Dark Pools Turned Out to Be Really Murky | Hacker News Reader

Dark Pools Turned Out to Be Really Murky

bloombergview.com

47 pointstptacek21 comments

Instead, it used a quantitative system called "Liquidity Profiling," similar to Light Pool's "Alpha Formula" and Crossfinder's "alpha scoring." These systems were all basically efforts to answer the following question: When you buy stock, does the price go up or down? If it goes up -- in the next very short period, say one second -- then you have positive alpha. If it goes down, you have negative alpha. If you regularly have a lot of positive alpha, you are Bad, and people shouldn't want to trade with you. If you regularly have a lot of negative alpha, people will be delighted to trade with you. But then, if you regularly have a lot of negative alpha, you should stop trading because you keep losing money.

This is so awesome.

Matt Levine has quite a talent for describing markets in plain English.

Absolutely. This reminded me to subscribe his complete rss feed, before that I had just the 'Money Stuff' feed in my reader.

I think it as bit disingenous. The key part about the alpha measurement is this

> [...] in the next very short period, say one second [...]

Most "institutional investors" (which I guess means longer-term in the context of the article) would not be affected by this.

There's a paragraph in the article that explains that when such an investor places an order with a broker, it gets broken up into many small orders, which are indeed at high risk of being tagged as "too much positive alpha".

You're right that the article states, that the AES SOR would sometimes be (mis?) classified as oppotunistic - so the metric is not perfect. I'm actually inclined to believe that order send/cancellation rate (see analysis in eg. [1]) would be better.

But the entire paragraph (in my reading, at least) seems to be arguing, that you will loose money, if you have short-term negative alpha (footnote 8 actually retracts the argument again). This is obviously not true, as this would only leave room for the HFTs in the market.

[1] http://www.nanex.net/aqck2/4661.html

Awesome, but quite standard for major broker dealer banks. Any major market making broker dealer will have tech designed to protect itself from being on the wrong end of trades with better informed market participants. And those better informed market participants may well be market making HFT firms. See recent discussion about BARCs FX trading [1]. Larry Harris' class Trading and Exchanges has an excellent explanation of the microstructure dynamics of markets makers and informed traders [2], which I comment on here [3].

[1] http://www.bloombergview.com/articles/2015-11-18/barclays-to... [2] https://global.oup.com/academic/product/trading-and-exchange... [3] https://etrading.wordpress.com/2006/08/31/adverse-selection-...

I don't think this is as horrible as it sounds. It would be expected that a long-term investor would have a negative alpha (for the period of a few seconds), especially in a shallow market encouraged by low-latency traders.

For example, if you're buying 100k stocks on an exchange, chances are that you'll move the price. Since the price movement was "random" (i.e. not caused by any fundamental news about the company, but rather about your decision to execute that trade exactly in that moment), it would make sense for the price to revert back to the previous mean. But then, if you're focused on the long-term performance of your investments, you won't care - you know in advance that you will have to pay a fee for executing large trades (which is only fair, really), so you factor that in when deciding which stock to buy!

The prevailing attitude on Wall Street is that there is no such thing as predatory behavior. Or rather, that predation is part of the market ecosystem and to be viewed in a value-neutral light. People used to say that any offer to buy or sell provides a service to the market by adding information about price. But now everyone, not just expert traders, has to acknowledge that most modern operations that are mature in their use of math actually seek to add disinformation to the market. I would suggest a yet more sophisticated view is that multiple sources of randomness should be better for the market than fewer. As you can see superficially in the article about Light Pool, attempting to limit the entropy in this pool does not help make the pool more "fair" or "rational". It would be better, as in more sustainable, for dark pools and electronic markets as a whole to refrain from dictating what is an acceptable algorithm.

Many people read Flash Boys and applaud the notion of a rate-limited or otherwise restricted exchange platform that removes the advantage of network proximity for low latency info. I think we will find in another decade that exchanges built on this principle introduce a host of other inefficiencies, both time- and information- based arbitrage strategies that appear unfair to some market participants.

Yes - the limit order book mechanism was designed as an automation of open outcry. It was not designed to cope with illusory liquidity from eg quote stuffing or market makers quoting away from top of book then pulling orders when TOB trades.

Each trade is information. Trading can only come from disagreement, so each time you trade with someone, it's basically as if someone was screaming "you're stupid" in your face. Of course, you hope they're wrong and you're right (they're hoping the opposite).

So I see the issue here is one of integrity - claiming to enforce one policy when you're really enforcing a different one.

But why are High Frequency Traders "bad"? Like, do they significantly destabilize the platforms they use? Or is it just that they make profit, and thus the other users of the exchange are losing money? And if it's the latter, how is that not completely illogical? In order for you to make a profit when trading, somebody else needs to lose money. If you ban everyone who profits from the exchange, then there cannot be an exchange, right?

This is probably ECON 101 stuff right here - I probably sound clueless, and it's because I am. Please help this poor, confused lad out?

It's probably not right to classify all HFT as unscrupulous. However, consider the kind of algorithm that picks up when you're looking for a big amount of stock X, and then proceeds to race you to the other exchanges in an effort to buy from under you; it's pretty hard to justify that as anything other than predatory.

Making money from speculation (i.e. investing) is one thing, but gaming the system for profit is something else entirely, and something that quite inarguably does not add value.

Thanks for the great example. It definitely clears up some of my confusion.

I'm not an expert either. I have know two successful traders. By successful, I mean they made a living, and were employed. Both barely passed the series 7 exam.

They both used inside information when trading. Both were in denial. Both had a lot of money. A lot of it in the Cayman Islands. I still see a lot of inside information being thrown around. Hell, the only real money my deceased father made in the market was off an insider tip from his father in-law. My father asked me, "Should I do it?" Me, "Dad--you literally broke your back with honest construction work, and only certain members of your family appreciate what you put your body through. Yes--do it, and enjoy the money!"

That said, I'm a little jaded when it comes to stock trading. I basically believe most are cheating. Since I don't believe there are honest traders(not day traders who did well in bull markets), doesn't mean they don't exist?

Why should an honest trader pay more for a stock just because Computer that is physically close to one of the exchanges sees his order?

I don't want to argue with anyone. Preventing HFT trading, or at least the guys who are trying to literally sit on the exchanges seems like an easy fix? I don't know why they don't stop it?

> Why should an honest trader pay more for a stock just because Computer that is physically close to one of the exchanges sees his order?

It is not possible to see any traders orders before they execute on any exchange that I know of. Regardless of colocation, custom built asics etc. That people believe this is possible is largely down to a lack of understanding of how exchanges work and the fact that many powerful market participants (hedge funds and ibanks) are natural competitors to HFT firms.

> I don't want to argue with anyone.

I completely understand and commiserate, that said I hope it doesn't count as arguing when you clear up factual incorrectness rather than opinions.

> But why are High Frequency Traders "bad"? Like, do they significantly destabilize the platforms they use? Or is it just that they make profit, and thus the other users of the exchange are losing money? And if it's the latter, how is that not completely illogical? In order for you to make a profit when trading, somebody else needs to lose money. If you ban everyone who profits from the exchange, then there cannot be an exchange, right?

Theoretically you make your profit from long-term value investing. If you think FooCorp has a bright future ahead of them and I think their industry is in decline, I can sell you my FooCorp shares. And theoretically we've worked for any money we make: maybe I've done some research and uncovered a massive fraud at FooCorp. Maybe you're an activist who's going to talk to their management about a major restructuring plan[1].

You can run an exchange just with value investors like that, but you won't get a lot of trades - see the mention of Liquidnet in the article. Maybe you finish your plan a week before I finish mine. So maybe there's a middleman who'll sell you some FooCorp shares (going short), and then buy the same shares off me a week later, taking a cut for themselves. On one level they're providing a valuable service - but you can see why you and I, who did all this research effort, might hate some people who sit there with a computer being a middleman and raking in near-riskless profit. In the same way that we might hate a real-world middleman - a salesman or importer or the like. It becomes particularly acute if the middleman is an HFT selling shares that we "could have got on our own" five seconds later at a better price than their price (because they have an information advantage from e.g. direct feeds being faster than the SIP). In one sense they are still providing value, but it's a strange, counterintuitive kind of value. And of course people remember the times when they bought and the price went down a second later more than the times when they bought and the price went up, and it's easy to blame the HFTs for it.

[1] Why is this not just zero-sum? Because ultimately the market cap turns into real investment. If the market overall thinks that BarCorp has a great business plan and FooCorp is wasting money, then the price of BarCorp goes up and that of FooCorp goes down, and maybe BarCorp builds a new factory and FooCorp lays some people off. Assuming the market got it right, BarCorp's factory is more efficient than FooCorp's, so they build more useful widgets with the same money, and the world has more widgets in it than in the counterfactual universe where the companies' prices didn't change - real value creation

The article reminds me of the book Flash Boys by Michael Lewis. Lewis also describes what was going on (and maybe still is) at the Wall Street with dark pools, HFT and normal traders. Definitely worth the read.

EDIT: maybe I should've read the article to the end before commenting. But my recommendation for Flash Boys stands. It is a good book.

Many of us felt Flash Boys got it wrong; Levine is careful not to be too critical of his colleague but certainly gives that impression. (The "Shilling for the buy side" article that was posted here shortly after it came out covers much of the criticism).

Thanks for the suggestion of 'Shilling for the buy side', really interesting read besides the many insults. I especially like the many links for even more information. I try to gobble up as much as I can, the stock market is extremely fascinating and interesting. Of course Flash Boys is quite one sided and as it is with most things you should read it with a grain of salt. It's entertaining though.

I didn't know stock trading can be so interesting. It tickles my inner detective.

Reply on news.ycombinator.com