"The Great HFT Debate" is a must-watch, following the release of 'Flashboys' just seeing how defensive the market players got.
I don't dislike Flashboys because of its conclusions, I dislike it because of its inaccuracy. It called inot question every other book of his. Which was really disappointing as I had enjoyed them all so much.
^^ a blog post where I collect this stuff as I answer this question so often.
Quote:
> I think this is the crux of this story: according to Michael Lewis and Katsuyama, we’re supposed to trust people like Einhorn who have been convicted of insider trading, people who are suspected of insider trading (buy side is by definition rife with this; particularly firms that do merger arb and special events), J.P. Morgan, Goldman and Morgan Stanley: we’re supposed to trust these guys more than we’re supposed to trust a bunch of tiny little market making firms who had been inconveniencing them by taking away some of their flow. Lewis tries to make this seem like a battle between the underdog “good guys” and the evil establishment. To believe this, you’d have to believe that Goldman Sachs and people like Einhorn are underdogs, rather than the actual establishment. To believe this, you’d have to believe the tiny industry of HFT traders actually rules the world and buys off congressmen and the SEC more than … J.P. Morgan and Goldman. [...] Yeah, I might believe that. I might believe that if I were a dribbling retard.
doesn't really answer why Flash Boys does such a terrible job of explaining HFT
The whole premise of the David and Goliath story Lewis wants to tell is that HFT is a device used by the Wall Street establishment to screw over the rest of the market.
But in fact it's more the opposite: HFT firms are upstart prop trading and market making shops that are pissing off giant banks like Goldman and JPMC by making it harder for them to skim profits off block trades for pension and hedge funds.
Also: the use of rhetoric like "dribbling retard" isn't what makes an argument "ad hominem". An ad hominem argument is one whose logic rests on a claim about its opponent and doesn't account for its opponent's argument. The logic behind the quote you're commenting on doesn't do that.
i do appreciate your explanation for how the goal of Lewis's book is based on a fallacy
The argument being made here is about Katsuyama's complaint that the prices he'd see in his terminal would suddenly move away from him if he tried to trade large blocks of stock at that price.
What makes Katsuyama sound incompetent is his apparent lack of understanding of the price pressure his own large block trades put on the market. Lewis makes Katsuyama out to sound as if he can reasonably expect that if FCOJ is quoted 29.78-29.80, he can sell 100,000 shares at the quoted price. Of course, not only can he not expect to do that, but the influence that large trade has on the market is the whole reason Katsuyama is paid by RBC to trade for its clients.
Hence: either Lewis doesn't understand his subject well enough to report on it, or Katsuyama doesn't understand the basic function of his job.
In any case: not an ad-hominem, sorry. There's a popular Internet fallacy that says any argument that uses insulting language must be an ad-hominem. Insults aren't usually productive, but they don't automatically produce flawed arguments.
I guess calling him an idiot is not ad-hominem in that he's arguing he doesn't understand what he's talking about.
A) I agreed with your last comment
B) was just pointing out that the snarkyness comes across strong in the blog post. So it's possible that frame of mind affects the writing and it affects the reader, argument aside.
Maybe I should change the post to be more explicit about that.
http://www.amazon.com/Flash-Boys-Insiders-Perspective-High-F...
First, people have long had questions about the societal effects and costs of trading, even when humans did it. Second, the automation of trading isn't merely replicating what humans do at a lower cost. I'd call it a fundamentally different activity. Third, one of the things that has become more and more attenuated is any sense of moral responsibility. Traders were never known for attacks of the warm fuzzies, but algorithms not only don't give a shit about how the affect humans, they can't. And fourth, instead of saving everybody money, finance eats up an ever-increasing share of the economy, and I think automation is part of that.
Automation isn't universally good. Sometimes it makes things better, sometimes worse. As technologists our job isn't just to automate everything. It's to use our professional skills and experience to improve the world, to improve human lives.
By a lot of measures, the financial industry captures a lot more money than they did in previous decades. But it's far from clear that they're delivering a lot more value, and given that computers have drastically reduced their costs, it's reasonable to expect them to be smaller, not larger. As a start, this has some good graphs:
http://esoltas.blogspot.com/2013/02/5-more-graphs-on-finance...
And here are a couple more general-audience articles:
http://blogs.reuters.com/felix-salmon/2011/03/30/chart-of-th...
http://blogs.wsj.com/economics/2011/12/10/number-of-the-week...
Google "odd eighths scandal" for a good starting point.
† "Front running" has a specific technical meaning, which very few people who throw the term around seem to know about: a front-runner violates a fiduciary duty they have to a client, trading against their clients for their own benefit. Market makers and prop trading firms aren't generally brokers for other people and don't have that duty; they can't "front-run" the people they out-trade.
Your real estate agent would "front run" you if they knew you wanted to buy a particular house, knew your maximum price, bought the same house for less than that price and then sold it to you at a profit. Same with your stock broker and, say, IBM stock; they have a duty of best execution to you.
By your analogy, the real estate agent knows you want to buy 5 houses, and suddenly after closing the first one, the price of the remaining 4 goes up.
If your own real estate agent bought a bunch of houses to resell them to you at a profit, they would be front running you. If another real estate firm noticed that you seemed intent on buying 5 houses and bought a bunch of them to resell to you, that would not be front-running; in fact, that would basically be a description of how the real estate market works.
I'm not saying it's your own broker or agent that is biding up the remainder of the lots you are intending to buy. It is "observers" who are manipulating the system in a way to make you reveal to them your intentions and price, so they can "front run" you for the next transaction a millisecond later at the next exchange.
And the reason I quoted observers was becuase in my limited understanding (if i had the book in front of me I would re-read this), the HFT traders would put up some token offers up on all the stocks for the purpose of exposing potential activity that they could abuse. A quick google of this suggests that this is called "Pinging".
For example: a commercial real estate developer may be putting together "an assemblage", which takes a bunch of disparate properties and tetris-style assembles them into a contiguous landmass to hold e.g. a larger commercial property with a signed anchor tenant and the plans for more. The developer has important, market-moving knowledge about the near-term prospects for properties composing that assemblage: they are all about to become sharply more valuable precisely because the developer has a use for them. The developer will send out someone who is isomorphic to my dad to discuss with the owners of each property, partially in serial and partially in parallel, to get them to sell to the developer. Their goal is to get the properties as cheap as possible and they are under no obligation to tip their hand as to why they want them.
It is highly, highly in the developer's interest to conclude all purchases before either a) the property owners or b) anyone else realizes what is going on. This is complicated by many factors, including e.g. planning commission approvals required in some places which are public records. A fairly common strategy among savvy local real estate investors is to a) identify assemblages before they are assembled and b) race the people putting them together.
"They bought the corner gas station? It's useless without Milly's house. Have they closed Milly yet? Quick way to check. calls Milly Milly, I was wondering, haven't you considered being closer to your children? Where did they live again, Florida? Yeah, yeah. Interested in selling your house if I give you a fair offer? Well, we can talk it over any time, but $300k cash and I can close immediately. Sure sure, let's chat."
Real estate developers hate, hate, hate when it happens, because sometimes the market maker here has correctly intuited "This deal doesn't happen without the property which was previously occupied by Milly, right? Man, that would be a shame. $700k." "That's an outrage. It's worth $250k." "You were certainly going to tell Milly that,which is why I gave her $300k. But, between businessmen, it's not worth $250k or $300k. It's worth whatever number your client is willing to authorize to get the assemblage done." "$600k you dirty dog." "$650" "DONE." "Pleasure doing business with you."
Now naturally, sometimes the market maker guesses wrong. He now is holding a house. That's the nature of the business.
This maps fairly directly to financial markets, except the "assemblages" are called "block trades", they're composed of fungible units rather than individually distinguishable properties, and this happens much faster and much cheaper because market makers are very, very good at their jobs.
You can imagine that the neighbor next door to Milly might notice the fact of Milly's transaction (e.g. via hearing directly) or the market maker's transaction (e.g. via public records). This is analogous to the tape being painted in a lit exchange. "Ooh spiffy, my property is more valuable than I thought it was!" The real estate developer hates this as much as Brad Katsuyama hates when someone spills the beans that he's putting a 100k trade together by painting the tapes with buy orders.
The real estate developer, and Katsuyama, should learn to deal with their disappointments or get better at executing on their only job.
Arbitrage doesn't really cover it because people were buying at the original price and may not buy at the marked up price.
Where it really is wire tapping, I would say the right term might be (high speed) industrial espionage.
And that they were likely a party to. They would place small orders on all the stocks so that they could detect activity that they could exploit at the next exchange a millisecond down the wire.