and finish degrees if you can, any science degree is generally useful in getting jobs.

As for quantum finance & program trading, its a high risk/high reward game when if you have enough capital. The trouble is you have nothing to fall back on if you plan doesn't work - in other words your personal risk management is wrong. Sounds like soon you will either be rich or bankrupt & homeless.

I've been in your shoes, I have a strong maths and programming background, but limited business sense, and lost $300K which virtually destroyed me mentally. Now I run businesses on a more conservative basis, buying business for 25% return on equity using Buffett-style analysis with big margins for error built in.

You are sounding a bit depressed and desperate, things that force you into irrational decisions just like I did. Think seriously about getting professional help & advice.As an aside, I suggest you listen to the entire Warren Buffett MBA graduation lecture. Here's a quote that always sticks with me.

Link http//www.tilsonfunds.com/BuffettUofFloridaspeech.pdf

I was going to do the same things when I had a little bit of money as when I had a lot of money. If you think of the difference between me and you, we wear the same clothes basically (SunTrust gives me mine), we eat similar food—we all go to McDonald’s or better yet, Dairy Queen, and we live in a house that is warm in winter and cool in summer. We watch the Nebraska (football) game on big screen TV. You see it the same way I see it. We do everything the same—our lives are not that different. The only thing we do is we travel differently. What can I do that you can’t do?

I get to work in a job that I love, but I have always worked at a job that I loved. I loved it just as much when I thought it was a big deal to make $1,000. I urge you to work in jobs that you love. I think you are out of your mind if you keep taking jobs that you don’t like because you think it will look good on your resume. I was with a fellow at Harvard the other day who was taking me over to talk. He was 28 and he was telling me all that he had done in life, which was terrific. And then I said, “What will you do next?” “Well,” he said, “Maybe after I get my MBA I will go to work for a consulting firm because it will look good on my resume.” I said, “Look, you are 28 and you have been doing all these things, you have a resume 10 times than anybody I have ever seen. Isn’t that a little like saving up sex for your old age?

When you get out of here take a job you love, not a job you think will look good on your resume. You ought to find something you like.

If you think you will be happier getting 2x instead of 1x, you are probably making a
mistake. You will get in trouble if you think making 10x or 20x will make you happier
because then you will borrow money when you shouldn’t or cut corners on things. It just doesn’t make sense and you won’t like it when you look back.

 

[meanrev]

I thank you for your kind advice. I really appreciate it. I'll elaborate on my decision to drop out and see if you still believe I have gone off track.

The (predicted) maximum drawdown on my portfolio requires me to hold a certain amount of cash equity for the strategy itself to recover the losses that can and will come my way. If I go back to college at the end of summer, my living expenses plus tuition will exceed my possible drawdown and make trading implausible for me, shutting out my only form of income (trading) during this time. Yes, I will still have enough to meet margin requirements and trade, but it makes no sense to hold insufficient capital to stage a recovery, so I *will* have to close my brokerage accounts. Not to mention, this is a full-time effort; I cannot be in school while I trade, I will fail at one of them. Assume I close my accounts. Given the rate of my living expenses, my family's savings AND my previous trading capital will go negative in my final year of college, before I can graduate. I have explored the possibilities of loans, lottery etc., but nothing works.

The only redeeming possibility would have been for me to find a job that would yield equal or higher rates of returns as my trading income (if not higher) during the summer, so that I have a chance at mitigating the large one-time expenses of tuition at the end of summer. Unfortunately, I didn't manage to find such a job. Going back to college if I cannot afford college, at this point, is like buying a one-way ticket if I cannot afford the return trip. I've seen people playing the guitar in the Boston subways, holding a sign that reads, "Penniless traveler". Now, certainly, I can hope to survive as a penniless traveler through my remaining college journey, but I am the type who would rather not travel if I cannot afford it.

Desperate, yes. But I am of very clear state of mind to accept the fact that I cannot afford college now. In fact, I wrote my last will and testament and will get it notarized later this week. Because now I am holding a substantial amount of money (not enough to see me to the end of college, nonetheless) in my brokerage accounts, and it makes sense to give my parents legal defense in their favor should I die in an accident.

It has also always been my character to make my own free decisions, and I would probably have done a startup after college, had I graduated.

I don't like making money. Most people have green and red graphs on their trading screens; my candlestick charts are all white. If I had infinite money, I will burn most of it - because inflation is bad for the economy. Then I will finish my degrees.

****
I just saw your new edits: I am sorry to hear about your experience. It is harrowing and does strike fear in me. I will tread with caution.

 

[Mépris]

Call me old fashioned but I think dropping out looks like a bad idea. "What happens to your money if you die?" is a reasonable question to ask but what about "What will I do if my trading endeaours fail?"?

Really, I wish you the best of luck and I hope you (keep) succeed(ing) in what you're doing but have you considered asking your school for financial aid? At most schools who do offer aid - my understanding is that most liberal arts colleges do -, if one's financial situation changes, their aid packages are adjusted accordingly. The caveat is that one must've been receiving financial aid since freshman year. (although some do allow one to apply at a later point)

Best of luck!

 

[meanrev]

Mépris - I explored financial aid options but nothing is available to me, unfortunately.

It's true that I need to have a backup plan. I don't intend for trading to be the centerpiece of my work, but haven't yet found the right plan given the little amount of time I've had so far. Thanks a lot for your wishes, I'm most grateful. I'll keep you guys updated on this thread!

 

[chill_factor]

obviously, not all dropouts are going to make millions of dollars. only 23% of people in the US have a bachelor's degree yet 70% go to college. We don't have 47% of the population be millionaires.

if you can drop out and be successful, you'll be successful no matter what since you were just born with it.

 

[meanrev]

@chill_factor: That's very true, and that's precisely the purpose of setting up this thread: we often hear about the very few successful dropouts, but we forget the vast majority of dropouts who are expected to make less than graduates that we never hear about. I'm not fooled by the confirmation bias. This is a thread for people to know what they're getting into.

I spent the last 5 days doing independent reading that's no different from what I'd be doing in a summer internship. Had 2 hours of sleep on Sunday shuffling my equities portfolio and building a spreadsheet for future weeks, analyzed the fundamentals on about 30 securities then went into about half of them. Subsequent days yielded 5, 4, 0 hours of sleeps respectively.

Made use of the code that I've written previously (see above post) from my project in my nonlinear optimization class to construct a pretty strong portfolio, up 1% on Monday yet down only 0.5~% over the next two days, outperforming the market in general.

On the other hand, my trading results are attracting some attention from other small investors, I had to decide on a performance/management fee yesterday. I now have several accounts - data, brokerages etc. and I lent out one to my partner/ex-roommate.

At this moment I've liquidated my US equities portfolio, exactly as the S&P peaked the third time around 10.08 AM. I will spend the rest of the day preparing a presentation and report for my girlfriend's final project on her quantitative finance class ("prepare an algorithmic trading strategy" - easy, as compared to a "profitable" trading strategy).

I have hundreds of hours of historical tests for different situations - and with some luck on my side, I guessed the movements correctly on the DAX for consecutive times, an unbroken streak for this week. Overall, I concluded the week spending (lavish lunch treat) much more than I earned. But I sleep well tonight.

 

[Devils]

I've played around with nonlinear optimization strategies (like simulated annealing) for finance as well. Its fun and may lead to money-making strategies, maybe not. Long Term Capital Management had some great ideas too http://en.wikipedia.org/wiki/LTCM#1998_bailout

What you are doing sounds awfully like day trading, or very short term trading. I bet you enjoy the excitement, but that says more about your personality than what's best for your future. Whether your trading makes money in the long term is another thing, the odds are not in your favour.Fro reading I'd recommend looking at Taleb's book http://en.wikipedia.org/wiki/The_Black_Swan_(Taleb_book)
or stock arbitrage ( you I know arbitrage would be boring for you)
https://www.amazon.com/dp/1439198829/?tag=pfamazon01-20

 

[meanrev]

Devils - thanks for the recommendations. I've studied LTCM in the quant finance class at my college, but I've not read Taleb's book. Is there a seminal paper on simulated annealing that you could point me to?

Definitely, the odds are not in my favor (it is said that some 75-85% of traders are losing money, and 95+% of newly opened margin accounts go bankrupt or close within a year), and I mentioned that my goal is to become !=(a day trader). I like excitement, but I've always been the unluckiest and therefore most risk-adverse person in my life. I even pick the airline to fly according to safety records.

So why this path?

Clearly, there has to be a chance for profit for me to decide on this. I need to clarify my stance on this.

1. Capacity. Most institutional traders have enough capital to "self-destruct" their positions from the liquidity risk alone. Well, with some leverage, even I have enough capital to make my order visible to people across the world, on some illiquid futures and low capitalization stocks, and just placing that order (even if it isn't executed) can change things. For this reason, the high-AUM traders are imposed the heaviest constraint - they can only trade "high capacity strategies", many of which promise fewer profits precisely because of a market saturated with buy-side firms with the same constraints, and it makes sense: it costs more resources and incurs much more operational risk if they were to instead engage in many "low capacity strategies", even if the latter promised more upside. There are other constraints (typically coming from the management) that I don't need to satisfy at this stage. Since we're both familiar with optimization, and you clearly more than I do, we should find it clear that a constraint in such an optimization problem reduces the objective maximum. I find it to my advantage that I can trade low-capacity strategies with fewer constraints.

2. The lines are easily blurred; a daytrader can be called a proprietary trader if he only trades his own money (and especially if that money happens to be a big sum), or if passed his Series 7 examination and has an account with a "proprietary trading firm", instead of a retail brokerage account, with access to professional leverage. An institutional trader is technically called a daytrader if he opens and closes the same position within the same session, yet we rarely call them that. A "quant trader" might also be a glorified programmer. While I have pointed out that my main goal is to find success mutually exclusive of daytrading, I have several prejudices against terming my current/prospective activities as "daytrading":

- I am managing other people's money, actively finding more client funding, and offering an asset management service. (Well, note that hedge funds aren't allowed to solicit for funding, but I don't have 20m of capital to be held to such a restriction, nor 1m of capital to be called an accredited investor, which is why you don't see such a criterion used to determine if you're a daytrader or not.)
- I have a certain growth strategy, even for my trading alone - and because my focus is on algorithmic trading, it doesn't just involve holding more positions, but chiefly involves improving infrastructure (co-location, better hardware) and hiring as many people as it takes (accountants, legal compliance, risk management, programmers). Now, the typical daytrader reinvests from holding more positions alone - there's nothing wrong with that, you can make millions a year from this - but for personality reasons, I'd much rather make less, but offer services rather than make millions in a fancy home. I think the ratio of planned allocation of reinvestments in infrastructure to position size makes a difference.
- Daytraders have a tendency to advocate a load of crackpot theory, and coming from a background where I've seen *really* good crackpots, I think I can evade such a future. My work involves a significant amount of modeling and programming, which does not fully disqualify me from being a crackpot, but does give me some chance of not being one.

To end it on a relevant note to the rest of the readers, it makes little sense to quit a physics program in college to do a menial task with minimum wage, unless there exist exceptional reasons for potential upside. From this point of view, I've considered the limited paths to success that laid ahead had I left college, and opted for this.

 

[twofish-quant]

I told my friend that I am now barely making enough to pay for my rental (Cambridge is an expensive place to live in) and day-to-day living without much growth in my operations - so I need to figure out a way to earn more, and fast

The fact that you are making anything at all is truly outstanding. Congratulations. The only piece of advice I have is to be careful how you expand. If it turns out that you are hitting the limits of your trading strategy trying to push it past it's limits can cause big problems.

If you need money, then you just moonlight at some random job while you learn more about the market.

I wrote my first goal in large font: do NOT become a daytrader. I always view daytrading as a job that doesn't require much ingenuity, and that doesn't run well with my physics background.

Most day traders are actually just gambling, and they are bad at gambling.

 

[twofish-quant]

Yes, I will still have enough to meet margin requirements and trade, but it makes no sense to hold insufficient capital to stage a recovery, so I *will* have to close my brokerage accounts. Not to mention, this is a full-time effort; I cannot be in school while I trade, I will fail at one of them.

Even if you fail at this, it will still look great on your resume.

Desperate, yes. But I am of very clear state of mind to accept the fact that I cannot afford college now.

Community college courses can't be that expensive.

 

[twofish-quant]

What you are doing sounds awfully like day trading, or very short term trading.

It doesn't sound that way to me. The reason it sounds to be like "real" trading is that he is barely breaking even. The thing about most day traders is that they don't keep very close track of the money that they lose. Our friend is dealing with a hard budget so if he loses money, he notices.

 

[twofish-quant]

1. Capacity. Most institutional traders have enough capital to "self-destruct" their positions from the liquidity risk alone.

Most of high frequency institutional trading involves essentially figuring out how to avoid shooting yourself in the foot. You are a mutual fund that wants to buy or sell $X million in stock. If you just show up to the market and say I want to sell $X million in stock, you will get killed.

So you have an experienced trader execute a trading strategy that will allow you to buy/sell $X million in stock without getting killed and you pay the trader a commission for doing that.

Even I have enough capital to make my order visible to people across the world, on some illiquid futures and low capitalization stocks, and just placing that order (even if it isn't executed) can change things.

Yup. There are some stocks in which the traders can figure out who is trading from the patterns.

For this reason, the high-AUM traders are imposed the heaviest constraint - they can only trade "high capacity strategies", many of which promise fewer profits precisely.

For the big investment banks, the profit comes from the brokerage fees and commissions. You aren't trying to "beat the market", you are trying to "hit the market" so that you have happy clients. Even *hitting* the market is a massive pain.

2. I have several prejudices against terming my current/prospective activities as "daytrading":

I don't think you are daytrading. Calling someone a day trader is a lot like calling someone an "astrologer" and "day trading" has as much to do with finance as "astrology" has to do with astronomy.

The thing about it is that you don't *sound* like a day trader. You are asking the right questions, and you seem to be aware of the risks. The one thing that I would do is to maybe to some more networking and show up an finance talks and area seminars. Get to know people and have them know you.

To end it on a relevant note to the rest of the readers, it makes little sense to quit a physics program in college to do a menial task with minimum wage, unless there exist exceptional reasons for potential upside. From this point of view, I've considered the limited paths to success that laid ahead had I left college, and opted for this.

And I think you are going to do pretty well.

 

[meanrev]

The one thing that I would do is to maybe to some more networking and show up an finance talks and area seminars. Get to know people and have them know you.

That's a sound advice. It indeed worries me most (my first post on this forum was about networking, too). I might show up at a few IAFE conferences to start.

I've thought of taking courses part time at a community college. My plan right now is to use Academic Earth/OCW to do some lateral learning on a consistent basis. The only shortfall is that it's hard to find the discipline to complete a course in continuous fashion. The good thing, however, is that I'm used to not attending classes regularly, to begin with. I'm looking forward to edX this fall.

By the way, are you the same twofish from Wilmott?

For the big investment banks, the profit comes from the brokerage fees and commissions. You aren't trying to "beat the market", you are trying to "hit the market" so that you have happy clients. Even *hitting* the market is a massive pain.

That's very true. I see this strange idea being shared by most daytraders, which is that GS is some kind of trading god, but I get the feeling that they're entirely missing the sell-side activities of GS. To put it in perspective, my futures broker holds $1.5b of customer funds, and from my account alone, I generate them $4000+ of gross (before they pay the exchanges etc.) income every month. Now, for the $22.6b of customer funds held by GS's futures brokerage/clearing services, that's a remarkable of money they're roping considering that they're market neutral. (cftc.gov/ucm/groups/public/@financialdataforfcms/documents/file/fcmdata0312.pdf)

I don't see good odds for staying in the market for too long. I'm really skeptical of sustaining this for a reasonable amount of time (even 5 years). If I were to have a fantasy business model, I'd prefer to be doing market-making in the manner of GETCO/Knight Capital, or perhaps HFT trading as in Jane St/Rentec.

Anyway, thanks a lot for your encouragement and kind words. I like the analogy between astrology and daytrading. I'd love to hear more of your thoughts as I update this thread in the future.

****​

Some update on my progress (I'll try to make my posts more useful for the regular physics readership):

With regards to networking, now that I've gone independent, I start to see some of the good things about being a student (previously). You get extremely discounted rates to conferences, regular updates on networking events near you, and I've never made good use of them. The other lesson I've learnt, is that your regular high school friends all end up in the working world with you. The guy that you never paid attention to simply because he wasn't in your "clique" might well be the guy you want to hire, or at least work with, now. Networking starts early, I suppose.

The other thing is that I get the feeling that you can fail and try again more as a student than out here. You can always fall back on the fact that you're in college to ask stupid questions, or to excuse your mistake and try again. There's a greater sense of opportunity cost and urgency, to choose something better to do at any given time. I like this notion of putting more weight on execution. One of the reasons I quit my college was that the modal view of people around you is that the key is to find some grand idea. And so when I call someone up at 4 PM in the afternoon saying that I want to do something, say, build a website, the resistance I get is because it's not a "good idea". On the other hand, I'm the sort who's never had a great idea - but I have many normal ones, and my emphasis has always been on the execution.

"I see a bunch of good choices, and there’s the one that you pick and make great."
-- businessinsider.com/executives-share-the-best-advice-they-ever-got-2012-3#marissa-mayer-vp-google-3

I spent the weekend helping my girlfriend with her trading project. I didn't come up with anything particularly ingenious, although I chanced across a decent way of evaluating chance in a trading strategy with a Monte Carlo method (burns-stat.com/pages/Working/evalstrat.pdf) I've seen basic methods like using naive buy-and-hold as a benchmark, or carrying out a walk-forward optimization packaged with trading software, and the idea of bootstrapping with random portfolios is as simple but more effective. I decided to use the chance to kill two birds with one stone; I wrote the code to carry out the Monte Carlo simulations, and used it to evaluate the trading strategy for my girlfriend's project.

Physics has helped a lot. One of the early lessons I received from friend while he still managing his hedge fund was that no one would write a book or teach you how to trade unless
(1) the method has already run out of profits
(2) he can't make enough money from trading to make it more worthwhile than teaching

The good thing, for me, is that I'm not afraid to do something even if it involves substantial math or programming, while others are the mercy of contract programmers, financial advisers, consultants etc. Now, I noticed that thousands of dollars of work is easily accessible for free on the MATLAB Central, and that all it takes is a couple of days (for me, less for others) to make it something useful. For instance, this is the cleanest description of cointegration analysis that I've ever seen (mathworks.com/company/events/webinars/wbnr55450.html?id=55450&p1=961661643&p2=961661661), pointing out some common mistakes in the use of the Engle-Granger/augmented DF/Johansen/Philips-Ouliaris procedures, it's all available for free - while some traders (won't name) charge $3000+ for a half-day course on the same material.

 

[twofish-quant]

By the way, are you the same twofish from Wilmott?

Yes.

To put it in perspective, my futures broker holds $1.5b of customer funds, and from my account alone, I generate them $4000+ of gross (before they pay the exchanges etc.) income every month.

And the fact that you are even aware (and keeping track) of this, puts you ahead of most people. Most gamblers don't realize how much the casino is making, and most casinos make it difficult for gamblers to realize where the money really is being made.

I don't see good odds for staying in the market for too long. I'm really skeptical of sustaining this for a reasonable amount of time (even 5 years).

You are learning lots of stuff which will be useful. The big problem that you'll have is getting past the gatekeeper. Without letters behind your name, the person in HR is just going to toss your resume. However, if you can keep this going for a few years, and "meet people" then I think you'll find someone useful.

In any case, one thing that's good is that if you blow up, you'll blow up quietly. It's not as if you'll take down the world financial system.

Physics has helped a lot. One of the early lessons I received from friend while he still managing his hedge fund was that no one would write a book or teach you how to trade unless

Yup. That's one reason I have sort of stayed out of high frequency trading and am mostly in the world of derivatives. If someone wants to sell you a derivative, there is pressure to talk about how the derivative was valued. High frequency trading is another world in which people don't talk very much.

Which is weird since there are very few secrets. People move around from firm to firm, any trading strategy is going to be widely known. However, I think there are some interesting game theory here (I know X. You know X. However, what matters is if I know that you know that I know X.)

The good thing, for me, is that I'm not afraid to do something even if it involves substantial math or programming, while others are the mercy of contract programmers, financial advisers, consultants etc.

The other thing is that you are small enough to take risks. If you write a MATLAB routine, and it turns out that there is a serious bug in it such that you lose a chunk of money, then it's just experience. Once you start dealing with huge sums of money, you end up with tons and tons of bureaucracy, and ten people looking over every equation.

 

[kenkhoo]

Sorry to threat-jack. I'm a masters student in physics and I'm pretty interested in trading too. I was quite an on hand trader but I've never looked much into algotrading and such - not much exposure on these matters due to the fact that I'm in a third world country.

Any suggestion (texts/articles) to get me started?

 

[meanrev]

You are learning lots of stuff which will be useful. The big problem that you'll have is getting past the gatekeeper. Without letters behind your name, the person in HR is just going to toss your resume. However, if you can keep this going for a few years, and "meet people" then I think you'll find someone useful.

The other thing is that you are small enough to take risks. If you write a MATLAB routine, and it turns out that there is a serious bug in it such that you lose a chunk of money, then it's just experience. Once you start dealing with huge sums of money, you end up with tons and tons of bureaucracy, and ten people looking over every equation.

Thanks a lot for the advice, I can't express how important it means to me. I'll keep your words in mind for the next - I don't know - 1, 2, 5, 10... years to come until I get there.

That's one reason I have sort of stayed out of high frequency trading and am mostly in the world of derivatives. If someone wants to sell you a derivative, there is pressure to talk about how the derivative was valued. High frequency trading is another world in which people don't talk very much.

Which is weird since there are very few secrets. People move around from firm to firm, any trading strategy is going to be widely known. However, I think there are some interesting game theory here (I know X. You know X. However, what matters is if I know that you know that I know X.)

*****​

kenkhoo: Sorry that I haven't had the time to update recently. I'd recommend

[1] Farbozzi et al (2006) Financial modeling of the equity market, from CAPM to Cointegration.
[2] Chan, E. (2008) Quantitative Trading: How to Build Your Own Algorithmic Trading Business.

Frankly, you could do without the 2nd text if you have retail trading experience - I finished reading the parts I needed in an hour. [3] Ruppert has a section in "Statistics and Finance" where he covers portfolio theory, which I think goes well with the early chapters of Farbozzi. I also found it easier to implement momentum strategies to get started than mean reversion strategies, because libraries for technical indicators are easily available, whereas retail trading software makers assumed something simple like an Engle-Granger test wouldn't ever be used by its clientele, so I had some learning and coding effort to do before I could build a MR strategy. I don't need to remind you to be be very, very careful to avoid snake oil and moonstone when it comes to literature on technical analysis, but a good one is

[4] Murphy, J. J. (1999) Technical Analysis of the Financial Markets: A Comprehensive Guide to Trading Methods and Applications

I'm obviously biased towards spread trading strategies, because that was what got me into the field in the first place, and good papers are:

[5] Avellaneda & Lee (2010) Statistical arbitrage in the US equities market.
[6] Thomaidis (2012) Statistical arbitrage and pairs trading (lecture ntoes)
[7] Gatev et al (2008) Pairs Trading: Performance of a Relative Value Arbitrage Rule

I estimate you probably have more physics AND trading experience than me - so you might want general knowledge that's useful in coming up with a trading strategy. [8] Hull's Options, Futures and Other Derivatives is a classic recommendation, but might be beneath you. You should obviously get the latest edition possible, and I found it rather expensive and couldn't bring myself to buy it and only got the chance to read it because my girlfriend was Hull's student. Otherwise, I hear [9] Joshi's Concepts and Practice of Mathematical Finance is a gentle introduction to the field in general. I think the material in [10] Tsay's Analysis of Financial Time Series will help you with [11] McNeil, Frey, Embrechts on Quantitative Risk Management, whose first two chapters are the cleanest introduction (with decently recent qualitative background about risk management and the Basel III framework) to historical methods like the ES, VAR, copula that I've found. [12] Ruppert's other book on Statistics and Data Analysis for Financial Engineering overlaps the material in McNeil et al, but isn't as well-organized or pedagogically written in my opinion. However, the additional material on Monte Carlo methods in Ruppert's book complements McNeil et al. There's a wealth of content past Chapter 10 that I didn't get the chance to read, but it seems to go out of quantitative risk management to overlap with the material in Farbozzi (recommended above).

I also like seeing things from the investment bankers' point of view, hedge fund managers' point of view, mutual fund managers' point of view and large daytraders' point of view. Respectively:

[13] Kuznetsov (2006) The Complete Guide to Capital Markets for Quantitative Professionals
[14] Stefanini. Investment strategies of hedge funds.
[15] Lynch. One Up On Wall Street: How to Use What You Already Know...
[16] Schwartz. Pit Bull (the back 10-20 pages especially)

A lot of ideas can be found by reading SSRN, Wilmott and Ernie Chan's blog. My recommendations above are really broad, once you decide to choose to go the route of analyzing Twitter messages for market sentiment or building a 20-legged spread across the maturation curve, there are much more specific readings.

I hope I've helped.

*****​

So what have I done for the past 2 weeks? I finished the backtesting and development for my first fully automated strategy. The risk measures were acceptable to my taste - so I finally managed to trade it on last Friday, where it was getting volatile. I lost 950 euros in less than 20 minutes and I was out for the day. It isn't nice when the first trade of your automated strategy is a losing trade, even if the logic tells you that it's perfectly normal. Fortunately, my ex-roommate (who's investing in me) was there to encourage me. I also finally got funded from him and I opened a separate account for his balance.

I got a chance to talk to one of the winners of last year's Rotman Trading Competition and he was pretty impressed by the amount of work I've managed to put into this, and suggested that we meet for dinner some time. One possible contact down the road.

My other time spent was mostly on getting ready for my trip to Tokyo. I took the opportunity to organize old files and build a private database online with a collection of e-books (I have the physical copies, but I can't carry them), journal articles for my research and code, which I figured will be hugely advantageous in the future when I have to collaborate with multiple people. Besides, it protects me against loss of data. My books are mostly on physics and math, as I've noticed that a higher frequency of finance/business/trading books are utter rubbish. My plan is to run only that 1 algorithm while I'm in Tokyo, while spending some quiet time doing lots of reading.

The next big progress that I made was also infrastructural. If I traded from Tokyo, then data would stream over from Chicago, then my order would have to be routed to the brokerage's risk management engine back in Chicago, before it goes out to the Eurex's matching server in Frankfurt.

Wissner and Freer have an excellent paper on the Physical Review E, "Relativistic statistical arbitrage" (http://www.alexwg.org/publications/PhysRevE_82-056104.pdf) which explains the importance of this delay, amongst other things. I figured I would be losing full seconds trading from Tokyo - but more importantly, I was worried that my connection would break down in a position where it's difficult to contact my broker.

So I made up my mind to colocate my algorithm in Chicago. Now, I've long planned to do this - but at this stage it's still too early. Moreover, my plan isn't to compete with the high-frequency traders (yet), so the 20-30ms times to Chicago from Cambridge sufficed up till now. Nonetheless, the whole Tokyo issue just pressed for me to colocate earlier.

And so I went through the machinations to build a virtual machine on a server in Chicago dedicated to trading. My initial plan was to construct a stateless virtual hard disk, upload it onto the Windows Azure cloud, and then do a persistent installation of my code and platform, then run it via a remote desktop connection. I reasoned that Azure is on the same internet backbone, and it's fast - much faster than other cloud providers.

The plan is flawless, but the execution was much more difficult than I expected and I didn't have the time accomplish this and I have to start packing for my trip. So I set Azure aside and looked for another provider. Long story short, I finally did. I've never managed a 'true' server in my life, so there was lots of learning curve to climb and when I was finally set up, I anxiously sent a few pings to the brokerage's server. It averaged 1775 microseconds, with 54 microseconds in standard deviation. I couldn't contain my excitement, because now it would only take $58 per month to run this full-time, and people pay in the order of $8,000-$15,000 per month to get a rack at CME's data center in Aurora for 250 microsecond execution rates.

My next step is to develop a second algorithm to complement this one, so I can fully utilize 12-14 hours of trading time. After that, I've looked around at availability of other means of colocation, commission costs, brokerages and liquidity of other derivatives and I started to eye a strategy for the Kospi 200 in the distance - because that would let me trade an extra few hours in the day.

Mission accomplished, and I set off for Tokyo in 12 hours with only a flight booking. I'll learn some Japanese on the way there. I assume the most useful sentence I need is to ask for directions, a la: "(name of place) wa doyatte ikimasuka?"

Because what's most important in life is to know where I'm going.

 

[phylotree]

I dropped out of my graduate school long ago. I supposed I could do better than was to be sticky to the courses. I could write up at least 4 publications in one year on my own. That was because I was taught full-time by greatest online professors ever at the time and so my level became a little higher than the current courses. I decided to leave because I had no money to continue my study while I wasn't supported by many professors around. My decision helped me learn a lot from people in the schools, from their "real" attitudes to their "real" human interactions that they "performed" upon me. Now I have grown up more than before and have got a full-time job...The job is just a reason for me to make up what I lost in the past. I am a programmer, and really love what I am doing.
I have done pretty much with static typed languages but have not spent much enough with dynamic ones. I am sure I am leaving the current job as soon as I could grasp all the basic structures/constructs out of them. I now can create desktop applications by myself, but I didn't realize how to publish any for them to be viewable on the web, which is the stage I am into now with dynamic languages. After I learn enough that I can grow my own wings to soar, making my applications to communicate with each other over the web may probably is the last phase I am in need of later on.

 

[kenkhoo]

I hope I've helped.

You've helped a lot - by getting me started! To be frank, I'm stucked in grad school for few years now... No outcome yet. And recently (particularly by reading your post) I've renewed interest on algotrading. Previously I've read on Quantative Trading by Ernest P Chan, but I've never put in solid actions (laziness, partly).

Because what's most important in life is to know where I'm going.

I can't help to agree more on this. Awesome job, btw... Do keep us updated & I wish you all the best in your tradings!

edit: I'm skimming through some materials now. I'll concentrate more once I have my article out.

 

[meanrev]

I landed in Tokyo two days ago. Both girlfriend and I missed our flights, funny enough. I flew on the same timezone for 28 hours or something. (Boston -> New York -> Toronto). Sad.

On the bright side, it was the first day to try out my algorithm and server - I put it on autopilot and went for some convenience store food. Came back and I was losing 250 euros. Then I had to check out of my hotel and travel to the city center, so I worried a little but decided that it's safe to leave the server running since I have a stop loss with the exchange. I finally settled down at my hotel 4 hours later, and saw a host of messages from my friend about how he couldn't bring himself to eat breakfast because we were hovering near stop loss and all that. Then I said wait a minute, let me see (this was around 11 AM Zurich time for the German DAX), and I noticed the highs were nearing the opening price (which acts as a resistance level, as technical analysts would say), which was a crucial point - it either breaks out or pulls back. So I logged in at the perfect timing- at this stage we were only up 200 euros - I said watch, the trade will close in the next few minutes and you can have your breakfast, whether we lose or make it. And voila, 1600 euros in his pockets. And I felt pretty prophetic.

I had warned my friend that I will be overleveraging with his account - but he has little option because he can only afford a small minimum investment, and he still wants to go ahead. And fortunately, the first trading day for the server was good, and I just increased his personal savings by 12% in the matter of 4 hours. Advised him not to celebrate though, he'll lose similar amounts on certain weeks, so the effective profit is much less than it looks.

On another side, I do make reasonably more than investment banking interns. Just enough to stay at the largest suites in regular 5 star hotels in Tokyo, but not enough to be a regular at the Mandarin Oriental. It comes with lots of frugal planning to get frequent stayer/flyer mileage with least cost. I have Excel spreadsheets just for that. After that, you only pay for a discounted rate on the lowest room rate. And to emphasize, I did eat only convenience store instant noodles. I just thought pretty cheekily when my college instincts made me hold the lift for random guests at the hotel and in return for the favor, they commanded me to press their floor for them (though I happily obliged, since I was going to ask courteously anyway). I'm happy with my physics sophomore appearance and still wear the same watch I've had since last year of middle school. I hope I can succeed in what I'm doing without ever sacrificing this autonomy; I don't like dressing up for others. Without seeming immodest, a photo of the fire exit plan for the room (yes, this is the top floor) and just a small photo of the study where I managed my server in Chicago remotely.

http://a3.sphotos.ak.fbcdn.net/hphotos-ak-snc6/181807_381643271893147_1177638806_n.jpg

http://a5.sphotos.ak.fbcdn.net/hphotos-ak-ash3/600599_381643318559809_1642201888_n.jpg

*****​

^phylotree: Very interesting. I don't know what to add, except that my impression is that there's a trend that 'decision-making' is becoming profitable. A lot of people mix this up with 'web applications'. Take it very simply, Google makes a lot of 'decisions' for you - it stores a predicted profile of topics that interests you, it gives you directions based on your location by default, it recommends pages and local businesses (e.g. food) to you and Chrome translates things for you now almost by default. Facebook's feed decides what are the most important updates about your social circle for you.

I notice this all the time on TechCrunch, VentureBeat whatnot - that there's a whole slew of startups FB/GOOG wannabes whose idea is simply a cool application, backed by a cool team with ambition and a laid-back tech campus corporate attitude. Now, don't get me wrong, their work is very remarkable and they've done a good job acquiring more funding than I ever would. However, many of the startups seem to be feeding a Web bubble 2.0. (Before scaring you all, though, I did a quick calculation of the market cap of all web startups as a percentage of US GDP and it has been on the rise since 2008, but not to the alarming levels of 2000~. Although their P/E ratios suggest that they're pretty diamonds, I was pretty disappointed as I took 2 hours to compile the data and write code to put it all together and was hoping for a huge short trade opportunity. I'll be honest though, there are many disappointing moments - whereas I found the return-on-effort effect much more consistent when I was still into academic research.)

Now, derailing a bit; I'm not a SV entrepreneurial specialist, so I will take it from the POV of capital markets, mixed with someone who is trying to start his own business out of college: The US tax structure currently supports this kind of venture speculation, just as the early government encouraged agarian wealth (Washington himself was one of the richest men from slave exploitation and land ownership), followed by industrialist wealth (the Rockefellers/Carnegies/Vanderbilts) and corporate wealth (nameless CEOs of the huge banks and conglomerates) starting around Roosevelt's time. We are now at a stage where there's a gold rush in entrepreneurial wealth with relatively low capital gains tax rates. From a trading POV, I don't like this entry point, because now everyone's doing it and it becomes overvalued, then now if the recumbent/new president makes a subtle shift in tax rates, we will see a correction, and I don't want to be caught in a downtrend. So yes, web applications are a popular launching point for a fresh college graduate or dropout like me, but I am not going for it.

So anyway, back to you, I think you're doing great - do figure out how artificial decision-making might benefit the consumers of whatever you're designing, I think this is a worthy venture.

^kenkhoo: If you have finished Ernie's book, the typical follow-through is Johnson's algorithmic trading and DMA book.

You might want to consider your goals and infrastructure available to your capital level before you go headlong into it. From an infrastructural point of view, Ernie's advice is mostly only applicable if you are trading semi-automated strategies or writing your own API-level solutions for a Newedge/GSET-execution account. Well, because he comes from an equity-trading background and typically uses MATLAB routines to generate a bunch of trades to enter into a spread trader from his brokers (IB and GSET). Nothing wrong with that, but it's not something you can do at higher frequencies and/or most equity retail brokerages (the average holding period on my trades is 70~ seconds, and the slippages on IB's fills will kill me). Besides, my goal was to make a living, and equities have too small leverages to make it at least possible.

Also, it doesn't harm to look at more esoteric areas, because you have the intellectual machinery to, and you might always find a new breakthrough, which this field is much in the need of.

An obvious suggestion is the area of regime-switching, where Markov models are popular. My suggestion comes from my math competition experiences (particularly, this year's MCM) - in particular, I see an operations research/multiclass queuing network problem for exchanges in dealing with illiquid products. Because now you have arrival rates for orders based on some model, and you want to figure out the optimal way in which to route these orders and match them.

 

[mal4mac]

I talked to one of my ex-roommates at Harvard, showed him some results, and he decided to throw in $30,000 for our fund.

A fool and his money are soon parted.