This page has been archived and commenting is disabled.

Haim Bodek's Presentation To TradeTech On HFT And His Controversial Findings

CalibratedConfidence's picture





 

HFT and Rule 610, here we go.  At Trade Tech this year, Haim Bodek of Decimus Capital shared his presentation on Reg NMS and how new regulations are going to change everything about high frequency trading.  ZeroHedge has a stash coverage of on Dark Pools and the story Haim has from just a typical automated trader to a market structure reform activist.  For those familiar with the depths of ZH, you may understand most of what is presented, though watching the entire 43+ minute presentation is highly suggested.  For those looking for compact video's on HFT also from Haim, see here, or get a book to read.

I won't even try to add any long-winded paragraphs.  Just watch this.  And one thing of note is at 25:45, the "speed" we keep hearing about on the the uniformed mainstream and slideshow websites, isn't the problem, it's who is smart enough to know what order type to send and when to send it.  Full Transcript below video.  Hope you all enjoy the weekend. 

From TradeTechWest:

 

Transcript:

So, to start out with, I think I’ll just tell you what I am hitting. I am hitting the basic High Frequency Trading strategy that’s been dominant in U.S. equities since an inflection point starting in early, or in late 2006, early 2007. Then I am going to explain how HFT in U.S. equities and regulation NMS are intimately intertwined. I am going to explain some of the properties of HFT-oriented special order types that give HFTs an advantage over institutions. I am going to hit on some of the landmines that you can encounter within order-matching engines, and then I am going to give a forecast on, at least from my perspective, what’s going to happen over the course of 2013 because most of things I am discussing here are huge topics in the industry and are also undergoing significant regulatory scrutiny.

So, scalping strategies, primary features, and you need to understand these features generally so you can understand how the order types REG and NMS and everything else plays into them. There is a connection between the features of this class of Algorithm Trading Strategy and what the exchanges have done in terms of evolution of market structure. The basic property is high frequency turnover, this in early 2000s, if you told me that I could get in and out of a trade in 10 seconds and I could get 5% of a name, I would think you were crazy, but that’s the goal of this type of strategy. The fundamental property that I think most of the quantitative property trading firms outside of HFT missed is the critical importance of queue position within your strategy itself, and I am going to explain that probably a little bit more detail in a couple slides here.

Of course, most low latency/high frequency oriented strategies are going to have precise and timely reactions to changing market microstructure events. We also have this class of strategy, especially maybe 2004-2005 starts leveraging special order types and order matching engine features, so you start seeing innovations in the mid-2000s that gets more and more complex, and many of those features were basically evolved to accommodate the needs of high frequency traders, specific firms in fact. This type of strategy compresses quite quickly, so the subsidy of rebate is kind of critical to this class of strategy, and we have post-only orders which control the rebate costs and make sure you aren’t exposed to taker fees, and then the last thing I think is very important to understand is that these firms in this class of strategy consider high frequency turnover, and exiting risk is basically an insurance against adverse selection and toxic order flow, so the idea is that they can’t forecast what’s going to happen in the future, and US Lit markets are quite toxic, They are going to rapidly try to get out of all of inventory, and that’s, you hear of these firms that trade 10% of the market and they are flat end of day, that’s connected to that concept. And so it's a very important concept because when you have half of the US Market have extremely low risk tolerance, you get derivative effects in the market itself, such as the mini flash crash. I’ll explain how that happens.

Next, these are just general bullet points on the strategy. You require price time priority market structure to exploit the queue position, so these guys have to be fast because they are just trying to get to the top of the order book. What’s important to understand is there are only certain moments in time when you can get to the top of the order book, which is why speed matters, and precision matters. These strategies are not joining a better offer and slowly rising up the queue. They are trying to set the market and get to the top first, and that can only be done, it's so precise, it’s repeatedly microseconds. So, low latency is queue for superior queue position. You get 5 or 6 firms in the market that have this class of algorithm, it compresses quite rapidly down to an effective zero-width market. What does that mean in practice? It means you buy at the bid, you are not trying to scalp out for a penny, you are trying to exit at that bid price, and that's … you kind of think of that as your best case scenario. And that would make a lot of sense if you think about the exhaust flow and the institutional order flow that’s constantly hitting the U.S. market, and trying to access liquidity, so this is basically a flip-out strategy, and if go you look through the data, especially after you learn some of these concepts, you can easily see in tick data, pretty much everything I am talking about there, but you kind of need a framework to understand why it’s happening, and that’s what I am trying to explain.

So, let’s talk about queue position, this is completely naïve, dumb explanation, but I think it’s compelling. Let’s say I have a naïve strategy, I have no view of the future, I am in a market that’s basically normal distribution of events with a completely neutral, random order flow, if I get to the top of the queue, and HFT’s care mostly about a one-tick winner or loser. The view of what’s going to happen beyond that is less important than the concept that if I make a penny or I make a one-tick winner market to market, I can get out, and I can lock in a profit. So that's kind of the key idea, the whole space of the future compresses into up-tick, down-tick. If I am at the top of the queue, I am likely to trade most of my one-tick winners, if I am top of the queue, it’s unlikely that the next tick is going to clear out, let’s say I am on the bid, is going to clear out the entire offer and move away from me. So I am going to get to trade a significant portion of my one-tick winners if I am at top of queue.

Let’s say I am resting there, keep in mind I am also going to trade all of my one-tick losers, right? This is random. This is nothing about information, right? I am resting, market goes against me, I trade everything. Market goes the other way; I trade some of the good trades, okay? That’s kind of an explanation on why spread and rebate subsidy is so important, because a naïve joining algorithm is expected loss, in let’s say a very very small tick size. Rebate and tick size will basically make a naïve strategy profitable. However, you know, this would only work if there was only one firm doing it, so I’ll explain later in the presentation what happens, how does this work when there are six firms and I will explain later, but the other situation, and think about a program trading desk, that's passively or electronic trading desk is what we are calling them now, passively joining offers, They are at the back of the queue. What’s going to happen, they are unlikely to trade, especially in a name like BAC, where they are way at the bottom, I think they have 70,000 shares ahead of them. They are unlikely to trade any one-tick winners, and they trade all one-tick losers. So, if you’re at the bottom of the queue, you hemorrhage. You bleed to death very quickly. Okay?

The other thing, which I will … now that I explained that it’s all about shifting the distribution of trade events, when that toxic trade comes in that’s going to sweep against me, there is another concept embedded in this, to try to detect when that’s happening and to cancel. You know, you hear everything about HFT is canceling all the time, cancelation is extremely important. There is just an opportunity cost. If I get out, I can afford to cancel 20-30 times, to get out of the way that big landmine coming in, I have a lot of false positives, lot of false alerts, then I can get out of the way, and it only cost me market share. So there is an emphasis on avoiding negative alpha, and by avoiding, that’s kind of a market making mindset. If I avoid that toxic flow coming in, my distribution is going to bias towards positive mean in terms of my PNL. So this concept of sweep risk, where a large order, that’s … you, guys’ maybe, comes in and takes out the whole price point that is bane of this whole strategy. That strategy will lose money on that trade if it trades against it. What is going to happen, usually, is when a sweep is detected; you try to cancel, as it’s sweeping the market. Let’s say I am on one market, I get hit there, I get off the other market, that’s why you see a rapid withdrawal liquidity, and what I am going to try to do is either hit some dummy behind me, let’s say I traded, I am going to hit the dummy behind me, and think about an inverted market like, you know, like what's it Ajay, right? I can hit a dummy behind me who is posting on Ajay, exit that bad trade, which I got rebate for, and I actually make a net profit on my bad trades if I manage sweep risk correctly, but what that does, since everybody does it in the HFT space, you know, the top five-six firms are all responding to the exact conditions, They are responding to each other, I’ll give you a very clear thing. If someone’s at the top of the queue and that's fast guy, and he is always at the top of the queue, and he cancels, well they can see that over the price feed, and they are like “Whoa, I am going to get out of the way, something’s coming in” so there is this concept of looking at the electronic crowd and trying to forecast that a sweep risk is coming in. Run away, and that, if there are no other algorithmic trading traditions in size in that name, that will cause a cascade effect, and in some cases, you’ll get the mini flash crash when there is basically no algorithmic strategy in there that can correct prices that are chargeable from kind of say a statistical standpoint. So if there isn’t a stat to dive in there to trade that bad price, you are going to get a mini crashes, a crash is when 6-7 participants immediately flip out of the market.

As I said, go back and look at your tick data and BAC, you’ll see this is how the name trades. So let’s talk about regulation NMS, this formulated in 2005, established in 2007, a series of initiatives designed to modernize and strengthen the national market system for equities securities. The term HFT was not widely used until REG NMS came out and specific firms had enormous change in their profitability. For me, everyone’s like “What’s HFT?” To me, HFT, in U.S. Equities is a class of strategies run by around 10 firms that no one heard of prior to 2007 and they were basically the market leaders after 2007. If you were making a lot of money in the early 2000s, and there are some firms that out there have high-frequency trading desks, it’s not the same type of high frequency trading that is dominating U.S. Equity volumes right now. So, the important thing to understand is the adoption of REG NMS coincides with a period of intense HFT volume growth. So, think about what’s going on as REG NMS is being negotiated and rolled out from 2005 to 2007. HFTs are starting to get more volume, they got some special order types, they are making $20 million net a year, and the banks are still ignoring them because that’s not an interesting enough number, right? But this legislation, this regulation’s coming in and it’s going to change the whole game. Every single aspect of price movement, what is permissible, what is not permissible, what trades can happen, and also the moments in time when I can get to the top of the queue, and what case … what might impact might be of cancelling orders, all of that is effected by REG NMS, the whole game changes. So this is kind of a crisis point for HFT, and certain individuals or firms were very involved in trying to assert or proactively argue for certain features within REG NMS. As you can see, these are the bullet points:
• Getting to the top
• Cancelling ahead of sweeps.
• Who’s the maker, who’s the taker?
• What signals can I get over the price feed to detect adverse fills?
• And the aspect of what speed advantages, where does speed help me and where does it not? I want microstructure conditions.
All of it is impacted by REG NMS.

I am going to give you two very specific examples:
Rule 610 of REG NMS banned black markets. A trading center must prohibit its members from engaging in a pattern or practice of displaying quotations that lock or cross any protective quotation in an NMS stock. What does that mean? There is a two-tick wide market, I could jump in the middle and be the best bid or offer, but I would likely lock, you know, have my price be the same price as an opposing bid or offer on an another market. Well, that’s inadmissible according to REG NMS. It was a critical constraint, that one piece there, Rule 610 completely changes the Top of Queue game, so it’s a critical constraint on Top of Queue strategies, now what are exchanges doing in that situation? They are actively changing your price and rebooking you, and I’ll talk about that later. They are not allowed to accept your order into a lock, so in other words, the legislation says you are not allowed to send that order and get to the top of the queue in precisely the conditions where an HFT would want to do that. Right? So just remember, the basic default behavior of these exchanges is to not let you post the aggressive price. They will change price on you, and some exchanges, which I think is kind of comical, will actually require that you call up the Helpdesk and ask for a port setting change, so that your confirmations coming back actually indicate that your price was changed, so you can’t even look at your fills and know whether or not, sorry, your confirmations and know whether or not your order was adjusted to comply with Rule 610 unless you call this particular exchange and have them like change some database codes, so this, a lot of people will look at data and say “It looks okay to me” and I say “Well, that’s actually not correct data, right?” You need to know exactly every single exchange in the U.S. has some variant of these practices, so if you’re not on top of the various specific features of each one, then your order is basically being treated to some kind of default behavior, and I would argue that some exchanges, the Default is tolerable, and in others, it’s completely egregious what happens to your order. That’s one of the reasons I became a market reform advocate.

Rule 611 mandated trade-through protections. A trade center must vent trade-through of protected quotations and NMS stock. Well, that also creates an important constraint. It tells you what prices are allowed to trade or not. What does that mean in practice? I don’t know, for IOCs that are marketable in fast market conditions, 30-40% in certain names might be rejected. You are not getting rejected because of the liquidity wasn’t there, you are getting rejected because you violated rule 611, you just have to look at the code on your reject, and you realize you were trying to hit that order, it was there, you missed it because of rule 611, well, guess what happens? If there is a slow SIP feed, think … the regulatory feed, think what NYSE just got fined for. If there is just one exchange that is slow in pumping bad data to that SIP, what’s going to happen is you are not going to be able to trade prices that should be trade-able, right and you are going to get immense adverse selection. Well guess what? Modern HFT has gotten around all of these inconveniences of REG NMS, right. So that’s the story, is that REG NMS completely screwed up the game, and they found clever ways in the exchanges created innovations to basically circumvent the entire regulatory framework in all of the areas that impacted HFT needs.
All right, so what happens after 2007? Well, the exchanges go on a complete innovation binge, they create a number of special order types, you know, the press says 2000, no; there is like 15 order types that matter. The 2000 order types, the way I think about it, is that’s the forest. You have to know what trees you’re going after to basically protect yourself or take advantage of the features I am explaining. There is also order modifiers, really quite interesting fields that you can set on your orders that basically dictate very precise order handling treatment during locked markets for example. And then lastly, new port settings. That means you have to call up the exchange, and know that there is a database code that if I set it this way, I get to treated the way an HFT would prefer. Again, all this stuff is not the default order handling of the U.S. Equity markets.

So what I am arguing here and here is the truth, without these features, HFT goes back to 2005-2006, and looks like a curiosity again, right. This are the features that supercharges the entire industry. The ‘secret sauce’ of HFTs often amounts to little more than the application of these innovations in their intended usage scenarios, so that’s my beef, right. It’s very clear once you are an insider what the purpose and intent and intended usage scenarios are of these features, they are all connected to precise microsecond events, right? The order type, if you just use it randomly doesn’t do anything for you. You have to actually use it in the moments that matter and that is not properly disclosed by exchanges, but over the last 18 months, there has been significant change because the fact that these features were not disclosed adequately and the usage scenarios of those features is something that regulators have been looking at for 18 months now.

Okay, special order types, let’s talk about some. There is actually a ton, but we are going to talk about just some of the cases here that might be more easy to understand is probably the best reason I am picking these. Okay, on locked and crossed market, exchanges employ a number of price sliding behaviors to comply with the prohibition against locked markets. First you got to understand what price sliding is, I hinted at it earlier, price sliding is an exchange is allowed to change, they would argue that you made the decision because you used that order type which is that price sliding treatment, or just no matter how you look at it, the exchange is adjusting your price based on logic that they are running, and shifting your order around in the order book. Rebooking you basically, that’s what price sliding is. And they are doing that when? When the price is moving. When you are locking away market, when there is a moment where you could get to the top of the queue. Basically, in the most advantage situations where you can get an edge over the market, that’s when the exchange is changing your price, and through this process of rebooking, they are also changing the time you execute.

So the way that one exchange explained it is that they have the flexibility in the sensible and convenient booking of orders, so they are doing you a favor. Their other choice is to kick your order back. Well, guess what, that’s an HFT mode. Don’t handle my order like that, kick me back. You can actually choose that as a port setting on one exchange, okay, so that’s how it’s described in documentation in one exchange. The thing to realize about price sliding is it occurs in this two-tick market condition that I explained, which is, and what I say here is “Note that all exchanges, according to REG NMS, in theory, must display a two-tick or greater gap before a one-tick price move and NBBO is permissible.”

So people ask me, “How frequent is this?” I am like this is how every price tick in the U.S. Equity market operates. This is the mechanics, the quantum mechanics of price movement, this is happening for, that is why HFT-oriented scalping methods that use these features are, in my estimate is true HFT, probably … I don’t even know now because it’s been on the wane, but at peak, I am thinking like 40% is my personal estimate of what that volume was, but you know, that volume is coming in, to these 2-tick wide markets, exploiting order matching engine practices like price sliding, and special order types to achieve certain goals, such as a superior queue position. So, the point here, at the bottom, is that HFTs are not waiting for that 2-tick wide market across every exchange. They are not waiting of those 13 exchanges to fade before they place their bets. All of the special order types in a class related to rule 610 are all oriented to placing your bets before the price move changes.

Okay, so price sliding is a key element in exchange order handling that governs what we call lighting a new NBBO price. Okay? That’s the moment in time when a new price is permitted to change the regulatory NBBO. Price sliding impacts that significantly. It also determines which orders are placed in favorable or unfavorable queue positions. It determines when and what price orders are booked at, and it thus defines eligibility for maker rebates and taker fees. That’s the thing I think is still lost on people is that these practices are intimately connected to who gets the rebate, and I think people generally understand that the HFTs are trying to get that rebate, and there is a reason for that, okay?

So I will probably make … I just want to make one more point there, the conditions in which HFTs are targeting to make rebates are also favorable trades. They don’t want to get a rebate and then have a 1-tick loser, so it’s actually a positive alpha in those conditions. It’s not just about the fee; it’s about doing the right trade. So, price sliding effectively picks winners and losers in the context of HFT scalping strategies. You could argue all day long “Well, I did it at this price, and the price five minutes later, it wasn’t really a winner under that five minute mark. I am explaining that it’s all about whether or not you could have gotten a better price in that very short interval. And I would argue that if you use HFT oriented features, you will get a better price.

So, again, I want to harp on this, for five years I had to listen to this whole narrative about everything being fast, my story is, I was led to believe it was a speed issue, and I spent about a year building out six ________ [00:25:54] and forcing my team to get faster and faster and faster, what I didn’t realize is by not leveraging the special order type features and knowing about the, what I would argue would be undocumented price sliding practices, by not knowing that, I ended up jumping into these bad, disadvantaged situations more frequently, so that’s … basically there is … in this dark hole is the story of me hunting for this bug, well I was basically building a system that would jump into what I would argue are order matching engine abuses. And I was becoming the first dummy to get in there and bleed, I was ahead of everyone. So it's kind of really quite paradoxical experience, I am getting faster and faster and faster and this thing is hemorrhaging now, and there was a moment in time where my partner was just yelling at me, “How can you lose…” we are trying to do 15,000 shares in SPY, “How can you lose 2 pennies in SPY? It’s a penny wide!” Price sliding and queue jumping, and these features have come up in the Wall Street Journal and I am explaining now. But that’s what was happening, and the story is I was finally introduced to the features by one exchange, so then as I went through the process, I learned that basically all exchanges had in some manner had to address these issues to attract HFT order flow.

So, the last thing is, speed matters, but don’t put your money into speed. It matters only if you know what order types to send and when to send it, because low latency trading only matters to the degree traders act upon the same events. If you are not competing against someone else, if they are not pulling your quote while you are trying to hit them, or you are two guys trying to go for the same queue position, you have got 200 milliseconds to make that trade; you don’t need any speed investment at all. So, you need to conjoin speed with HFT-oriented knowledge, period.

So let’s talk about two of them: Well, hide in light orders, which there wasn’t really a name so I just picked that out of a hat and started using it, various names on the different exchanges. They basically circumvent the adverse impact on the ban on locked markets, using clever regulatory workarounds. So, another way to say it is HFTs hated price sliding, they wanted another way to get into the center and lock markets, and that’s what these order types do. They are kind of have your cake and eat it too order types, because they exploit regulatory kind workaround where they get to be a hidden order when that’s convenient, and they get to be a what's called a protected quotation when that’s convenient. They are like a hybrid order. So those orders are quite useful, and there is also Rule 611.

Rule 611 actually prevented a type of order called an ISO, which was meant for large traders to sweep across the order book, well what happens is you can leverage the exemptions of ISOs, and you can leverage fast price feeds to basically place orders in trade at prices that you can justify. Like I knew NBBO better than the regulatory feed, I am ahead of everybody, so you can use this low latency speed advantage to actually trade at prices that the public customer is trying to trade at and getting rejected, so those are two key classes, there are a lot of other interesting things that happen, some of them are market maker orders that can discriminate effectively, I don’t actually have much of a problem with those, but there is also mid-point orders. Some mid-point orders are oriented towards institutions, others are oriented towards HFTs, you have to know which ones are HFT ones on which exchanges.

So these two order types, what do they do? They provide a shortcut to the top of the queue and they avoid price sliding, not all price sliding is bad, just certain types of price sliding that HFTs would consider bad, these order types tend to avoid. Okay, one exchange explained the purpose of the orders was “It eliminates the need for traders to re-trade multiple times in rapid succession, trying to be high in priority at the next NBBO price. Doesn’t that sound like HFT scalping? So when you read in the press that these NBBO order prices were generally available for everyone, and they weren’t catering a specific class of participant, well, most electronic trading desks at banks are not trying to rapidly re-try orders multiple times in rapid succession, trying to be high in priority at the next NBBO price. So, these are HFT-oriented orders, and I will just maintain that until the end.

With the orders that hide or light, orders that lock the NBBO are hidden instead of rejected or price slid, and see these orders at the bottom? Those are actually your limit orders, right. You know, I get some criticism, it’s not what you might expect, HFTs say to me, offline, I have e-mail chains when I became public on this, that you know, they would never use regular order types, are you crazy? Right? These! Of course you have to use these! So I love it, like HFTs for a very long time were like “That’s the bread and butter order types of the U.S. Market.” You show that list to another crowd, they are like “What is that? That’s Greek.” Okay?

So, let’s talk about order matching engine practices. The complex sequence of microstructure events tied to price moves dictate HFT profitability. So, as I learned at trader machines exchange order handling is now a primary factor impacting execution performance and slippage. We went, you know, there was a group at my firm that was trying to address the slippage problem by forecasting the market, you know, certain leans, and what you could get out of that was nothing compared to what you could get out of using the order types correctly. So HFT special order types avoid and in fact benefit from disadvantages that frequently impact traditional order types, so I only have five minutes here, so I think I am only going to do two more slides, so I am going to explain again the dark side of REG NMS price sliding. Orders can be disadvantaged by being kicked to the back of the queue, orders can be promoted to the front of the queue, maker/taker fee allocation can be attributed in an opaque and unusual manner. Certain innovations that were introduced, I am thinking about one in July 2011, creates unexpected collisions between sophisticated participants and unsophisticated participants that otherwise would not have occurred. The impact on price sliding on order type precedence is not properly disclosed by exchanges. Okay, I am going to give you an example of queue jumping here.

When I show this slide to any institutional crowd they are doing large orders, they are just like “Oh my God! They get it,” okay? So we are going to show how an institution improperly accesses one exchange here. Assume the market for XYZ is 20.01, 20.02 offered 1000 up, institution wants to send a non-routable order to buy 10,000 for 20.02, that order is also a locked away market, but it’s marketable on this exchange, right? Now, again, they say “well you shouldn’t have sent a non-routable order!” If you know how electronic trading desks and banks operate, they don’t like to pay exchanges for routing fees, so they send non-routable orders. The institution buys only 1,000, right? That’s all that was there, at $20.02. And wants to pay $20.02 for another 9,000. But he used the default order handling treatment and he was price slid to $20.01, when he locked the away market. He wasn’t allowed to post at $20.02 after he cleaned out that 1,000. The trading firm, let’s say that you were an HFT on the sidelines, there was a neutral market, there was no information at all. The trading firm sees the 1,000 trade at $20.02, so he sees this 2-tick wide gap, which should be considered to jump in the middle, and he sees a new price slid bid for 9,000, so he sees the offer cleared out, and 9,000 on the bid, now you have an asymmetric market, 2-ticks wide, 10,000 on the bid. Could you have, do you understand how self-insurance works, and the fact that this guy is just trying to flip out at the same price? You know, getting in front of that 10,000 bid on a 2-tick wide market is a quite obvious condition to consider.

So, the trading firm then steps ahead of the institution by posting a “Hide & Light” hidden order, he locks the market at 20.02 ahead of the institutional order, and this complies with Rule 610. Meanwhile, a phantom order, a completely unreal order is displayed at 20.01 on the price feed, and that looks like it’s seemingly behind the institutional order. How did you get there? Well actually the order was ahead of you, but the price feed said the order was behind you. Okay? So, if the institution was smart enough to look into the screen to see what was going on, he would think, according to the feed, that there was order behind him, not ahead of him, because the hidden piece is ahead. Okay, so let’s so the away market unlocks and now the 20.02 bid is now permissible to light up, well, the Hide & Light order is rebooked, that’s very important, the Hide & Light order doesn’t just to sit there and have its cake and eat it too, it has to be rebooked. As a matter of fact, the phantom order that represented it is also pulled off the tape, so the Hide & Light order is rebooked as well as the institution’s original order. They are both pulled off the book and then put at the same price 20.02 but in this particular exchange they decided that they were going to do a swap-a-roo and put the Hide & Light order ahead of the institution but if you think about … if you think about the ethics here, that institution as the full information content, created the condition where the Hide & Light order could jump in, right, he is expecting because he sent an aggressive order, he is expecting to set that new price but this Hide & Light is going to go ahead of him. Well, guess what, it's not just one Hide & Light order that's going to go ahead of him, it's going to be the whole crowd of HFTs, okay. If there are six guys, he could have couple of thousand shares or more ahead of him when the market finally lights, so he sits there how am I going to getting it done, you know, What's going on? What's going on? I can't get it done, I can't get it done, this market is so toxic. If you just use the right order types and understood they are already handling treatment of that exchange, it wouldn't be toxic at all, actually what would you do? You jump in front of this guy. Okay, we are almost done I guess.

So in this particular example, I said that the lighting, that's important to understand, it's the lighting process when the market moves, okay. Lighting unfairly advantages HFTs, you know, and so I explained this all but the exchange on fairly lit HFT oriented orders ahead of orders commonly used by institutions … because I am running out of time here. Well, that's not all, that's just the only one that's public, okay, there are actually five major classes of order what I call order-matching engine abuse, these types.

Queue Jumping – the unfair order handling practices that permit HFTs to step ahead of investor orders, it's not just Hide & Light orders, that phenomenon applies to other order types and by the way it's different on every exchange.

Cannon fodder –the unfair rebooking and repositioning of investor orders that permit HFTs to flip pout of toxic trades. Okay, you could debate every some of these and say that's okay, or that was the intent. But when you look at the collective and you look at how each one operates on every exchange you get a quite asymmetric market where the public customer is egregiously disadvantaged, okay.

Fish Food – this is the one that really got me actually, the unfair conversion of investor orders eligible for maker rebates into unfavorable executions incurring taker fees. What does that mean? Well, when you can get rebooked and have the order slid around, you can basically be on the book eligible for rebate and then in some very strange interesting situation, just be pulled off the book, dropped back on like a market order you paid the taker fee, and guess what you executed exactly the wrong time in that case. That's … that was … if anything was my bug, it was actually that one.

Bait & Switch – the unfair insertion of HFT intermediaries in between legitimate customer to customer matching.

Catch Me If You Can – the unfair and discriminatory order handling of investor order during sudden price movements.

I won't say much more now but, you know, from my perspective the article on me, in from a page of Wall Street journal in September that was just one of like 30 things on my list that irritated me, but again that was one thing that the Wall Street Journal fell and they had a hard time doing it, felt that they could explain to the investing public. So, we are done with that. How much time do we have? We are done?

Male Speaker 1: We have another 5 minutes.

Haim Bodek: Okay. So, industry forecast of 2013 in one sentence, that things are changing and have been changing for at least 12 months but most of these things here are not well understood by the greater investment community, so lot of things that really irritated me that have been cleaned up probably will never to light, but some will. That’s it.

Male Speaker 1: Just like to point out that Haim has a number of pieces posted on the tab forum talking about these very topics, would you suggest they all be read?

Haim Bodek: Yeah, actually I collected everything into a little book called the problem of HFT. I was concerned that some of the arguments were being diluted, being spread across, and I have some new articles in there, so it basically gives the entire blue print. If you want to do your homework I gave you the Cliff notes. It is 100 pages but it is the Cliff notes for this whole thing. There is going to be a think outside, I mean these little bookmarks, I don’t have books today, but there are bookmarks which are kind of glossy and nice, you can pick them up and they will tell you a little bit about the book and also some…it will point you to my webpage. Also, I think yesterday or day before, I am forgetting, the Tab forum did put a video up where I discussed…

Male Speaker 1: Yesterday. It was put up yesterday.

Haim Bodek: I disused this section here. So if you want to get my industry forecast it is kind of embedded in that, so 12 minute video.

Male Speaker 1: You want to take one question?

Haim Bodek: Sure.

Male Speaker 1: Does anybody have one question? No? Oh okay, right in the middle there.

Female Speaker: You are focusing a lot on the passive strategy of HFT. My question is how many strategies you believe are strictly passive versus active or other types of strategies that are not based around that scalping as you described.

Haim Bodek: Well I don’t think it is passive at all. I think that’s what everyone is…I think people don’t understand that these are aggressive collision oriented strategies. They are trying to come in and execute against someone in an effective mid. I mean a lot of the trades they do, if you actually looked at it at the time of execution it would like they were giving up edge. They are really ahead of the crowd. They are not passively joining the crowd at all.

Female Speaker: But is the purpose of this the primary rebate capture?

Haim Bodek: The purpose of it is to trade a significant amount of volume and this is the mechanism by which it is done, it’s not…think about HFT oriented…these strategies existed in futures market which do not have rebate capture. This is just a class of high frequency, it is aggressive, it is opportunistic, it is in effect predatory because it gets ahead of winds, right, it is not a passive market making strategy at all. And that’s actually why these features are actually should be quite useful for liquidity secrets to employ if they were well know.

Female Speaker: Thank you.

Male Speaker 1: Okay, well, thank you very much.

Haim Bodek: Thanks.

 


- advertisements -

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Sat, 06/15/2013 - 18:06 | Link to Comment Mediocritas
Mediocritas's picture

It's a shame how few people seem to comment on nanex and calibrated confidence threads. Makes me think that either readers don't understand what's going on or don't grasp the significance of it. Every time a trade house reports a perfect quarter, that's code for superannuation contributors having their life savings stolen from, to the tune of billions of dollars per quarter. Where's the popular outrage?

Sat, 06/15/2013 - 04:53 | Link to Comment BidnessMan
BidnessMan's picture

The only way to "win" in this game is to not play.  Off to my next boating accident...

Sat, 06/15/2013 - 02:48 | Link to Comment Mediocritas
Mediocritas's picture

The example Haim gave...that was goddamned me to a tee. Well, they won in the end, I quit as I could obviously see I didn't have the tools to compete. I'm not alone because in every contest, once participants realize the result is rigged, they quit playing and then the season goes to shit.

So we end up in a world where it's just bots gaming bots for sub-pennies with no bigger picture in mind, leading to prices and quotes no longer have any economic meaning. That's another thing that really shits me about this whole situation, the IP theft, if I can call it that.

I might spend weeks researching an issue to come up with what I consider a sound quote but then when I try to get a fill at that price, my "unsophisticated" quoting would see a bunch of bots simply hijack the quote and get filled with me, the originator, somehow last in line. At least guys like Haim now explain the mechanism of how it happens, but fuck it, this shit completely breaks what a market is supposed to be about, prices cannot reflect reality without many human participants placing informed quotes without being jacked.

Exchanges shouldn't be privatized, period, all it does is create a moral hazard to break the game. Like allowing sports teams to bid for the ref.

Fri, 06/14/2013 - 21:59 | Link to Comment Abbie Normal
Abbie Normal's picture

So it's like cheating at poker by being able to see your opponent's cards, except that if you don't have a winning hand, you can also take from your opponent's cards until you win.  In other words, there is no way for them to ever win or even to fold.

Sat, 06/15/2013 - 21:23 | Link to Comment CalibratedConfidence
CalibratedConfidence's picture

Add the dimension that most of the players who "can't win" are willfully ignorant and that "to win" one must educate themselves to take advatage of the disclocations caused by the structure

Fri, 06/14/2013 - 21:21 | Link to Comment ToNYC
ToNYC's picture

Thousands and thousands of words to describe the microstructure of absolutely fraudulent price orders that could be reduced to de minimus by establishing a 1 second cancellation fee and a matching a non -proprietary margin audit engine. You get the market results you desire, losers give compulsively and winners choke on easy money. It's a beuautiful thing in the matrix you play.

Fri, 06/14/2013 - 21:02 | Link to Comment Downtoolong
Downtoolong's picture

it is aggressive, it is opportunistic, it is in effect predatory because it gets ahead of winds, right, it is not a passive market making strategy at all.

In other words, in the world of HFT we have eliminated the true market makers and specialists. All we have left are computer algo equivalents of floor traders trying to scalp everyone at the speed of light. If that isn’t bad enough, the volume of activity they generate isn’t true liquidity either.

Nice. Every time I place a market order for a trade these days I apologize and salute it first. It’s like sending it on a mission through a mine field onto the beaches of Normandy.

Sat, 06/15/2013 - 18:22 | Link to Comment Mediocritas
Mediocritas's picture

When I send a market order I feel like the fat kid in a game of chasey, lumbering around trying to catch a quote, being made fun of by nimble little rats who pretend they don't see the fat kid coming, then slip out of reach when he does. And filling any sizeable order feels like an old hippo being taken down by a pride of lions, one bite to the backbone at a time, impossible to turn quickly enough to fend off all of them.

Do NOT follow this link or you will be banned from the site!