Hacker News new | past | comments | ask | show | jobs | submit login
Trading bots are running wild on cryptocurrency exchanges (bloomberg.com)
181 points by chadmeister 8 days ago | hide | past | web | favorite | 139 comments





frontrunning (n): the practice by market makers of dealing on advance information provided by their brokers and investment analysts, before their clients have been given the information

The prototypical example is a large investment bank hearing about a large client buy order, then putting in a buy order first on their own book, knowing it will go up due to the client order. It is a violation of your fiduciary duty to the client.


The origin of the term is kind of comical to modern ears:

The terms originate from the era when stock market trades were executed via paper carried by hand between trading desks. The routine business of hand-carrying client orders between desks would normally proceed at a walking pace, but a broker could literally run in front of the walking traffic to reach the desk and execute his own personal account order immediately before a large client order. [1]

[1] https://en.wikipedia.org/wiki/Front_running


Same old same old. Only these days it gets done by building data centres in across the river New Jersey and as-the-bird-flies microwave links between New York and London...

Not really. They're reacting to someone actually trading in public on a public market. That is not front running. Someone has to react to a thing that actually happened in the public market, quickest. By definition someone has to be first. Cheating to get there, eg breaking before the starting gun, is illegal. Reacting first is not.

Front running is where you get information about a trade before the trade happens, illegally, and get your own trade in first with advance knowledge that you should not have. That's still 100% against the law whether a computer is involved or not. Do it and you will go to jail. If you got your information from the public market because it is a thing that already happened, then ok. That is really just fine. Other laws might be being broken somehow in other ways or they might not but seriously that's the whole point of a public market and using public information. Everyone has the chance to react as fast as they can.


You're both correct. Front running has legal and illegal versions. Getting any non-public information and using it for trading purposes is illegal, however you can still front-run with public information. Markets react at different speeds and so do indexes, if your machine is set up closer to the exchange and has a faster algorithm you can front-run an index by using information about the changes in the index's stocks.

In the case of this sort of information-speed arbitrage you're not making any stock analysis decisions, you just know an index has to respond a precise way to the stocks within it, and you're able to beat this information to the exchange and buy it ahead of the change.


As others have pointed out, taking a term that has a specific legal definition to relate to an illegal practice and broadening it to cover both legal and illegal activities doesn't seem like a great idea if your aim is to try and increase people's understanding of the markets. It just adds confusion. This is a shame, as your point about markets not being a level playing field because different participants are able to execute on information in different ways is worth making, especially in the context of new types of exchanges.

An analogy would be to redefine all killings (lawful or otherwise) as "murder". It's not helpful. Words have specific meanings in specific contexts, so lets stop overloading them so much when we don't have to.


Ok well technically front-running is a term used in the industry to specify a type of trading, it's not illegal in itself. However, you can front-run in an illegal way. Like driving a car isn't illegal but you can do it in an illegal manner. The act of front-running becomes illegal when you do it based on insider information, much like the driving becomes illegal when you speed.

I hope this clears things up.


To be very clear to all readers: front running is illegal. No one in the automated trading industry refers to the non-illegal trading that they do with a reference to illegal trading.

As discussed elsewhere in this thread (search my nick), "no one" refers to frontrunning as encompassing both legal and illegal cases except the author of the definitive textbook on market microstructure.

Thank you, I gave up arguing it and was getting massively downvoted for my comments.

ok

Wow! Didn’t know about these “secret” microwave networks.

Eg https://arstechnica.com/information-technology/2016/11/priva...


They're not really secret, they're standard practice. Not all that widely known, but that's because they're geeky trivia to most people. There are multiple microwave relay networks between Chicago and NYC. It's neat stuff.

I've worked for an options market maker, chasing nanoseconds of latency improvements. If you want racks in the exchange data center, you have to pay more to be on one side of the room vs. the other, because the latency improvement for being a couple hundred feet closer actually matters.


If you are really curious, I highly recommend this blog!

https://sniperinmahwah.wordpress.com/ (Many people who work in HFT are quite familiar with it.)

It's an entertaining and fascinating "view of the inside."


Some hft firms are rumored to be using short wave radio. Everything old is new again.

https://hackaday.com/2018/05/12/hft-on-hf-you-cant-beat-it-f...


Reminds me of the term "sneakernet." I wonder if "front running" was a thing back in the days of manually moving changes onto a source control server located in-office. I can imagine a mad dash to get code in first on a Friday afternoon.

sneakernet is still a useful thing for large transfers (even in literal walking distance, sometimes). like if i'm sending a large file to a person in the same office or house, i might prefer to just put it on a thumb drive and hand it to them, instead of copying it to some network share or something.

i imagine this will be true for a while (i.e. the practical throughput of sneakernet in many cases will keep outpacing the practical throughput of wired and wireless networks, probably for a long time).


It was only 30 years ago that manual trading ended in London. Similar time period for the US. Other exchanges and countries lasted much longer. See Eddie Murphy's Trading Places.

https://www.bbc.co.uk/news/business-37751599


There's actually still a couple hundred or so traders in the pits at the CME in Chicago. Certainly nothing like what it used to be, though.

> The origin of the term is kind of comical to modern ears

It's interesting, but I don't think it stands out as "comical" from the finance realm, given that programmers unironically talk about "race" conditions during the "handshakes" from "tunneling" with a "certificate" in order to to "serve" up "pages".

On a far-more-obscure note, the etymology of "computer" itself can be traced back towards "something which counts using small pebbles of limestone."


It is comical in the sense that software bugs being named after a literal bug (a moth to be exact) is.

If you’ve ever cut and paste with scissors and a glue stick...

> It is a violation of your fiduciary duty to the client

It is a violation of being put in a position of confidence with respect to privileged information. Front-running clients is always bad. But one need not have a fiduciary relationship with a counterparty to front run them. Practically speaking, however, when non-client relationships involve front-running, we usually go after them as insider trading. (For example, suppose I trade for a hedge fund. I hear my PM is going to buy X. So I buy X. One, I front ran my own shop's order. But two, I also violated their confidence.)


This expanded definition --- of the possibility of "frontrunning" when in possession of confidential information --- is the context of the Khan & Lu paper the paper coauthor cites below to justify his use of the term in this Ethereum paper, and, of course, does not really apply here.

It's of course worth mentioning that in the US, "insider trading" is itself defined largely in terms of fiduciary duty. There's nothing wrong with developing and exploiting proprietary sources of tradable information; in fact, as Matt Levine points out, that's kind of the point of a market (to aggregate information and discover prices). What's really problematic is when you're given access to private information in confidence and then use that information to trade against the owners of that information.

(I imagine this is not new to you; just for the thread).


In even succint-er Levine-y terms: its not about fairness, it's about theft.

The "duty" part is why this definition doesn't apply well to this case. Arbitrageurs here are working with the same public information of all other participants, and have accepted no obligations towards the folks whose orders are informing them. Although this is clearly a structural issue with the exchanges that participants should be aware of, the term "malfeasance" is being thrown around a bit loosely in the paper.

A large investment bank Knight Trading would queue orders after the market closes, and then in the morning they had software that would insert their own buy/sell orders based on what was sitting in the queue from the previous night. Doing this they were guaranteed to make a certain amount of profit every day, and from what I was told it was perfectly legal.

I never put in trades to be executed at the next opening. One always gets an adverse price that way.

Is that superstition or is there a logical basis to that strategy?

Knight was a market maker, not an investment bank.

does that make it ok?

A market maker's job is to have orders sitting on the book to provide liquidity to the market.

sure.. but does that mean it's ok if they stuff them with their own orders in a way nobody else can and that gives them guaranteed profit at the expense of others?

It does. And exchanges and other venues are continually updating to prevent market micro-structure from being able to be easily exploited.

Whoever told you this was yanking your chain. US equity markets (which is where Knight traded) don't close overnight. Trading volumes get thin, but trades can occur at any time of night. You can't just put an order down and wait till morning.

I have been trying to simply visualize how a front runner makes money. Now I think I understand it.

Basically, a front runner looks for orders which will move the price on an exchange. So if an exchange has the following sell orders:

0.1 BTC @ $5100

0.1 BTC @ $5110

0.1 BTC @ $5120

And someone places either a market buy for 0.3 BTC, or a limit buy for 5120+, then they would expect to get 0.3 BTC @ 5110 average price.

However, the front runner, if they can see this order is about to be placed, could theoretically buy 0.2 BTC @ $5105 average, and instantly relist them for 5120, knowing they will be immediately sold to the $5120 limit / 0.3 BTC market buy order about to be placed, in this case pocketing $15/BTC


> in this case pocketing $15/BTC

minus fees and commissions (trading being negative sum is always worth highlighting)


Market makers aren't agents/brokers. They do not have clients.

Right, and races between mutually disinterested market participants aren't frontrunning, which is what I assume the subtext of that comment was.

Author of the paper here. I justified our use of the word "frontrunning" on a Twitter thread: https://twitter.com/phildaian/status/1116698738343460864

There are a few different points here. Firstly, there are plenty of unambiguous definitions of frontrunning that are accepted by the scientific community that this blatantly does fall under; see the Clark SoK we cite in our paper for this in the blockchain space, or the definition Wikipedia uses, which was from an economics article by Khan & Lu (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1140694): "trading by some parties in advance of large trades by other parties, in anticipation of profiting from the price movement that follows the large trade."

The second point here is even if you want to restrict the definition to require privileged information, privilege in p2p networks is not binary: there are some nodes that are better peered on the topology, have better mining equipment, etc. who can use these privileges to extract market rent, harming the security of a blockchain.

This is a complex and dense paper that took 8 people 18 months to produce, and I recommend you read it with an open mind, rather than having a gut reaction to whether we use your preferred definition of "frontrunning". I also recommend reading the primary source material directly, or at least my summary on Twitter - https://twitter.com/phildaian/status/1116155236433956865

Happy to answer any questions!


The definition implied by Khan & Lu is incoherent. It refers back to Larry Harris (Trading & Exchanges most definitely does not define frontrunning this way!) to suggest that, in attempting to execute trades for large blocks, traders might expose their orders and "create an opportunity for frontrunning". But that's basically the fundamental challenge of shopping a large block; it implies that essentially all of institutional trading is "about" frontrunning. That can't be right.

But if 51% of Wikipedians believe otherwise, who am I to argue?

I read the whole paper before I commented, but really wasn't remarking on the paper so much as this subthread. If I was going to pick nits, it'd be with the title of the paper and the allusion to _Flash Boys_, which is just a terrible book.

Late edit:

If you want to bring this back to computer science, I think it's helpful to think about this stuff in terms of the CAP theorem; according to our science, the opportunity to arbitrage latency in a continuous market basically has to exist.


FWIW, I have a copy of Harris' book on my desk right now and every explicit citation of Harris in Khan & Lu's paper appears to be accurate--supported by the text in Harris (Chapter 11).

edit: > But that's basically the fundamental challenge of shopping a large block; it implies that essentially all of institutional trading is "about" frontrunning. That can't be right.

That doesn't make the Khan & Lu definition wrong, just that there is a lot of "front running" that occurs. Larry Harris even explicitly mentions this point:

"When large traders recognize that they cannot trade as well as a professional trader can, they commonly hire a professional to help them trade. They may hire a block broker to act as their agent, or they may ask a block dealer to facilitate their trades. Block dealers who facilitate their customers' trade trade their blocks at uniform prices. The dealers then try to profit by trading the block in the market at a better average price. In a sense, block traders are front runners whom large traders hire to help them solve their trading problems"

This is related to the idea of sunshine trader discussed earlier in the chapter before that blockquote.


I have to concede you're right, and that the distinction that Harris draws is not between front-running and "jellyfishing" or whatever we want to call it, but between legal front-running and illegal front-running. Here's Harris's standalone definition, which is what I remembered:

Front Running: Front running occurs when a broker improperly allows one order to trade ahead of another. The order that goes first usually profits from the price impact of the following order. Front runners hurt the traders whose orders they front-run because they take liquidity that the front-running traders otherwise would have taken. These orders then fill at worse prices than those tat which they would have filled.

Front running is most common when a broker holds a large order that will likely move the market. The broker then trades for his own account first, or he tips off a confederate who does the front running.

Front running also hurts the brokers who represent the orders that are front-run. Broker clients who pay close attention to how well their brokers perform will discover that their brokers who knowingly or unknowingly allow others to trade in front of their orders do not trade effectively on their behalf. When their poor performance becomes apparent, the clients often direct their orders to other brokers. Firms that employ brokers must therefore be vigilant to ensure that their brokers do not cheat their clients and thereby lose business for the firm.

This seems pretty cut-and dried.

But, a little later in the book, there's a whole little subsection on "Legal Front-Running By An Observant Trader", ending with "[the front-runner's] trading is legal. Her profits come from recognizing [the broker's] shortcomings as a broker, and from noting that [the broker's] clients tend to split their orders. She is a profitable trader because she is observant and because she acts quickly on her information."

I fully concede this point.

It's of course important to retain the distinction between legitimate front-running and FINRA-style illegal front-running.


In fairness, it's a very Harris-y kind of chapter which cuts against how you might want to understand or talk about front-running with people outside the industry. (off topic: having met the guy I'm not at all surprised by how he constructs this chapter or the book in general -- which I love).

In the dystopian timeline in the movie, they find The Citation's lifeless, sundered husk at the bottom of the ruins of what was once a massive thread. About DNSSEC.

OK I give up what is The Citation a reference to

It's not a "preferred definition." Your definition of frontrunning is simply incorrect.

FWIIW Michael Lewis is considered to be a serial fabulist by most people on Wall Street; his "flash boys" was a particularly egregious example of this, being a marketing effort for IEX, which was funded by friends of Michael Lewis.


I linked you to several sources for my definitions and you haven't commented on any of them, so no idea how you could reach that conclusion. I've explained to you that even with your restrictive definition, our paper covers analogous behavior in DEXes.

Yes, it's very surprising, no, shocking, that people on Wall St. are reluctant to have behaviors like this called out publicly....

I am not Michael Lewis and cannot speak to the accuracy of his facts; our work is a scientific paper backed by a lot of data currently undergoing peer review. The views on HFT and market design in the paper are similarly informed by cited work that is quite extensive in its data analysis, and peer reviewed. I recommend you read it!


Help me understand what you're trying to say with that middle paragraph, about how "shocking" it is that people on Wall St. are reluctant to have behavior called out? To what behavior are you referring?

I'll add: on Twitter you noted that virtually every HFT trader, lawyer, and quant pushed back (either to your definition or to your allusion to Lewis's unfortunate book). You seem to think that's added credibility for your argument. Isn't a near-universal rejection of that argument from subject matter experts sort of damning for the position you're trying to take?


You pointed out no such thing, and regardless; no such bona fide definition of "front running" exists. You got it wrong: admitting your error would be the honest thing to do. It's OK to be wrong; that's part of science and rational enquiry. It's not OK to not admit that you're wrong.

FWIIW I'm not on Wall Street; I work in tech. Wrong again.

Your results are both trivial and basically linkbait. Everyone who knows what a time priority order book is knows there are stale orders on DEX exchanges. The fact that people are arbing them is markets doing what they're supposed to do; creating one price across disconnected markets. Anyone who knows what they're doing and uses a DEX knows this; DEXes are for people who for whatever reason can't use central exchanges, or for smart contracts that are less concerned with getting the best price marked to the millisecond. In fact I pointed out the stale order issue to people who created the one of the more popular DEX protocols over lunch. They pointed out what DEXes are for. I guess nobody sent you the memo. It's OK to not know this; I didn't know. Personally I think they're bloody stupid, but you making their use out to be something sinister or poorly understood is ... well, a lot like Michael Lewis paranoid lunacy.

I realize you're not Michael Lewis, but you're taking on the title of his execrable book to sell ... something much less science fictiony and exploitative than his tall tales, but still fairly off the mark.


Yes, the people who created DEXes do know about and advertise this issue, because this is work we've been doing since early 2017 and discussing with them even earlier. In fact, I advise a lot of those projects, and many consider me the SME on DEX frontrunning. I'm also the reason 0x e.g. released a bot to explicitly do this.

To claim this is what they were designing for is just ignorance of the space though. None of them knew about or even anticipated this issue; just because it is obvious to you does not mean it was obvious to them.

Just because DEXes are only for regulatory arbitrage (your words, not mine) doesn't mean that we should stand up crappy market designs.

The majority of the paper is about the formalization of and data on PGAs and miner OO, which was also not known in the blockchain community until now. If you point me to a reference to the contrary, or that contains the same results as e.g. Sections 5-7 of our paper, I'll gladly cite this and reduce my claims of novelty. Otherwise, this is something I've been working on for years, I'm going to stick with it, and it's up to peer review in the scientific community to decide :).

I hear what you are saying, but your preconceptions about Flash Boys being a bad book seem to be severely coloring your understanding of the work!


They aren't "preconceptions". A preconception about the book would be if they hadn't read it, weren't familiar with it, and had opinions about it anyways. Flash Boys is a bad book. I'd argue that your desire to immunize the book, which is largely unrelated to your paper, from criticism is harming the case you're trying to make about distsys flaws in DEX's.

The paper is good! The surrounding context you're trying to provide about how real-money markets work, perhaps not so much.


I'm not saying he has preconceptions about the book, I'm saying his opinion of the book is creating unwarranted preconceptions about our work.

The paper doesn't comment on money markets much at all, and cites pretty unambiguous peer reviewed work which is highly substantiated when it does.


Sorry, I was confused because you referred to his "preconceptions of Flash Boys as a bad book", and also talked at some length about how industry insiders don't like the book on Twitter, and also alluded to practices described by Flash Boys that insiders want to hide upthread.

Virtually all existing market designs are crappy. Maybe CBOT is an exception. Michael Lewis' book is a big giant marketing document for a shitty market design, FWIIW: not a preconception; I read and panned his ridiculous book.

I didn't think you made the case in your sections 5-7. 7 in particular looked like speculation to the point of paranoia (which, admittedly, is justifiable in blockchain, but for the kinds of dollar amounts you're quoting -rather unlikely). I'll read it again to make sure.


Well, we show (among many other results) that today's system is potentially unstable assuming miners are rational and profit maximizing. I don't think that's paranoia in a system that gives you 0 security guarantees unless you assume miners are rational and profit maximizing.

> Virtually all existing market designs are crappy. Maybe CBOT is an exception. Michael Lewis' book is a big giant marketing document for a shitty market design, FWIIW: not a preconception; I read and panned his ridiculous book.

I'm not saying anything promoted by Lewis is a good market design, nor that existing continuous limit order books are. Merely that from the PoV of the efficient markets hypothesis, blockchain markets are worse (and have the potential to be better, e.g. by exploring batched designs).


Yes, blockchain markets are worse in many ways; that's obvious. But I don't think your speculation about miners has any merit.

Your statement in section 7 seems to not understand how blockchains work. I could be wrong; it don't pass the sniff test. Since I actually have a day job, I may not have time to look into this in enough detail to thoroughly pull it apart, but hey, maybe they'll catch it in "peer review."


If you find something wrong with it, consider this an open offer for me to revise or withdraw the paper as a result.

That being said, I've been studying blockchains for years, and this paper has been reviewed by some very high caliber designers of blockchain protocols in both academia and industry. It also uses very simple models as in http://randomwalker.info/publications/mining_CCS.pdf which in my view are extremely unlikely to be incorrect (of course, anything is possible).


The reason you're getting pushback on using the term front-running to describe this is that it is a blatant misuse of words for the purpose of painting a behavior as bad without needing to say why. You've taken a legal term that describes a crime and extended its definition to something that is not a crime to make that non-crime look bad. There might be some good ideas in there, but relying on this deceptive rhetorical technique implies otherwise. If you have something worthwhile to say, why the deception?

The "scientific community" does not get to define terms used by the financial industry and financial regulators.

Except trading doesn't have a "scientific community"; they have legal definitions not scientific ones.

Except one party doesn't realize it is a race.

That doesn't matter. The actual offense of frontrunning involves an agent-principle relationship between the parties.

I think that the definition in the parent's comment is erroneous. You don't have to be a market maker in order to front run; any broker-dealer can do it.

That said, I think you're off a bit on the nomenclature, too. Market makers can indeed have clients. For example, [1] is a list of the Chicago Mercantile Exchange's market makers for equity options. Citigroup, which has clients out the wazoo, is on the list. It's (presumably) just that the market making operations themselves are not done on behalf of a client.

You're maybe conflating the concept of a market maker with that of a proprietary trading firm. It's very common for market makers to be prop shops.

Prop shops, by virtue of not having clients of their own, are perhaps some of the least-well-positioned firms to engage in frontrunning. But a player like Citigroup or Goldman Sachs (which does a fair bit of currency and interest options market making) could have a relatively easy time of it.

[1]: https://www.cmegroup.com/trading/equity-index/e-mini-equity-...


> Market makers can indeed have clients

Market makers generally don't have clients.

Once upon a time, I was a big-bank algorithmic equity derivatives market maker. We only had counterparties. Sometimes, those counterparties were our own bank's sales-trading desks. We were isolated from them, in terms of supervision and information access. (The sales-traders had clients.)


I work on a market making team at a major investment bank, and I can back this up. There is a very clear and serious separation between those who trade bank funds and those who trade client funds. We are not allowed to know anything about their positions, accounts, strategies, etc. (and vice versa).

Not only are there very stringent access controls in place to prevent cross-contamination, but there are physical boundaries as well. When the guy who runs the business side of our team started collaborating on a project with a team that did client trading, he had to move off our trading floor and into an office, and we had to revoke this access to, well, pretty much everything. He wasn’t allowed to go near our traders’ desks. It stayed that way until the project ended.


Just going to second this. I used to work for Virtu, which bought Getco and the smoldering pile of ashes left from Knight Capital’s meltdown almost solely for their Designated Market Maker status for so many symbols.

There is a miscommunication here, arising from the fact that bunderbunder thinks that 'Citigroup' is an entity, and is clearly both a market maker and has clients, whereas you (and other professionals in the field) think 'Citigroup's equity option market making desk' and 'Citigroup's prime brokerage division' are entities, and are clearly only one of those things each.

People at a distance from organisations see them as monolithic. People closer to them see them as coalitions of distinct entities.


You can't front run as a market maker, because a market maker literally cannot have clients. Thats like the first rule about market makers in most regulatory environments. Saying otherwise excludes you as a person who knows anything about the details here.

A parent company could operate both a market maker and agency/broker businesses side by side. Leaking information from the agency/broker to the market maker operation is illegal.

Perhaps you think I am nitpicking here, but the point is the bad guys in OPs example are the broker/agency which illegally sell/give away their clients information. Their illegal behavior is not predicated on the existence of the market maker. The market maker is simply an easy bogeyman for people who don't know what the hell theyre talking about.


In the case that they also run the exchange where they are making the market, they most certainly do have clients.

This doesn't happen; you have no idea what you're talking about.

Are you still talking about crypto exchanges?

A lot of people, including the author of Flash Boys, would dispute how narrow that definition is. You don't need a client for there to be a systemic problem with trading on everyone else's advance information when the advance information is the buy order itself.

Wikipedia states, for instance: >Front running, also known as tailgating, is the prohibited practice of entering into an equity (stock) trade, option, futures contract, derivative, or security-based swap to capitalize on advance, nonpublic knowledge of a large pending transaction that will influence the price of the underlying security.

No client needed.


Follow the cite from the Wikipedia definition you just quoted and you'll find, in the linked paper, that it goes right back to Flash Boys.

The definition is probably cited in a lot of places because it's literally just the definition of insider trading. Which frontrunning is a form of.

A nice neutral primer on the HFT frontrunning debate can be found here, which contains a note regarding the HFT industry's attempt to advance the 'fiduciary' requirement. https://fas.org/sgp/crs/misc/R43608.pdf

Matt Levine, for instance, is very firmly on the 'keep frontrunning to only mean brokers interceeding before client orders' side of the semantic battle. Charlie Munger, by contrast, had no issues calling out the entire HFT industry as 'evil' and 'legalized front-running'. All this to say, there's clearly a lot of fight over the semantics here.

If you want to look at the official industry position, look at the communication between SIFMA and FINRA during the relevant time period wrt frontrunning regulations. You'll see SIFMA makes repeated reference to recognizing they have a duty in respect of customer orders, while the FINRA rules (5270, for instance) do not track this language. SIFMA attempted to conflate Rule 5270 (Front Running of Block Transactions) with Rule 5320 (Prohibition Against Trading Ahead of Customer Orders). I know a few people who have admitted in private that this strategy was not a mistake.

Rule 5270 notably has no exclusionary requirement that the non-public knowledge is in respect of a client trade. Rule 5270 also contains a definition of Publicly Available Information, which requires that the trade be COMPLETED prior to it being public.

To note how much of a gong show this situation was, note that in requests from SIFMA to delay the implementation of Rule 5270, FINRA said they'd grant the extension because firms needed time to train staff and retool systems - but FINRA explicitly noted that the activities prohibited under 5270 were almost all already illegal and covered by other FINRA rules.

So insider trading on non-public information is illegal and that's not news. Unless everyone does it with computers for a while. The regulator would look like an idiot for charging practically the entire industry, so instead they tell people they have a few years to knock off all that silly business while giving them a stern finger wagging.

But they only 'fixed' it for large orders, and then only in an official sense. The SEC hasn't done much in the way of enforcement on this point.


> So insider trading on non-public information is illegal and that's not news.

This is a frustratingly common misconception, but it is still incorrect. If you ascertain material nonpublic information without breaking a confidentiality agreement or fiduciary duty, you are fully allowed to trade on that research.

Price discovery is a core function of a financial market. Without information asymmetry, there exists no meaningful way to signal new prices to a market. In the abstract, the only way to achieve alpha is to have an informational advantage over the market consensus.

This is why the SEC is generally very careful in its terminology; to wit, you will usually see the terms "insider trading" and "illegal insider trading" distinguished. The "illegal" prefix is significant. Insider trading also refers more generally to the mundane, non-criminal trading activity of any corporate insiders. Legally trading with material nonpublic data is not insider trading by definition, because you can't legally trade on that kind of information as an insider.

Take a look at the SEC's spotlight of insider trading cases[1]. Note the commonalities - a confidentiality or agreement or fiduciary duty was broken directly (e.g. by an employee of the firm involved) or indirectly (e.g. by friends or family of the employee). Likewise, consider the SEC's definition of illegal insider trading[2].

As for what Munger has said about HFT - it frustrates me to hear that he's said those things, because of the coherent criticisms you can levy against the practice, frontrunning is not one of them. It's extraordinarily frustrating to me that you're characterizing this as something debatable - it isn't! "Frontrunning" has an established, precise definition which the SEC will happily and aggressively use for litigation - HFT or otherwise. Whatever Munger's problems with HFT are, his argument would be strictly better if he dropped the inaccurate jargon and simply criticized the practice directly.

In fact, I can't find a single citation in which Munger actually provides an argument for why HFT is bad or frontrunning. There are legitimate reasons you can be concerned about HFT, but Munger doesn't seem to mention them. Instead, every source I can find has him claiming HFT is wasteful or unproductive because he doesn't like the idea of people competing on nanoseconds. His overall dislike for the practice is extremely underspecified.

In contrast, Levine is actually far more neutral about the activity than you've given him credit for. He has written extensively about HFT, Flash Boys, IEX, the various players involved and the peripheral industries and incentives. He might not be as personally successful or wealthy as Munger, but his take on the subject is - objectively speaking - much more well-constructed and coherent.

___________________

1. https://www.sec.gov/spotlight/insidertrading/cases.shtml

2. https://www.sec.gov/fast-answers/answersinsiderhtm.html


That paragraph was meant to be glib, sorry if it wasn't precise enough. Money Stuff #2 has cratered before it really got wings.

That said, your response is half accurate. You're right that not all trading on non-public information is illegal, because price discovery is important (note, however, that in common parlance, 'insider trading' refers to illegal insider trading, not trading on non-public information). Anyways, regardless that's not the case here. First off, frontrunning does not help with price discovery. Second, trading ahead of a trade that's imminent and non-public is not legal. The opening subsection of rule 5270 is fairly straightforward:

"(a) No member or person associated with a member shall cause to be executed an order to buy or sell a security or a related financial instrument when such member or person associated with a member causing such order to be executed has material, non-public market information concerning an imminent block transaction in that security, a related financial instrument or a security underlying the related financial instrument prior to the time information concerning the block transaction has been made publicly available or has otherwise become stale or obsolete."

Recall, FINRA went to lengths to point out that this was ALREADY illegal by the time this section was put into force after the flash crash.

As for Munger, his views are fairly straightforward.

https://www.cnbc.com/video/2013/05/03/why-charlie-munger-thi...

https://www.cnbc.com/video/2014/05/05/hft-equivalent-of-rats...

"It does civilization no good at all. It's the equivalent of letting rats into a granary. I don't like it."

I think your note regarding Levine shows a point you've missed. I'm not saying Levine is for or against HFT activity. I'm saying he prefers a narrow interpretation of the term front-running like you do. However, even he capitulated, because both usages are acceptable. Hence my original note at the start of this chain where I note there's a dispute regarding the use of the term.

"In Money Stuff yesterday I said that "I seem to be losing the fight against semantic drift in the term 'front-running,'" which (I thought) used to mean a broker's breach of fiduciary duty by trading ahead of his customers, but which now seems to mean anyone's trading ahead of anyone else. Reader Bill Bremse pointed out by e-mail that some form of the second sense has been in respectable use for quite a long time, including by Larry Harris in the 1990s. I am not sure there is any higher authority on market-structure usage than Larry Harris, so I will cheerfully confess error on this one. "Front-running" can indeed mean something other than an illegal fiduciary violation, though at least Harris seems to have used it to mean trading ahead of exposed limit orders, not racing to trade on one exchange after seeing executions on the other. Though I don't know how much that was a thing in the 1990s."

A bit of an explanatory note: Prof. Larry Harris literally wrote the book on exchanges. He was also the Chief Economist for the SEC in the early 2000s, and later the head of their economic analysis unit.

Here's Order Exposure and Parasitic Traders, which he authored in 1997, well before HFT really took off. He refers to the activity we're discussing as front running, and explains systemic issues with the trading strategy before the HFT industry even existed: http://www-bcf.usc.edu/~lharris/ACROBAT/Exposure.pdf


Weird, I remember an earlier version of the Wikipedia article distinguishing frontrunning (client orders, you buy for yourself, you fill order) from tailgating (client orders, you fill order, you buy for yourself).

Edit: And that version claimed tailgating is legal. Looks like it was in this version: https://en.wikipedia.org/w/index.php?title=Front_running&old...


Yeah, not sure why they accepted that edit. Tailgating is something entirely different.

Sounds similar to what Robinhood does.

Since we're close to the topic, let's take a moment to all remember that Flash Boys is a crappy book that gets a lot of stuff really really wrong.

Read the rebuttal!

https://www.amazon.com/Flash-Boys-Insiders-Perspective-High-...


Additionally, I can personally vouch for Peter Kovac, the author of the rebuttal to Flash Boys: https://www.amazon.com/Flash-Boys-Insiders-Perspective-High-...

I can vouch for Peter because I worked with him for close to 5 years at Madison Tyler Trading / EWT, LLC (which later merged with Virtu Financial).

He’s a very talented technologist who build the real-time clearing and risk system along with an encyclopedic memory of financial regulations and nuance. He always stuck out to me as an “honest to a fault” kind of guy, and that is one of the reasons we got on so well, as I was never much for nuance.


Dark Pools is an excellent alternative, on a slightly different topic, although similar material is covered.

Strongly recommend Dark Pools for people interested in this topic from just a sort of history-of-technology perspective; it also covers a much longer span of history. Dark Pools also takes kind of a gimlet eye to algorithmic trading, in case anyone thinks the recommendation is because the author thinks the opposite thing to Lewis.

Michael Lewis is a former trader himself. I find it hard to believe he could get things that wrong.

Wasn't Lewis more of a pit bond trader? When I read Flash Boys, I read it with the understanding it was kind of like reading a book from a Horse Drawn Carriage salesmen after Henry Ford released the Model T. He has a bone to pick, and his bias oozes into every word of the book. Are some of his problems with electronic trading valid? Absolutely! Is he horribly biased and lets that show in his writing to make a more "shocking" story? Absolutely!

There used to be a finite number of spots in the trading pits. If you didn't get into the pit, you couldn't be in on the trading and potentially millions that were made. Now the trading went from the speed of human communication to the speed of electrons moving over fiber optics. However, instead of a limited spot on the floor, literally anyone can buy space in exchange co-located datacenters. You and I could chip in and buy some servers to put in the Weehawken, NJ datacenter used for NYSE and it would be affordable. Electronic trading brought a lot of changes, some good, and some bad, but it is arguably more "fair" now than it has been for a very long time.


Lewis was a coffee getter at Sollys by all accounts that ain't his. Good writer though. As long as you remember it's fiction.

He wasn't a trader. He was in investment banking, which is not the same thing. If anything he was more into sales and not trading.

Not really the same kind of trader, was he? Wasn't Lewis a bond salesman?

Bond salesman.

I'm worried about articles like this are because people want to try to regulate crypto currencies. I really like the idea of crypto being the wild west. If you use a cheap/crappy broker or exchange you're going to get a worse price than if you use a good (expensive?) one. Deal with it. If you want to deal with a heavily regulated currency use EUR/USD/JPY etc.

> I really like the idea of crypto being the wild west.

Some people liked the idea of the wild west being the wild west. But it turned out that a lawless frontier favored a few people over a lot of others, and just because something is beneficial to you doesn't mean it's equally beneficial to others, or to society at large.


Tell that to these US "helpful regulations" that prevent me from buying what I choose with my money.

Anything that's illegal to buy with fiat cash, it's illegal to buy with cryptocurrency, too.

Like what, a hitman?

It's inconvenient to move more than $10,000 at once.

There are good reasons to own a three necked flask other than making cocaine (although if you are an amateur chemist and actually buy one with bitcoin I'd be surprised if you didn't get in trouble.)

There's gray stuff like porn, although I don't necessarily support that one.


> It's inconvenient to move more than $10,000 at once.

I realize that this is subjective, and everyone's got their own acceptable risk levels, but I would feel very, very paranoid about trying to move $10k via cryptocurrency. This whole thread's about the Wild Wild West, and trying to move $10,000 in cryptocurrency is like tossing some gold bricks in your bag and taking a hike, instead of paying Western Union to ship it - insured - via one of their secured wagons.

(I'm also honestly not sure how convenient it is to convert $10k to bitcoin, and then convert it back. How long does it take for it to make it through the exchanges?)

> although if you are an amateur chemist and actually buy one with bitcoin I'd be surprised if you didn't get in trouble

And that's the whole point I was making elsewhere in this thread. Paying for something with Bitcoin doesn't magically make it legal.


Cryptocurrency will never achieve any meaningful status without regulation. We have anti money laundering and other financial laws for a reason.

Unfortunately this is the sacrifice that needs to be made if you want permissionless money. Regulation is not necessary.

Permissionless money isn't useful to the large majority of the world. It's a nice to have for rich people -- or "temporarily poor millionaires."

More useful is to ensure that governments have the rules that their populace wants. And even most useful is to ensure that the human beings who don't have money (for the multitude of unfair reasons that may cause such a situation) receive enough to survive.


Fortunately the populace wants regulated money.

Crypto advocates are speed running the history of financial systems. They’re getting pretty close to the “regulations might be nice” section of history.

Note: they can really only "frontrun" you if you are making a massive trade. Even then, from the other side it's not really perceived as frontrunning, rather as "getting out of the way" of a huge order that's going to change the price. You don't want to be the one that sold them some stock for $1 that they will immediately push to $10, which then you will have to buy back for $10, as you are the market maker...

It isn't just getting out of the way. HFTs see the big order and realize it is too big for one exchange, so they buy from the other exchanges before the rest of the order gets there.

Coincidentally, there is a paper and talk about this called:

Flash Boys 2.0: Frontrunning, Transaction Reordering, and Consensus Instability in Decentralized Exchanges

https://arxiv.org/abs/1904.05234v1 https://www.cbspringconference.org/


However, that's not quite front-running in the established sense.

That paper is about a side-effect based on the design of "Decentralized Exchanges" on Ethereum and similar systems. That's why people should always dive in the details of those blockchain solutions; not only you have implementation issues, but even design issues where the "solution" falls apart when game theory is accounted for all actors and interactions with the network.

Unfortunately, the complexity of such "decentralized solutions" makes it easy to miss a lot of issues. Even with audits that cost thousands of dollars per case, you still have several cases where serious vulnerabilities go unnoticed.


Here's a talk about this research that Philip Daian gave at Devcon 4:

https://slideslive.com/38911605/smart-contract-security-ince...

Probably my favourite talk from that conference.


Non issue. Of course. Just put in a limit order and be done with it. If you are really scrapping pennies in front of steam rollers you should find another gig because algos will eat you up.

There is no steamroller in this approach.

Perfectly exploited, you act as an infinitesimal holding-time middleman that has zero probability of not being able to sell, since the future queue is already known. You buy enough so that the real buyer has to pay the upper limit for their limit order every time they buy.

Imperfectly exploited, you still make a very reliable, very low risk profit.


The OP was referring to the person affected by the front-running, not the front-runners. "Picking up pennies in front of a steam-rollers" is from Taleb, and refers to short-term traders who buy/sell a large amount on margin (or backed by loans), and hold it for a very short period. Since the market tends to go up on more days than it goes down, on average these guys make money (before taxes and trading fees). The steam-roller is the black swan event, the large, unpredictable event that will drop the price faster than you can get out, causing you to lose more money than in the years leading up, and quite possibly bankrupting you due to margin calls or the loans. The OP is saying that the only way someone front-running you causes you problems is if you are doing many short-term trades, because otherwise the percentage you lose is too small to worry about. If the front-running loss (which is small) is a large percentage of your gain, then you must be picking up pennies, and if you are doing that, you are risking being flattened by Taleb's black swan steamrollers.

The only way someone front running causes you actual profit margin problems is if your profit margins rely on trading a large volume in a given time period, yes.

This essentially means you're profiting off an understanding of volatility. As Taleb showed, this requires care, since the standard deviation of a security cannot be reliably quantified from historical data, yes.

But there are almost always ways to trade large volumes on a prediction of volatility without running into Taleb's steamroller. The easiest first step is to never write a call option on a stock you don't already own (or short sell stocks in any other fashion). Then your losses are bounded since you can't go below zero dollars. Then you've just got to make sure you don't bust when you're left holding the busted security. This can be done with ordinary bankroll management like a poker player might use. But it's also what hedging is for. With hedging, you control the risk distribution of a given trade. All you do is simultaneously make other trades that boom in the case that the first trade busts.


I always thought that the phrase refers to doing a lot of small upside, but high probability of profit trades like selling far out of the money options for pennies. One bad trade is enough to wipe you out including hundreds of successful trades.

Any reference to flash boys is commonly an article looking for click bait. If you look at modern HFT, most of the profits have been eaten away among the competition, its no longer some hugely profitable industry. It is contracting. Sure a few firms are making millions, but its really not a big deal. Its really getting tired, most people love to get all up in arms about this, at this point high speed trading is run of the mill.

The same will happen to cryptos. There is always early profitability because theres no competition, but the profitability will erode.


I am not as familiar with the crypto trading industry but I did work in equity markets for a number of years and for a while on short term, high volume strategies. I still have a hard time trusting any article that uses Flash Boys. I still find that book to be one of the worst representations of the electronic markets. It paints this evil picture similar to a chemtrails kind of conspiracy where a single entity is causing harm. In reality it is a complex organism where each entity is try to gain an advantage.

Since a lot of comments are talking about the equity markets I figured I would add my thoughts.

There is definitely a need for regulation. No we don't need to get rid of electronic trading, frontrunners or algs but we do need to pay attention to whats happening. I remember that window of time where the NASDAQ was selling order information subseconds before it happened. This is bad. Or the exchange (forgot which one) that was creating special order types for just big clients that were unpublished, this is also illegal and was stopped.

I love frontrunning. People seem to forget the era pre computer market making where there was a guy that made the market for a specific stock. Or how slowly information traveled. People on the floor would know the moment something happened and be able to trade on it right away compared to everyone else. We had dollar size spreads. Today we have penny or even sub-penny level spreads. Sure someone is making money off of me when I make a tread, but their margin is A LOT less than what it once was. Is someone manipulating my order...probably not, I am too small of a fry. Is someone manipulating Vanguards orders, people are probably trying but I suspect based on literature that firms like Vanguard have put out...they overall like it, cost of trades are drastically lower than historically.

TLDR if someone has a way to trade across markets faster than the rest...thats great, they are serving a useful job in making prices across markets even and if they are doing it well they are earning their money for the risk and cost associated with it.

Edit: I highly recommend the book Dark Pools. Its a much more unbiased book about electronic markets and how they came to be. I feel that it paints a much more realistic picture compared to Lewis' hype.


Am I wrong, but don't all the crypto "whales" know each other, and can't they synchronize buying and selling? Essentially their bots all start buying together, wait for retail and other bots to join the uptick and then again all sell at the same time. Profit. Repeat.

I'm pretty sure they don't all know each other unless you're referring to fake ones used to scam users of centralized exchanges (in which case they're the same entity.)

How is this any different from algo trading on the stock market?

I read or watched somewhere where the algo traders would skim a tiny percentage of every transaction of mutual funds because they know exactly when to strike.

There are lots of sharks in the muddy water...


Algorithmic trading is about building an algo that determines how a market or stock will move based on news, economy, other factors.

In flash boys the front-running that these quant funds are running isn't about an algorithm, but literally all about having high capacity pipes that can see an order come in and then fill it between different exchanges and based on the minute differences in stock prices pocket the difference.

So an algo fund can lose money if the algorithm is incorrect. While front-running you never lose money.

It's basically like being a tax collector.


Great explanation! For an in depth and relatively formal (and mathematical) look at this from an economist I recommend https://faculty.chicagobooth.edu/eric.budish/research/HFT-Fr...

>all about having high capacity pipes that can see an order come in

I don't understand how they can see the order come in. Is this something that only happens in decentralized exchanges?


You can pay the NASDAQ (or any exchange) to colocate your server beside their server with a low-latency network link and see a live feed of the orders as they hit the book.

So isn't the problem with the exchange revealing data before it is "committed"? How does one frontrun if the exchange doesn't allow it?

I know about colocation, but I was under the impression that once you see an order in the book it is in the book. I.e. your order will be placed after it.

If there are multiple exchanges trading the same (or very similar) products, then you might be able to see an order on one, and quickly send an order to the other one. In the US, equities trade on multiple exchanges.

People paying taxes, in theory at least, get a benefit (to them or society).

Isn't it more like paying off mobsters, or just having someone steal a fraction of a penny from everyone?


Anyone can watch crypto going into the exchange. Usually people don’t (shouldn’t?) hold it on the exchange for excessive periods of time, so once it confirms, they try to sell and withdraw.

If you can see this, you can front the sell (before the deposit even confirms) and then be the buyer shortly after it confirms.

But I don’t see this as being much different than providing liquidity and taking the risk that someone else doesn’t want to take.


That depends on the underlying cryptocurrency, for example I think that there is no way to design exchange such that BTC deposit would be readily apparent on blockchain before it is credited to the exchange account. So it is simply another argument why "smart contracts" does not solve any useful problem.

Why would DEXes (Decentralized Exchanges) allow priority ordering by paying higher fees? Wouldn't they be aware that enables front-running?

I'm pretty sure the author was referring to "gas" on the Ethereum blockchain. The more you pay, the faster your transaction completes.

More info here: https://ethereum.stackexchange.com/questions/3/what-is-meant...


Surprised that nobody has taken the effort to explain "how" this is happening.

The article talks about flaws in DEX design, opening the risk of front running. What's happening is that DEXs work on chain, therefore orders are submitted on chain and thus visible in the mempool while waiting for miners to include it in a block. Front runners view these pending orders and then submit a similar order (but with a higher gas fee), thus incentivizing miners to take their Ethereum transaction instead of the earlier order (for higher fees).

While the article in question is correct with regards to what I explained above, I hate how generalizing it is. Not all DEXs are designed the same way. The first generation of DEXs did everything on chain which also has the downside of limited throughput (limited by blockchain transactions per second). The next generation DEXs are working to solve this with off chain solutions which would also solve the front running problem.


How would one go about becoming a flash boy, if it were still feasible?

By design.

[flagged]


Some people value other things above rationality.

What a crazy world it would be if people acted morally even when not being watched by "nanny-state big government"; such government might be entirely unnecessary then.


Word.

Financial system setup by those ideologically opposed to current regulatory regime suffers from the same problems the current regulatory regime was designed to mitigate. Report at 11.

Where are the people that told me that front-running is only used to describe your actions in response to your own client's orders?

Just because you were taught not to use it in email doesn't mean that certain forms of arbitrage are not front running.


Anyone who is a serious participant in the crypto markets knows they are nothing but a pure Wild West free for all. That is, like all other markets except without the pretense of regulations that purportedly protect the little guy. (But usually achieve the opposite.)

The point is that it's not a free-for-all. The biggest players completely manipulate the markets so that all the small fish end up losing to them.

That’s exactly what I would expect if a system was described as a “free for all”, the strong preying on the weak.



Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: