Spoofers Tricked High-Speed Traders by Hitting Keys Fast
bloombergview.com
bloombergview.com
I mean, if the initial outstanding orders were at risc of actually being bought and he had to deliver, then I think it is just fine that the market is naive enough to judge the worth of a company on some anonymous seller, rather than company performance and market-place condition of said company.
What's next, you start moving stock prices through twitter bots?
I think stock traders should be punished by the market place for acting on poor signals and acting like sheeps.
If he could in fact not deliver the initial sales, then not only is it fraud, but it is a broken system that allows you to put up stuff for sale that you don't own.
Compare that to high frequency traders where it is alleged that 'flash' market orders are submitted and cancelled so fast that no-one could actually take them. How can that be considered OK and yet these human traders are labelled fraudulent?
If that is the case, am I breaking the law if I just buy some shares and, during the phone call with my dealer, make a comment that I hope my share purchase drives the price up?
Why is it different? Because the law says it is.
And if it's not, why isn't it? Basically it's a double standard, because (most of) the markets are too cosy with HFT firms.
Most high frequency traders are trying to make money by providing liquidity, i.e. offering to buy and sell at a better price than the rest of the market. This benefits both the HFT (because they make money) and other market participants (because they can trade at lower spreads).
What these guys were doing was entering orders that they actively didn't want to trade, so that they could push around the price of the stock and repeatedly scalp a small profit. They had no intention of improving the quality of the market (either by narrowing spreads or improving price discovery). In fact they were actively making the market work less well, by impeding the process of price discovery with their "bluff" orders.
The whole point of a market is to be an efficient mechanism for matching buyers and sellers at a fair price. If someone is deliberately trying to distort the fair price, they are undermining the value of the market (and other market participants, not only HFTs, will suffer because their trades will no longer take place at the 'fair' price, but instead at the 'fair' price plus whatever direction the scalpers happen to be pushing it in at the moment).
I don't buy the "providing liquidity" defense of HFT.
http://blogs.wsj.com/moneybeat/2014/04/03/schwab-on-hft-grow...
So what. 99.9999999% of the pixels blatted onto your screen aren't looked at. Why do you care?
Let's say I'm making a market in a derivative product (A), one where the price is 'derived' from the price of another product (B). By a simple equation. Let's say A = 2 * B. No one likes product A. No one trades it. Lots of people trade B. All day long B moves around. B ticks up, I have to cancel my bid and offer on A and move them up 2 ticks (I do it quickly, cos I'm an evil HFT). B ticks down, I have to cancel my bid and offer on B and move it down 2 ticks. All day long I move my quotes. No one fucking trades. 100% of my orders are cancelled and replaced on a different price.
Why do you care?
A limit order is a limit order. It isn't doing anyone any harm by being there.
>Some high-frequency traders have claimed to be profitable on over 99 percent of their trading days.
McDonalds are profitable on 100% of their trading days. Casinos are likely profitable on 99% of their trading days. If they weren't profitable they wouldn't be doing it. Market-makers like Virtu are more service providers than they are 'traders'. They are the middle men who allow others to trade and take risk, and extract a small fee for doing so. Here's a good analysis to read: http://blogs.wsj.com/moneybeat/2014/11/13/virtus-losing-day-...
The cheeseburger market is rigged.
However, thinking logically about the whole HFT business: As I understand it it's something like this:
Let's say I want to buy 100 apples at $1 a piece. A middle man comes and says: I see that you want to buy 100 apples. Let me buy them for you. Here they are (an apple is now $1.001)
Why do I need this middle man to steal from me? You will probably say that they are providing liquidity. It's still stealing and the markets had enough liquidity before the whole HFT gang came.
Maybe I'm wrong, if so please explain me why.
The HFT is your counter party. When you say you want to buy apples, you are buying them from the HFT. Like any reseller, they are hoping to have bought those apples at a cheaper price than what you want to buy them, because they are a profit making enterprise. Like any reseller, the service they are providing is being able to sell you those apples right now, when you want them. They have taken on the risk and expense of finding cheaper supplies of apples and held them for the appropriate time to sell them when the price was right.
Like any reseller, you do not have to buy directly from them and pay their mark up. You just need to invest the same energy and expense that they do on supply chain to be able to get apples at cheaper prices than people want to pay. Of course, you are probably not in the apple reselling business, so it might not make sense to do this.
And how does that compare to the pit trader order cancel rates HFT replaced?
> Some high-frequency traders have claimed to be profitable on over 99 percent of their trading days
Are these same traders potentially working on an IPO deal? How about being acquired.
Lets put it this way, it is trivially easy to come up with a strategy that is profitable (also just FYI they mean the buy/sell made money, they dont count their operations cost in that) on 99% of days. Its that 1% that wipes out your whole company that is hard to avoid.
I know we used to have a granularity on price that was pretty thick, what would be the issue with having an order have to be on the market for at least a second? I have my doubts about sub-second trading increasing liquidity and making the markets more effecient
They weren't genuine, that's the point. They were offers made with the express intention and hope that they never get filled.
The orders from high frequency traders in contrast are honest orders, when they are sent they honestly express that the trader wants to trade at that price. The fact that they may then be withdrawn 800 microseconds later when the HFT recalculates its price in response to some stimuli and changes its mind about the price, doesn't change the fact they were honest. They're just fast as well. That's what control systems do, they update their outputs when their inputs change.
Either he actually was willing to make good on his offer or he wasn't. His reasons are irrelevant.
"hope that they never get filled" is not a reason to make this a crime. If he actually refused to fill them, then sure. But hoping?
What makes one worthy of legal rules and the other not? Don't know.
It's really hard to prove intent when trading is done in microseconds and the truth lies somewhere in the algorithm. Those who got caught got caught mainly because of their email trail
//spoofing logic goes here.
It is usually expressed in a variety of algorithms in a variety of places. Further, even getting the code is very far down the path of legal proceedings and having the expertise to suss out intent is outside of the skill set of most regulators. Looking through emails/requirements materials and/or having an informant is much more in their skill set.
He is putting orders in with the intention and hope that they never get filled, but as long as they are both fillable and filled when matched, what fraud has occurred?
I can send you an offer for something with the intent and hope you won't accept it - maybe it's a lowball offer to sell my car - but as long as I sell it to you when you accept and an accord has been reached, no fraud has been committed.
(edited for tense)
Suppose an HFT has an order on the market from 0.30 to 0.40. If you submit an order at 0.00, your order will arrive at some time uniformly distributed between 0.00 and 1.00. With probability 0.1 you'll get a fill.
Always hard to prove though, since it goes to the intention of the market participant.
After an order is matched, you can't cancel it except in exceptional circumstances (for example, you mistakenly entered an order very far from the bbo - in that case, with the cooperation of your broker and the counterparty to your trade, you may be able to unwind it).
However before the order is matched you are free to cancel it if you decide you no longer want to trade at that price (for example, you see some change in the fundamentals of the stock that causes you to no longer want to buy, or simply that the price moves in such a way that a buy/sell at your original price no longer looks like good value).
But offering a trade, and then withdrawing the trade manually a few seconds later because I never intended to execute the trade... that's not legal?
What was the original rational for creating this class of thoughtcrime? Why does it criminalize intent, while the action is perfectly routine, and the outcomes are the same either way? There is, after all, the risk of someone actually calling your bluff and buying your sell order, or selling into your buy order, which is cleared instantly... so where's the fraud?.
Doesn't HTF presuppose that we are hard-coding this behavior, this tactic, into the trading algorithms of automated traders?
Yes that's right. So quickly you may as well assume that they are updating the price they are willing to buy/sell continuously in realtime. Which is the idea. And because they are updating their prices continuously and accurately, they can keep the gap between those two prices as small as possible. For some bizarre reason, this is the only field in all of technology where HN contributors get freaked out by this. Name a control system that doesn't update its outputs frequently.
>But offering a trade, and then withdrawing the trade manually a few seconds later because I never intended to execute the trade... that's not legal?
Yes that's right. Markets work on trust, if you don't actually want to buy/sell what you're claiming you want to buy/sell then you are lying. More importantly, if you don't want to trade, what are you doing putting the order in? What does that leave as your motivation? Manipulation, that's what. Which is illegal. It's very hard to define market manipulation, but that's just about the easiest way I can think of - can you think of a simpler definition? They have to settle on some way of describing it, and it's not unusual for crimes to be defined by intent.
>Doesn't HTF presuppose that we are hard-coding this behavior, this tactic, into the trading algorithms of automated traders?
I don't follow
No, that's patently not true. If you are UNWILLING, or UNABLE, to execute a trade when matched, then you are lying.
Plenty of people "don't actually want" to buy/sell, every day - using that phrasing, you could point to people who are forced to cover shorts or dump a stock as it plummets to recoup some of their losses.
"I don't want to sell this stock at this price, but I WILL" should never be considered fraud - for the very least reason that it opens a deep rabbit hole into a form of thought crime that really, do many market participants want opened - when they're asked to explain, with a straight face and credulity, what their market strategies are really intended to do.
A) it isn't fraud, it is order spoofing, which is precisely defined to be based on intention.
B) lots of our legal system is based on intention. It is a common aspect, not a rabbit hole.
If any amount of time passes, it's considered to just "be sitting there", so you can change your mind and cancel it. Microseconds isn't the same as no time at all, when you can pre-issue buy or sell orders that will get filled at the very instant a matching order goes in.
> If he could in fact not deliver the initial sales, then not only is it fraud, but it is a broken system that allows you to put up stuff for sale that you don't own.
Real world commerce works on trust and sometimes retroactive enforcement. If I order ten tons of timber, the world doesn't force the person selling it to prove they have it in stock, I just get to sue them if they don't deliver. Somehow it all manages to work out.
People who enter many quotes that they never trade on would be punished by this system (human traders and "flickering" high frequency traders alike) whereas people providing genuine liquidity, in the form of long-lasting quotes at a good price that they intend to trade, will benefit.
One proviso is that I've never worked on US equity markets, so if this rule is in place there then I wouldn't know about it.
There are also rolling windows and disconnections for people who go crazy with quote spam or have low fill rates.
On the highest level both AI/HFT and Monkey do this for their own earnings. Society decides that profits out of investment are well earned because they serve some important mechanisms that allow companies to operate. HFT used this as an important argument: they add volatility to the market thus allowing more realistic prices. But the same argument can also be made by the Monkeys.
On a lower level we now start to distinguish: the Monkey techniques are manipulative and were used out of malicious intent. Some AI/HFT techniques may be manipulative but intent disguised: if the technique or decision to "manipulate" was devised by AI, who is to blame?
For example, imagine a bot programmed by some 5-10 person team at a hedge fund. They find that running sentiment analysis on twitter and news comments can accurately predict whether a given security will rise or fall in response to a fed press release. The profits are good, so the team manager moves a couple million into the algorithm. One night after work, a team member tells his programmer friend about the algorithm, and mentions some of the most profitable stocks. This friend goes home and programs another bot to corrupt the sentiment analysis dataset, by posting fake comments with properly tuned sentiment. The hedge fund bot reacts as expected, and now the friend has the power to manipulate the bot. He has outsmarted the bot and can take advantage of the high volume trading.
That might be a bit of a contrived example, but corruption in machine learning data is a very real problem. People are just starting to study it. [0]
[0] https://www.usenix.org/system/files/conference/usenixsecurit...
Boo hoo hoo! Won't anyone please think of the HFT systems?
What is wrong is doing something expressly illegal to do it. Oh and then putting it into an email chain with an FBI informant.
http://www.amazon.com/Flash-Boys-Wall-Street-Revolt-ebook/dp...
Also, everyone who has any expertise in the way markets and especially electronic markets work, thinks that "Flash Boys" is a terrible book on the topic. You literally are worse off in understanding for reading it.
Try "Dark Pools". Still sensational and occasionally wrong, but at least it gets close.
If you are trying to learn about electronic / high frequency trading by reading "Flash Boys" you are actively working against yourself. You will literally be more incorrect after you have read this book.
Your humble opinion is factually incorrect. At the heart of this story it doesn't have anything to do with tricking HFT (how ever much that may appeal to you). They broke a basic law, one that everyone who trades knows, and then they put it in an email chain to an FBI informant.
Is it all because of the motive? And bots can legally do the exact same thing because it's harder to prove any motive behind a bot's actions?
The part where he did this with the expressed intention to deceive the market. This is one of those cases where your intentions are far more important that your actions.
You are not allowed to do this if you are a simple guy from the street.
Terrorism, or something.
The reason "spoofing" is illegal is because it gives a huge information advantage to one party that the other parties do not have. Similarly to insider trading. I have no opinion on whether those should be illegal, but if the basic premise of the market changes from "we assume everyone here wants to trade" to "this information is garbage", it will definitely have a negative impact on the price of the traded instruments.
In a spoofing allowed word, that volume is meaningless. Market makers can only differentiate on how quickly they can pull their fake orders and how much risk they are allowed to play with. Not something I'm sure we want to be optimize the rules for.
That said, given the difficulty in enforcement of this particular rule, it may be worthwhile to just throw our arms up and say, "we can't enforce it, so lets make it legal and everyone moves forward on an even playing field".
Would it be illegal? Sure, the reason the strategy works maybe price manipulation, but it is hard to ascribe intent to it all the same.
If you had an email chain stating that you were going to train your bot to layer, that would show intent, which would make it illegal. Otherwise it isn't.
Intentionality is at the heart of the law.