Stock Market Manipulation By Millisecond Trading
cfa22 writes "Nice piece in the NY Times today on ultra-fast trading on the NYSE and other markets. The 'algos' that make autonomous trading decisions have to be fast, but I wonder: Is network speed ever a bottleneck? Can anyone with inside experience with millisecond trading provide some details for the curious among us regarding hardware architectures and networking used for such trading systems?" According to the article, high-frequency traders generated about $21 billion in profits last year.
I interviewed at a millisecond (market-making) trading firm in Chicago. They claimed that when a hedge fund, etc. would buy or sell a stock, that one large purchase or sale would typically signal another. Whichever firm could get their quote up the fastest would make the buy or sale, and it's a winner-take-all system. The first market-maker to adjust their price would benefit. Thus, server speed is THE essential bottleneck. Needless to say, they keep the location of their server a secret.
Ultra fast trading is an interesting idea and done right it can lead to successful short term returns, but if you take a Ferrari around a hairpin at 120mph, you're still going to hit the wall and die.
Here's what happens when that particular Ferrari hits the wall:
http://tech.slashdot.org/tech/08/09/10/203233.shtml
A firm I worked with recently tore down an arbitrage network (they were getting out of the business as it was not core) which comprised of a great deal of Layer 2 dark fiber between sites in NYC and an external data center in NJ, Force 10 fabric switches with multiple paths to server clusters, and a great many Sun X-series servers running Linux. This arbitrage network bypassed the standard corporate (i.e. Cisco-based) network as they wanted exclusivity, higher bandwidth and as much speed as possible. Still, there were issues and the whole environment was scrapped since the actual returns did not match the expectations or cover the costs.
When I looked over the shoulders of the designers (they didn't want too much support from the regular network engineering team) they were concerned with raw performance and not as much with security or other daily operational issues. I would characterize it as the difference between, say, a NASCAR Sprint Cup car and your regular transportation. The former is purpose-built solely for performance while the other has to contend with safety requirements, daily functionality, and a lower common denominator for use.
I think, therefore I am - Rene Descartes; I yam what I yam, an' that's what I yam - Popeye
Can someone explain to me the benefit to society of this kind of activity? I get how the stock market is beneficial, generally allocating resources according to the merit of the business ventures involved, investing capital where it will produce goods/services/jobs, and so on. So despite being a social lefty, I'm not anti-capitalism or anti-stock-market; it has risks and flaws but it works. But how does this kind of stock trading benefit anyone other than the traders themselves?
http://alternatives.rzero.com/
Allston Trading occasionally speaks at my university, and they've said that network bandwidth can be a big bottleneck. They needed to install servers across the street from the NYSE to attain the edge they needed.
As far as who the profits go to, ALlston (and I suspect many similar organizations) keeps most of their profits internal, and exercise big profit-sharing programs for their employees. It's actually quite an interesting idea, as this group of almost entirely Computer Scientists are using their expertise to make some good money as a cell in an atmosphere dominated mostly by business-types.
As someone who works with people who do this, I can tell you they spend a lot of money on very powerful machines, and then try to place said machines within walking distance of the exchange's computers. I have been told that running a server at the office is too slow, even if its in the same city. Also, millisecond is the wrong word. Their trading is measured more closely in microseconds.
"Going to war without the French is like going deer hunting without your accordion." ~General Norman Schwarzkopf
Anyways, back on the original question, no, network speed is not so crucial once all of your packets are properly timestamped.
RTFA:
One second after the market opened, shares of Broadcom started changing hands at $26.20.
...
While markets are supposed to ensure transparency by showing orders to everyone simultaneously, a loophole in regulations allows marketplaces like Nasdaq to show traders some orders ahead of everyone else in exchange for a fee.
...
Automatic programs began issuing and canceling tiny orders within milliseconds to determine how much the slower traders were willing to pay. The high-frequency computers quickly determined that some investors' upper limit was $26.40. The price shot to $26.39, and high-frequency programs began offering to sell hundreds of thousands of shares.
On the one hand, you can call this 'perfect price discovery'
OTOH, that specific set of behaviors fundamentally breaks the traditional way that the imperfect markets have worked.
And just as importantly, it represents an unfair trading advantage that you or I will never have.
Allowing this behavior doesn't further market activities, it just allows a few players to accumulate wealth at everyone's expense.
Seems to me that the solution is to close the loophole that allows this to happen.
[Fuck Beta]
o0t!
Pensions and 401(k) plans largely. There's a huge amount of wealth in those. Ironically, the only pensions that act like they give a damn about what they own are the union plans.
You actually could have a life of comparative leisure relative to the past, but humanity spends huge amounts of time competing with status displays which have become vastly more important to personal happiness in the more "relaxed" world. If you're willing to live a 1950s lifestyle you should be able to have far more leisure time than the 1950s person. Remember though that you'll never leave your home country for travel, live in ~300 sq/person house, share a household immediately after college, and you would probably only own 3-4 suits of clothes.
Degaussing scares the bad magnetism out of the monitor and fills it with good karma.
GS sees both the buy order and the sell order.
They can sell before large sell orders. Then buy again after to avoid a loss.
They can buy before large buy orders. Then sell right after to take a gain.
It's called front running and it's illegal.
But the government is not enforcing the law.
My theory is they are letting illegal activity go on to hold up the market prices.
I could be paranoid. But I've been in this thing for 28 years now and I've never seen such goofy market behavior and repeated evidence of price manipulation at market close on thin volume-- and yet there are no investigations.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Thought I'd chime in on this. I have a trading account with $150K+ in it, and my per transaction costs are extremely low (I trade futures though, YMMV when it comes to equities, commodities, etc). The provider I use exposes an API that I can interact against to directly execute trades (although, you're always going to be fighting the speed of light). I wrap around this API with Python and Postgresql. If you're smart and your algorithms are rock solid, you can do well (I don't have to work if I don't want to). On the other hand, you have to make sure a human is *always* in the loop somewhere, otherwise you end up with the United Airlines fiasco someone pointed out above.
>But how does this kind of stock trading benefit anyone other than the traders themselves?
How does any trading benefit anyone but the trader themselves? This high frequency trading issue is a moot point. There is nothing sinister about it. What bothers people is that a few are doing it and making oodles of money. I work in this industry and have worked at investment banks where you do high speed trading. But now I work at a hedge fund and we just avoid it. It is an arms race and what bothers people right now is that there is a break in technology capability.
I am not kidding here. I know for a fact a few houses have been able to get micro-second trades, whereas the rest of the industry is still dealing in milli-second trades. Thus what is happening is that some very very big houses are getting whalloped in a major way. Hence the sour grapes...
"You can't make a race horse of a pig"
"No," said Samuel, "but you can make very fast pig"
We did trading similar to this a few years ago, though we used a simple, pure-arbitrage strategy. We were trading the morning of 9/11 (pre market-open - the market never opened that day), and didn't pull the plug when the news hit. We were actually testing feasibility, and had not yet installed our server in a rack in Manhattan. My partner was trading (actually, "overseeing" is a better term) in a day-trading room about 30 miles N of NY. (Later we moved our server to the same building as the Island ECN, which at the time owned the ITCH technology mentioned in the article.). Everyone in the room stopped trading, their jaws dropped, staring at the monitor in the corner. BTW, the consensus in that room that day was that the market was not going to re-open. EVER.
My partner, though, is sitting in the back of the room, trading away (or, watching the computer trade away), wearing headphones, which already had the other guys suspicious. (The headphones didn't have soothing music, but vocalized order flow information - it was too fast to keep up with visually. My partner did have to intervene when things went awry, and the headphones told him when he needed to act or if things were humming along nicely.)
I figured later that we did about 5% of the pre-market NASDAQ trades that morning.
This gave me some consternation for some time. We made a bundle that morning, and the subsequent increased market volatility was the start of a winning streak. We did thousands of trades a day, and simply did not have losing days. Was there a purpose or value to our trading?
I eventually concluded "yes". We provided market liquidity when it was sorely lacking. The people who sold to us at, say, .50 outside of the market would have sold to somebody else at $1.00 outside of the market had we not been there as a counter-party. We stood there catching hot knives and handing them off, while providing the benefit of a quick and certain sale. This helped reduce spread and volatility.
Indeed, over the next couple of years, our typical profit/share shrunk from .10-.50 to 1-2 pennies, and competitors entered the market and the war of technological escalation started.
Now, as for the "flash orders", which apparently started last year, I don't see the benefit of that to the market. It just front-running in exchange for a fee paid for the privilege.
It's a bit under 10% as measured the new way created under clinton.
It's over 16% the way they measured it until then.
http://www.shadowstats.com/alternate_data
And it's over 20% if you include people who have run out of benefits and people who used to make six figure incomes who "have a job" for a fraction of their former pay.
These are really historic times... the banks are refusing to take ownership of houses that should be foreclosed on. The banks are refusing to list houses they have foreclosed on to artificially pump prices. Cities are tearing down thousands of houses rather than let them sink to their true price (which may actually be zero anyway since without a job, you can't pay anything for a house).
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.