How I Built a Startup High Frequency Trading Firm(howtohft.blogspot.com)
howtohft.blogspot.com
How I Built a Startup High Frequency Trading Firm
http://howtohft.blogspot.com/2011/01/how-i-built-startup-hft-fund-part-1.html
12 comments
Part two: http://howtohft.blogspot.com/2011/01/building-trading-system...
the posting really doesn't have much meat. pretty vague, not that much specific information. i was hoping for something a little more in depth...
does this blog remind anyone else of one of those "how to make money fast at home" advertising thingies you see in pop ups?
does this blog remind anyone else of one of those "how to make money fast at home" advertising thingies you see in pop ups?
It's an introductory post for a new blog. Obviously the author intends to go into much more detail in future posts.
Well you've whet my appetite - I look forward to learning more, particularly to see if you managed to overcome the dual chicken-and-egg problems before going out of business.
His firm already went out of business. He states that he will write a post-mortem at some point in the future.
I like the post-mortem meme sweeping startup world; I find them very enlightening.
I like the post-mortem meme sweeping startup world; I find them very enlightening.
Yes, he said that in the post. I was wondering if he solved the problems before going out of business, or if it contributed to going out of business.
I believe he solved them both at least temporarily, but if either Egg 1, Egg 2, Chicken 1 or Chicken 2 ceases to be true for long, then the business won't survive.
Egg 1: trading lots of shares (check, tens of million shares daily)
Egg 2: track record on trading strategy that is scalable and profitable (probably was profitable then stopped once volatility disappeared or the market changed, but definitely scalable, trading tens of millions daily).
Chicken 1: Volume-based discount (we don't know for sure, but tens of millions of shares daily is enough to qualify as a high-tier class on many houses)
Chicken 2: Clients (check, billion dollar hedge fund)
True: Egg 1 and Chicken 2
Probably true: Egg 2 and Chicken 1
Egg 1: trading lots of shares (check, tens of million shares daily)
Egg 2: track record on trading strategy that is scalable and profitable (probably was profitable then stopped once volatility disappeared or the market changed, but definitely scalable, trading tens of millions daily).
Chicken 1: Volume-based discount (we don't know for sure, but tens of millions of shares daily is enough to qualify as a high-tier class on many houses)
Chicken 2: Clients (check, billion dollar hedge fund)
True: Egg 1 and Chicken 2
Probably true: Egg 2 and Chicken 1
From what is written in the post it seems that they managed to solve both problems.
"Fresh out of college... trading tens of millions of shares daily for a billion-dollar hedge fund." "The mid 2000's when I finished college..." "My firm finally went out of business in 2010 for a number of reasons."
"Fresh out of college... trading tens of millions of shares daily for a billion-dollar hedge fund." "The mid 2000's when I finished college..." "My firm finally went out of business in 2010 for a number of reasons."
The going out of business part wil be quiet interesting, from the first post it does sound like the business conditions have changed. Would it still be possible to get something like this off the ground today or was he fortunate enough to get in before HFT got such players employing the best of the best?
I would say it would be extremely difficult to start a HFT program nowadays. I actually created a profitable system in 2009.. I ended up making around $500k but by 2010, I was making less and less money each month no matter what I did to try and improve the system. I finally turned it off completely about 3 months ago. I think the problem is there is simply no "dumb" volume in the market and everyone is just competing against other HFT arbitrage programs.
Yeah, this doesn't smell right. There are two posts on the blog as of my reading. Neither has much content and they do not sound like someone who has actually been through this.
This does not sound like someone who is really looking to share information. It is too superficial.
The first post was an introduction. The second was a nice overview of how to structure your trading system. It brought up points I've never thought of.
He did say more detail was coming in future posts. I'm not sure what you're expecting; this doesn't seem oriented towards developing models or trading strategies per se, but rather all the ancillary things that goes into building a HFT startup, including the execution of those strategies. Which was what the second post was about.
That is far more useful, since there is enough How-To-Be-A-Quant literature out there.
He did say more detail was coming in future posts. I'm not sure what you're expecting; this doesn't seem oriented towards developing models or trading strategies per se, but rather all the ancillary things that goes into building a HFT startup, including the execution of those strategies. Which was what the second post was about.
That is far more useful, since there is enough How-To-Be-A-Quant literature out there.
The idea is great but the quality of the posts lacks until now. I just hope, you can improve that over time. It really must get more from guessing and ideas to concrete models supported with diagrams, further readings and maybe code where appropriate.
Interesting. Even though my personal opinion on HF trading is that it is a disease that is at the core of rotting out our financial system.
I think there is some truth in that. Markets in which HF trading occur are extremely efficient in the short term, but the information they act on in the short term is primarily about trading and not information which should affect the long term value of the stock.
The Efficient Market Hypothesis seems to break down over the long term when the market is efficient in the short-term, because bubbles can easily develop - and so genuine pricing information gets diluted by complex emergent effects, destabilising the market (in crop futures markets, for example, this has lead to artificial food shortages, and it probably played a role in the Financial Crisis).
I think a transaction tax is the best way to reduce excess short-term market efficiency; if you have to pay the government a small percentage of each trade, people will trade less frequently and only on better quality information.
The Efficient Market Hypothesis seems to break down over the long term when the market is efficient in the short-term, because bubbles can easily develop - and so genuine pricing information gets diluted by complex emergent effects, destabilising the market (in crop futures markets, for example, this has lead to artificial food shortages, and it probably played a role in the Financial Crisis).
I think a transaction tax is the best way to reduce excess short-term market efficiency; if you have to pay the government a small percentage of each trade, people will trade less frequently and only on better quality information.
Care to elaborate further?
Well in my opinion and how it was founded, the stock market should be a place for 'speculation' based on the success of a company, analysis of the market, micro and macro economics etc.
Companies in need of capital and broader ownership 'go public' to secure funds and take in investors with a monetary and intellectual interest in their business model etc.
(I know this is a 'romantic' view of the stock market)
High Speed Trading does NOTHING for owners, investors, the company or anyone but the traders themselves. Essentially it works on the same principle as a the 'worm' in Office Space (geek reference) attempting to make very small amounts of money in mistakes, miscalculations and very short-time discrepancies in valuations of financial instrument.
I feel very old voicing this opinion but it just seems like the very edge of greed and 'money for nothing' attitude to create giant data clusters aimed at just nibbling away at the corners of financial markets for profit.
Companies in need of capital and broader ownership 'go public' to secure funds and take in investors with a monetary and intellectual interest in their business model etc.
(I know this is a 'romantic' view of the stock market)
High Speed Trading does NOTHING for owners, investors, the company or anyone but the traders themselves. Essentially it works on the same principle as a the 'worm' in Office Space (geek reference) attempting to make very small amounts of money in mistakes, miscalculations and very short-time discrepancies in valuations of financial instrument.
I feel very old voicing this opinion but it just seems like the very edge of greed and 'money for nothing' attitude to create giant data clusters aimed at just nibbling away at the corners of financial markets for profit.
Disclaimer: I work at an HFT shop.
A very basic question: if you wanted to buy 100 shares of MSFT right now, who takes the other side of the trade?
High Speed Trading does NOTHING for owners, investors, the company or anyone but the traders themselves.
This is a false statement. Owners, long term investors, etc. value the option of immediacy. That is why options have intrinsic value. If there were no market makers (speculators), it would be almost infeasible to enter or exit a stock position without considerable cost. HFT market makers actually decrease the transaction costs of long term investors by tightening the bid-ask spread (for hundreds of stocks, the spread is as tight as legally possible: 1 penny).
A common misconception is that high frequency trading is like operating a money printing machine. This is also false. High frequency traders take on risk every time they take the other side of your trade. On average, if they're intelligent, they'll be compensated for that risk. In the end though, there is no such thing as a risk free trade. Even pure arbitrages have risk inherent in executing all legs of the trade at once. Pure arbs are very hard to build a business off of in practice.
One other point I'd like to make is: what is the point of this ridiculous speed? If you're confident in your ability to adjust the prices you're willing to buy/sell at very quickly in order to react to new information, then you can make tighter markets. Making tighter markets (if you're intelligent and fast) is desirable for the market maker because it allows him to capture more order flow at what he believes is a fair price. Tighter markets also lower transaction fees for end users of the market. In reality, all this HFT cuts profits away from all market makers (per unit), especially compared to when markets were insanely wide back before electronic trading.
Greed has nothing to do with it, and as I pointed out before, "money for nothing" is the complete opposite of what's going on. High frequency traders take on risk in the expectation of some small payoff. The compensation (on average) is the natural result of risk transfer. I don't think high frequency traders are greedier than people in any other business. Are they profit motivated? Of course. But so is Wal-Mart, GE, and almost every person doing a startup. I think "greedy" is an unfair assessment.
A very basic question: if you wanted to buy 100 shares of MSFT right now, who takes the other side of the trade?
High Speed Trading does NOTHING for owners, investors, the company or anyone but the traders themselves.
This is a false statement. Owners, long term investors, etc. value the option of immediacy. That is why options have intrinsic value. If there were no market makers (speculators), it would be almost infeasible to enter or exit a stock position without considerable cost. HFT market makers actually decrease the transaction costs of long term investors by tightening the bid-ask spread (for hundreds of stocks, the spread is as tight as legally possible: 1 penny).
A common misconception is that high frequency trading is like operating a money printing machine. This is also false. High frequency traders take on risk every time they take the other side of your trade. On average, if they're intelligent, they'll be compensated for that risk. In the end though, there is no such thing as a risk free trade. Even pure arbitrages have risk inherent in executing all legs of the trade at once. Pure arbs are very hard to build a business off of in practice.
One other point I'd like to make is: what is the point of this ridiculous speed? If you're confident in your ability to adjust the prices you're willing to buy/sell at very quickly in order to react to new information, then you can make tighter markets. Making tighter markets (if you're intelligent and fast) is desirable for the market maker because it allows him to capture more order flow at what he believes is a fair price. Tighter markets also lower transaction fees for end users of the market. In reality, all this HFT cuts profits away from all market makers (per unit), especially compared to when markets were insanely wide back before electronic trading.
Greed has nothing to do with it, and as I pointed out before, "money for nothing" is the complete opposite of what's going on. High frequency traders take on risk in the expectation of some small payoff. The compensation (on average) is the natural result of risk transfer. I don't think high frequency traders are greedier than people in any other business. Are they profit motivated? Of course. But so is Wal-Mart, GE, and almost every person doing a startup. I think "greedy" is an unfair assessment.
> it works on the same principle as a the 'worm' in Office Space (geek reference) attempting to make very small amounts of money in mistakes, miscalculations and very short-time discrepancies in valuations of financial instrument.
Just to continue this discussion a bit further, how are these mistakes supposed to be corrected otherwise? :)
Just to continue this discussion a bit further, how are these mistakes supposed to be corrected otherwise? :)
I think HFT can actually serve to smooth out inefficiencies in the market.
That being said, it can, and almost certainly has, been abused. I am thinking mostly of 'quote stuffing'.
Also, there is the potential for a feedback loop between trading systems that can send prices to levels that are way undervalued or overvalued. I think overall these fears are probably way overblown but I suppose they do exist. Basically, I think it HFT fine but I hope the SEC is catching up technologically to prevent both catastrophic situations in time and bad actors from taking advantage of structural exploits.
That being said, it can, and almost certainly has, been abused. I am thinking mostly of 'quote stuffing'.
Also, there is the potential for a feedback loop between trading systems that can send prices to levels that are way undervalued or overvalued. I think overall these fears are probably way overblown but I suppose they do exist. Basically, I think it HFT fine but I hope the SEC is catching up technologically to prevent both catastrophic situations in time and bad actors from taking advantage of structural exploits.
Agreed. But you can't blame traders for greed, since all the owners, investors, and companies are also driven by their own kinds of greed, some of which benefit no one else but themselves.
I think HFT is quite hard and challenging for individuals. I won't feel comfortable trading huge amounts of money that can disappear in a second because of an algo mistake.
I'm looking for some tutorials and may be strategies to make money from Forex. Anyone know a good blog or book?
I'm looking for some tutorials and may be strategies to make money from Forex. Anyone know a good blog or book?
If you think HFT is quite hard and challenging, then Forex isn't any easier: you can still lose a ton of money over an unforseen event somewhere in the World.
More seriously, as someone who trades for a living, here's what I've learned (the hard way, i.e., losing my own money makes for expensive lessons):
Forex is one of the most volatile markets on this side of the "Milky Way". Only Commodities (Natural Gas for instance) beat it. The average range of noise (noise, random movement, not signal) is enough to loose huge amounts of money.
Forex brokers offer a stupid amount of leverage. 50x, 100x (and even 200x) leverage is not trading. It's gambling. Which means that if you want to stay in the game you need a lot of trading capital (want to make a million trading forex? Start with a billion).
Depends on your experience but, if you don't have a lot, start with stocks (CFDs, for instance) or indexes for deep markets.
If you still want to go the "Forex route", some reading material has to include: Macroeconomics and Monetary Policy. You don't need a PhD on it, but you do need to grasp the basics of interest rates, currency parities, inflation, growth, central banking, capital movements.
Statistics. Again, no need for a PhD, but the basics are useful/helpful.
Money Management. People want the "holly grail strategy indicator" that gets you 9 out of 10 profitable trades. That's a myth. The best traders in the industry usually lose 2 out of every 3 trades. The point here is: you make up what you lose with the winning trade. So the real point is not how to enter (though it's still important) but to know when to "exit" the trade.
Basic trading strategies. Some apply better to Forex, other to Stocks, but in general the same principles apply.
And do take care: the Forex market is full of "Win x times your inicial amount in n days with our y fullproof strategy/platform" scammy proposition.
More seriously, as someone who trades for a living, here's what I've learned (the hard way, i.e., losing my own money makes for expensive lessons):
Forex is one of the most volatile markets on this side of the "Milky Way". Only Commodities (Natural Gas for instance) beat it. The average range of noise (noise, random movement, not signal) is enough to loose huge amounts of money.
Forex brokers offer a stupid amount of leverage. 50x, 100x (and even 200x) leverage is not trading. It's gambling. Which means that if you want to stay in the game you need a lot of trading capital (want to make a million trading forex? Start with a billion).
Depends on your experience but, if you don't have a lot, start with stocks (CFDs, for instance) or indexes for deep markets.
If you still want to go the "Forex route", some reading material has to include: Macroeconomics and Monetary Policy. You don't need a PhD on it, but you do need to grasp the basics of interest rates, currency parities, inflation, growth, central banking, capital movements.
Statistics. Again, no need for a PhD, but the basics are useful/helpful.
Money Management. People want the "holly grail strategy indicator" that gets you 9 out of 10 profitable trades. That's a myth. The best traders in the industry usually lose 2 out of every 3 trades. The point here is: you make up what you lose with the winning trade. So the real point is not how to enter (though it's still important) but to know when to "exit" the trade.
Basic trading strategies. Some apply better to Forex, other to Stocks, but in general the same principles apply.
And do take care: the Forex market is full of "Win x times your inicial amount in n days with our y fullproof strategy/platform" scammy proposition.
Haven't read it yet, but it's sitting in my to-read statck:
http://www.amazon.com/Quantitative-Trading-Build-Algorithmic...
Sounds like just an intro. Some of the other books Amazon recommends on that page might be worth checking out too.
http://www.amazon.com/Quantitative-Trading-Build-Algorithmic...
Sounds like just an intro. Some of the other books Amazon recommends on that page might be worth checking out too.
I looked into Forex a while back and came across http://www.babypips.com/, which seems like a pretty thorough introduction/guide for beginners.
I'd be interested to know which firm, and to see the post mortem about why they went out of business first.
I've added another post, this one very in depth and technical on how to build a limit order book.
http://howtohft.blogspot.com/2011/02/how-to-build-fast-limit...
http://howtohft.blogspot.com/2011/02/how-to-build-fast-limit...
This just screams link bait and spam.
Just look at the domain: "howtohft.blogspot.com".
1) Post fake "how I built X startup to do X" post, add juicy terms like "bootstrapped" and "college".
2) Post on HN
3) ????
4) Profit.
Just look at the domain: "howtohft.blogspot.com".
1) Post fake "how I built X startup to do X" post, add juicy terms like "bootstrapped" and "college".
2) Post on HN
3) ????
4) Profit.
I don't see any step 3 - what's he selling?
Don't see the obvious (4) profit part yet either, so (3) must be extra mysterious.
he is hoping to build his reputation to land his next opportunity.