Tradebot had zero trade-lose days in last 4 years
cnbc.com
cnbc.com
Remember how much Dick Grasso was paid as head of the NYSE? Typical exchanges charge substantial membership fees and reap massive profits.
In the age of computers the traditional exchange becomes a lot less valuable, since it's easy to match buyers and sellers over any computer network.
The law requiring all trades to be filled at the lowest price was motivated by major exchanges not wanting to be cut out of the action (even though a trader might be OK with getting not quite the lowest price in exchange for much lower transaction fees overall).
Also, the majority of HFT platforms run off of the major exchanges on private networks that only sometimes pass orders on to the rest of the market. The more liquidity/volume occurs on these smaller networks, the less relevant the older established players (NYSE, NASDAQ) become. Exchanges have traditionally been gatekeepers charging fees to use their monopoly product.
This battle has been going on for a long time (and so far the upstarts have been winning in spite of the hurdles imposed by the influential old timers).
FUD about algorithm based trading strategies is just a part of this larger battle and it's 100% nonsense.
I do think that there seem to be some problems with the "PRICE TIME" rule. My understanding is that is supposed to mean that if two orders get placed at the same price, then the one placed first gets filled first. However, it appears possible for some players to work around this rule and jump in first. That seems unfair to me.
Tradeworx says this too: Jumping ahead of an order that was placed earlier at the same price by another trader is an UNFAIR practice, because it undermines the principle of PRICE-TIME priority on which our equity markets are premised
Unfortunately, this UNFAIR practice is widespread, due to a deficiency in Rule 611 of Regulation NMS
HFTs should not be blamed for exploiting it – in fact, many HFTs who exploit this deficiency do so unwittingly
Instead, the regulators should work to correct this deficiency in the market structure ASAP ! (page 17, http://sec.gov/comments/s7-02-10/s70210-129.pdf)
So it doesn't bother you that this "investment" doesn't have any relation to building productive enterprises of any sort?
One might might think that the nastiness of 2008 would wake people up to the idea that markets can generate destructive speculation as well as constructive investment. But fixed ideas mostly resist reality.
And contrasting the NYSE profiteering to high frequency profiteering is kind of false argument. It's more a matter that deregulation has been a disaster and the situation cries for a modern version of re-regulation - except that the inmates are still well in control of the asylum.
As the article points out, these firms provide liquidity which makes trading faster and cheaper. Theoretically, more liquidity should also make it easier to move capital to the most profitable enterprises.
Actually, the article suggests that the firms pulled out of the market, and that their actions amplified the crash. HFT is a zero sum game, and these firms just demonstrated that they exist to extract value from the market, not provide it liquidity.
[edit]:Though there are obvious differences, I feel like HFT in general is just another form of side betting, alarmingly similar, in principle to CDOs and other derivatives, in at least a few ways. They're exotic, and theirfore unregulated, they dwarf and therefore distort "real" market activity, and they appear to provide value, but the geniuses behind them underestimate the risk because they can never account for all the possible black swan events.
When "real" markets depend on these educated guesses, and the guesses turn out to be wrong, the real markets suffer. Sure we know how to create fancy new financial products to back up these guesses, and exploit the leverage provided by developments in math and computer science, but the fact remains that we still can't even identify, let alone account for what we don't know.
The good news is that because these particular bets are shorter term, in this case the immediate damage is reversible, though their were definitely consequences that were irreversible, and if we keep allowing our equity markets get pushed around by HFT, the next time the system hiccups, the irreversible consequences may spiral out of control.
But I don't have a very deep finance background. I'd love to hear thoughts from some people who do.
Without high frequency traders, the market became far less liquid. This was considered a bad thing by almost all involved. It's almost as if high frequency traders are actually providing a valuable service for the rest of the world!
Also, HFT is not very vulnerable to "black swan" events. HFT doesn't usually use leverage, there are very few counterparties, no leverage, and almost no contagion.
I totally agree.
>This was considered a bad thing by almost all involved.
What I'm trying to propose is that the regulators involved in overseeing markets are becoming more and more out of touch with the techniques being used to improve their efficiency. What's important for long term economic health does not align with what is important for short term market efficiency. And as a result of this, the long term economic policy cannot keep up with the increasing degree to which short term regulation can affect markets.
Yet long term economic health has an impact on other metrics of "utility" beyond itself, and we're in trouble if the power of the people in charge of making the rules can't address the needs of the people who depend on them.
In other words, our regulators actions are predicated on the assumption that we knew how to improve the efficiency of the equity market. Lately, the influence they have on the markets has been increasing, but the amount of oversight they have been subjected to is decreasing. The skills that are relevant to getting politician into office are diverging from the skills and attention that are required to selecting these regulator in the first place. The politicians wind up screwing the people who both elected and depend on them.
--
Let's consider the current, using arbitrage to improve the market's liquidity case, as an example. The most effective arbitrage techniques are made possible in part by simultaneously exploiting developments in computer science theory with those financial theory, and combining our improved ability to process this newly discovered information with our improved ability to implement our improved knowledge. In other words, in this case our increasingly sophisticated understanding of finance and computer science is what allowed us to more efficiently implement and test our hypotheses in the first place.
Previously we used CDO's to improve our ability to assess risk. We did this because we assumed that assessing risk was important to the efficiency of the markets. However, CDOs were made possible by developments in financial theory and implementation, as well as statistics, and the politicians in charge of influencing our dependance on ability to improve this sophistication were out of touch with the long term effects that their decisions have.
Why are crashes bad?
I think the point that you may be missing is that there will always be booms, busts, panics, euphorias, etc. Those are byproducts of human psychology.
The real problem is when people blindly expect the market to always be stable, to always increase in price, etc. Most of the people who lost money in the recent financial crisis were people who couldn't afford to lose on the bets they made or who had essentially let it all ride without understanding that sometimes share prices go down too.
It's a self-serving argument at best. I sometimes wonder why the whole thing couldn't be replaced with a giant daily auction.
If I put in an order to sell one share of GOOG for $1, would you start to believe that the market price of GOOG was $1 per share?
A market price is only as valid as the degree of consensus. Increased liquidity means that there is a high degree of consensus about the current price. This means that one can feel additional confidence that one is getting a fair price.
The main concept to think about here is an order book. Think about the psychology behind orders that are a ways away from the best price (both bid and ask). Those parties are willing to transact, but at a premium. In other words, they prefer to hold unless a sweet enough offer is made.
Thus, for a very large order one has to pay that premium.
The investors get a daily price, and they have to notify us they day before.
There seem to be two obvious problems here. First, what actually happened was that at the point where that service would have been most valuable they stopped providing it. There seems to be a difference in value between idealized theoretical HFTs that provide liquidity at all times, and real HFTs that provide liquidity when there's plenty of it around anyway.
Second, "they abruptly took X away and the results were bad, therefore X was providing something valuable" is an unsound argument -- e.g., coming off a drug cold turkey can be very bad even if the drug was doing you no good at all. It doesn't seem implausible that a sudden drop in liquidity during a market crash might be a very bad thing quite regardless of the optimal level of liquidity in general.
That said, there's also traditional leverage used in many cases because you may have a fully hedged position across different exchanges that don't recognize each other's instruments in margin calculation. For instance, you can be long Jan 11 Brent Crude Oil on CME and be short just as much on ICE. The risk is practically zero, so firms borrow as much money as they can to put on as much position as they can. If you've got more margin than others in these sorts of trades, you win huge when everyone else is backed up to the wall.
ALL companies exist to create wealth for their shareholders. Google does't provide search out of pure altruism, but providing their service and making money go hand in hand.
Incidentally this sort of trading is only zero-sum if you ignore time. Taking a position based on events that have not happened yet is in fact buying or selling risk.
We do this because by facilitating companies we get employment, goods, services and wealth for owners. Of these the utility from wealth for owners is the least significant.
From the point of view of a retirement saver who invests at a single price per day, high frequency traders do hold positions for close to zero time. So in this zero sum game hft traders are just reducing our income.
On the other hand I'm not sure how you police it, because some home-investor with a trading program that kicks in and saves him from losing several thousand while he's in the bathroom is engaging in algorithmic trading too, and I wouldn't deny an individual investor such safety mechanisms.
In other words, if we rule out insider information, isn't all trading, on some level, "algorithmic"? The real-ness comes from the financial statements of the corporations being invested in.
Speculation is an inevitable property of markets and is indistinguishable from more "concrete" behaviors such as hedging. Suppose a farmer buys rain insurance but not hail insurance. He's hedging one and gambling on the other. Yet he bought no contract for the latter.
Such transactions do help build productive enterprises, as they help with price discovery of the shares and add liquidity (which makes the investment more sound for everyone).
General attacks on speculation and finance go much further back than anti-Semitic worries about "Jewish finance", to long before the medieval era when European Jews became associated with finance--- the Jewish scriptures themselves, and writings of the early Christians, both have a bunch of attacks on finance in them. Surely the Talmud itself isn't anti-Semitic?
And a modern quantitative understanding of the role of speculation is pretty unconnected to Jews (most financiers are atheists or Christians, anyway).
First, Talmud itself may well be anti-Semitic in this sense, only this is irrelevant. For example, the Bible says, quite clearly and redundantly, many things that Christians reject. Bible doesn't matter--Christians and their actions do. Likewise, Judaists matter and not the Talmud.
Second, Jews, in the sense of targets of most anti-Semitism beyond the troglodytic variety, are not a religious, but rather a social construct. Hitler wrote this in Mein Kampf, and this is why it's not okay to talk about anymore. The thing that makes it tricky is that while the racial aspect is relevant, it is not relevant in the way that makes criticizing Jews racist. That would have been Jew-by-culture <=> Jew-by-race. Instead, we have Jew-by-culture => Jew-by-race.
"Talmud itself may well be anti-Semitic in this sense" In what sense? In the sense of attacking speculation since it's been pre-decided that all attacks on financial speculation are anti-semitic going back to Roman times?? The possibility of the Talmud being anti-semitic was simply assumed and implicitly rejected by parent as a standard rhetorical device so I'm not sure what you're driving at here. (and while usual the homily about actions mattering more than words is appropriate in some context, here it's apropo of nothing given that the overall discussion concerns the origin and implications of various criticisms).
Modern antisemitism may indeed involve the cultural racial equation you describe but this doesn't demonstrate that either modern finance theory and early Jewish scripture, the criticisms of finance mentioned by parent, have any relation to modern antisemitism.
Edit: I see that you've edited your post.
_delirium's post had two things in it that are incorrect. It was highly rated, so many other people are presumably wrong in the same way. I pointed out why they're wrong.
On most other topics, it wouldn't occur to anyone to object to my pointing out why some things are wrong, just because it wasn't apropo [sic] the origin of the thread. This is why I don't talk about this and a few other topics publicly unless I'm anonymous, and mostly even then. People just tell me to shut up, in so many words (and that's if I'm lucky).
When I described your points as "Apropo of nothing" I meant related neither to the thread's topic nor Delirium's post. You should be able to read above why.
Congratulations on the lamest troll I've ever seen on HN.
It confuses what is necessary - taking risk, with what is reviled - extracting profit at the expense of others from the deficiencies of markets.
... and I don't think you can draw any lines between Minsky and anti-Jewish conspiracy theories. Indeed, "all criticism of speculation is anti-semitic" is as much a "conspiracy theory" as any other hot-button generalization.
Speculation sometimes has negative consequences for the economy. For example a few years ago oil prices were driven higher than they would otherwise be by speculation and this reduced GDP.
It would seem to me that if this were a problem, private networks would emerge that either disallowed the look-ahead option or allowed the party placing an order to declare it off limits to look-ahead (which would probably result in an increased transaction fee).
So in other words, the only time this could really result in exploitation is if one party lacked the wherewithal to use a different trading network (or a large trade broker) when trying to do a large transaction a few shares at a time.
This seems like the sort of thing that would attract innovation and would be fixed before anyone could blink, as trading network firms realized they could make more money by offering both options. My guess is that the lowered fees available in the sort of ecosystem that exists today would limit the demand for a trading network that disallowed it. After all the feature was added on as a desirable product fairly recently and nobody complained about it... until HFT became the scapegoat du jour for the financial crisis.
HFT provide liquidity. With enough of them, you really don't need a NYSE any more. After all, they're doing the same thing the exchange does.
Now all the attention to HFT makes a lot more sense. The logic never added up, but the politics obviously do.
Thank you.
I mean, a bug in my coding means someone potentially doesn't get an email with a download link. A bug in theirs potentially means millions of shares are tossed around, possibly even confusing other "well-coded" terminals to trigger a feedback loop.
I am interested to know if the stop command was a "save your own skin" move or a "save the market" safeguard.
Definitely "save your own skin." The consequences can be dire when trades are "busted" (as many were). Consider this hypothetical situation:
Shares of a stock are trading at $50. After that plunge on Thursday, it's trading at $18, and an algorithm jumps in. The shares rally and at $22, the algorithm sells the shares for a quick profit, and calls it a day. The shares go back to $45 by the end of the day. (This is basically what happened).
The next day, NASDAQ rules that all trades at a price below 40% of the price before the plunge [are invalid](http://www.streetinsider.com/Insiders+Blog/NASDAQ+Busts+A+Nu...). Now that purchase at $18 didn't happen, but the sale at $22 did, and the high frequency trader is short the stock at $22, with the current price $45. That's why lots of high frequency trading shops shut down when trades are in danger of being busted.
Note: I don't do high frequency trading, this was just from interested reading after that plunge happened last Thursday. It seems bizarre to me that they can rule a purchase invalid but keep the sale, but I believe that's what happens.
Bernard Madoff used to have a reputation like that. I rather suspect that the only way to be a winner that consistently is to do something as shady as what Madoff did.
After edit: noting downvote, I'll give one example of a possible shady practice in this industry mentioned in other comments on this thread. Not following the price time rule, which the trader mentioned in this thread says can happen "unwittingly," despite a regulation requiring the first order placed to be the first order filled.
You put quotes around "unwittingly" like you don't believe it. It really is unwittingly - the high frequency shop submits a bid at the same time as another place (or even afterwards), but the exchange fulfills the HF order first. See page 17 in http://sec.gov/comments/s7-02-10/s70210-129.pdf for how it occurs and page 18 for some real world examples.
"Empirically there is a 1.7 cps difference in profitability for a posted share that is first in line vs one which is last in line" (pg 17 from the document linked above. You have read, that, right?)
There's what sounds like a reasonable proposal in that document too.
What you are arguing is that because they are successful they must be shady, not that it is impossible to be successful and that therefore the claim must by shady.
Can anyone explain what the first half of this sentence means?
The NYSE has guys (mostly guys) who stand around on the floor and act as market makers for various stocks. They call 'em specialist. They stand around and run the auctions for shares, match up bids and so on. Basically, they make sure that if you want to sell a stock, there is somebody buying, and vice-versa. This is their business, and they can't just throw up their hands and wander away, ruining the market in some given stock.
The high-frequency traders provide a similar function in electronic exchanges. However, they all threw up their hands and walked away, exacerbating an already bad situation as the liquidity left the market.
Every few months a post like this comes up on HN or Reddit and the general consensus seems to be that Goldman Sachs etc are too smart, too fast and too good to compete with.
I got a bit curious and did some digging, and it turns out that yes, some people are doing it themselves, quite successfully. See http://www.elitetrader.com/vb/forumdisplay.php?s=4a039395d01... for example.
I for one think this looks doable.
The biggest hurdle is initial capital. You could probably start a semi-respectable HF operation with 5-10mm in tradeable funds.
Its just how much you want to scale. Also, if something goes wrong, you can lose your capital very quickly. An easy risk to cover for a multi-strategy prop firm, but it could sink an independent trader. I've spoken to a lot of these guys, and a new strategy mysteriously losing 1MM+ in a couple minutes is pretty common.
Back to access: The types of connections these firms have to exchanges are very expensive. If you dont have the scale necessary to cover 10k per month per connection than you can't compete with all of the other players in the market. Then it goes down to execution speed, colocation, etc - those are the things that make non-bank prop firms like Tradebot or Getco successful. The strategies are only a part of the equation, the trading systems are what really separates the pack.
Most of these firms have way more developers than traders, and the developers they have would make Google jealous.
I'm not convinced about the access thing, though. http://www.interactivebrokers.com has decent looking connections for non-crazy prices, and their brokerage fees are good too.
I don't think you'd be able to do a pure HF strategy using it, but it's possible a strategy based on volatility trading could work maybe(?)
Volatility trading is more math-focused, but requires discretionary trading as well. Its less automated, and requires more specialized knowledge than other strategies. Volatility strategies aren't run by random people coming out of school or other industries. They are almost always run by a trader with a very well developed back-end including a quant (who is trained in volatility modelings - not as easy as the books lead you to believe) and still a very sophisticated trading system with access to both the options exchanges (which are unwieldy) and "normal" equities exchanges.
Its all possible, but can you make a living at it? Its a big maybe. Can you actually make more than your normal job or even a fraction if you were with a well endowed quant fund? No way. Not unless you have been in the industry for a long time.
The mechanics:
IB wants to buy X at $100, you want to sell $X at 100. Normally your order would be routed to the exchange and so would theirs. Both would go to the back of the queue: if Foo/Bar placed buy/sell orders at $100 before you, they get filled first.
Instead, both you and IB get to bypass the queue by trading directly with you. The only real cost to you is that this probably incurs some latency.
Don't get me wrong: These firms do provide liquidity, which is great for anyone trading in the markets, but if you are going to try to play these firms' game of running liquidity providing strategies or even stat-arb style strategies, on their home court you just aren't going to be very effective. That is unless you are running something totally new. (Side note: I knew a guy trading MBS based off of an obscure NY state housing law. The strategy was totally unique, and ran like clock work. If you have something like that, than you could be marginally successful. He was making like 300k a year off of 60k in capital).
BTW, that high frequency trading is good for the market because they provide liquidity, is bullshit. High frequency traders do not trade in stocks with low volume (because it is risky), so they don't add any liquidity that wasn't already there.
There's no legitimate reason, AFAIK, for high frequency trading to be allowed. All they do is syphon away our milkshake in transit from Alice to Bob, because their speed allows them to.
We should be taxing these quick buy-sell trades. At least that way more of the money goes to benefit society. Some equation that taxes at the rate of:
80% profit tax for selling stocks held 10 minutes
50% profit tax for selling stocks held 1 hour
10% profit tax for selling stocks held 24 hours
no tax for selling stocks held 1 week
seems quite fair, would have zero impact for real investors (who are investing in wealth creation, not money creation), eliminate these crazy price fluctuations (that could damage unsuspecting investors' portfolios with stop-sell orders that were meant to protect their investments), and generally help everyone live more meaningful lives.
Utter nonsense. Many HFT firms trade low volume stocks - low volume stocks have a smaller number of trades, but a higher margin per trade.
My argument remains that whatever benefit of liquidity that is injected into the market by HFT is offset by the margins that they take away.
If people could not profitably take on short term positions most of the liquidity in markets would dry up. Market Makers and HFT trading groups comprise the great majority of displayed liquidity (and this is precisely the liqudity accessed by 'real' investors when they wish to buy or sell).
If that liquidity is removed spreads would widen and 'real' investors would take the hit.
On a larger scale imagine a state pension fund or a large mutual fund that has $X billion which it needs to invest in the newly illiquid markets how much price impact will their 2M share order for Ford have now?
I don't understand what you mean. What is the difference between wealth creation and money creation?
Same goes for gold.
We could all be working towards a better world where food / shelter / medicine / sustainability is abundant, but people confuse this real wealth with money, which is an artificial interest bearing / government&Fed inflated accounting construct.
There is still some way until they are there in fundamental analysis (http://en.wikipedia.org/wiki/Fundamental_analysis). I think there are some pretty good opportunities in automated fundamental analysis and trading, though. If I had the capital I'd build a trading bot which traded on the basis of automatic news analysis.