CQS has had issues in the past. You should know this.
More importantly, why wouldn't you invest in colocations and collect the data yourself using direct feeds (with GPS clock synchronization etc to validate the data)?
We do. You are grossly misinformed. The charts correctly show the sequence and times of this event. I'm not going to engage this discussion further, though pmail is fine.
As a bystander who does not have the background necessary to evaluate either of your claims at face value, I would really like to hear your explanation for why you disagree with his explanation.
A simple analogy: you are an advertiser using Facebook. You tell them you want to get 1000 impressions, and facebook gives you a report at the end of the day saying that they gave the number of impressions you paid for.
So you look at yesterday's view count and today's view count and notice that somehow the view count only increased by 900. Something is afoot!
Nanex's argument is tantamount to saying "that means we must have lost 100 organic views today"
My argument is tantamount to saying "I actually bothered to look at our access logs (which we record on our servers) and only saw 900 that we could definitively attribute to real Facebook users. Is it possible that the report is incorrect or falsified?"
Back to the current situation. There are many sources of market data. Each individual exchange generates its own feed, and with the major exchanges (NYSE, NASDAQ, BATS, DirectEDGE, ...) you can colocate in the exchange data centers (NYSE and ARCA are in Mahwah NJ, NASDAQ is in Carteret NJ, and various other exchanges are located in New Jersey and Chicago) and record the data yourself. There is a unified tape (CQS/CTS) which combines and disseminates a combined record (across all exchanges). This is used to determine the "national best bid/offer" -- the prices people are willing to buy/sell at.
The process of CQS generation is fraught with problems, but lots of older traders and academic types use CQS data because its much cheaper to get that data than to get data from individual exchanges directly. However, you are subject to the quirks of the combination process, including subtleties regarding timestamping data (since this data includes trades and quotes from Chicago and from New Jersey, the sequence of events may appear different if you record from chicago or new york or philadelphia or some other place; if you ask the exchanges to timestamp directly, you have to worry about clock delay and skew between the exchanges' servers).
nanex is saying that it is acceptable to depend on that data and any anomalies must have occurred outside of the recording process. I am saying that the recording process can create the types of anomalies that nanex is showing, and that the only way to be sure is to record the data directly and carefully synchronize your recording machines. AND when you do that you see that there really is no anomaly.
Just to emphasize how sloppy the exchanges are with regards to timing: on the BATS exchange they use multiple servers to run trades and generate quotes, and every once in a while you see messages appear to be out of time order because the individual machines weren't properly synchronized (although, if you filter for a single ticker, messages are always in chronological order)
There's quote feed solutions that do co-location at the exchange servers and deliver the quotes to you in individual channels (ARCA/BATS/EDGX-A etc) and also SIAC feeds.
That looks like a hardware product. You still need to purchase market data access to the relevant exchanges to get that data, and those are expensive (NASDAQ costs, for example, run upwards of 20K/mo to get the lowest-latency data)
They should just spring for the data before making accusations -- the problem is that when you cry wolf all the time no one will take them seriously when a real case comes around.
Honest question: How are timestamps different from any other user-supplied data? Is the timeline of events ever recorded using an unsynchronized clock for relative comparison?
hmm, its not only the activity just before the official announcement, but also, just after it. There is no way possible that people can just make make a trade 'after' hearing a report 100-200 milliseconds after the report.
The trading before the official announcement is a concrete proof, if this is official, and bug/error free. But then again, this is not my forte.
Using "unsigned ints" for values that will never be negative allows you to perform one validation test instead of two. Use a typedef to avoid writing "unsigned" everywhere and you end up with less clutter in the code (for humans) and in the binary (for machines).
I am Eric Hunsader from Nanex and created this animation using our own custom software tools and our NxCore data feed.
Our position on HFT can be summed up in the first lines of text below the image: "It's not high frequency trading (HFT) that concerns us. It's high frequency quoting". Links are included for details. Our latest paper on HFT can be found here: http://www.nanex.net/aqck2/3532.html
Anyone who has taken the time to really look at the contents of the tsunami of data that some HFT creates will come to the conclusion that something very wrong and harmful going on. We've found this is nearly impossible to convey in text or graphics (with the possible exception of this animation), and sooner or later as more and more academics get up to speed, they will generally come to the same conclusions we have.
Our primary business is providing a real-time (and historic ) data service for U.S. Stocks, Options and Futures. In the course of monitoring our feed for our subscribers, we run across anomalies that we think need to be published for the public good or long term health of our markets. We have not received a dime from any of our analysis and not a week goes by that we don't regret opening that Pandora's box.
I will check back later today to answer any questions. If you don't receive a reply for a few hours, it's because we are working at our "day job".
You don't like high message rates because it makes your job harder. You guys are a great data provider (I used to be a customer) and you're doing a pretty good job (if not somewhat biased, perhaps unconsciously so) of informing folks about the happenings of the market. However, message rate limits are already managed by exchanges, what more would you ask?
Theories related to quote stuffing to cause ticker plants to stall and other gaming related items just don't jive, to be honest. It's always been the responsibility of the trader to ensure they can manage the data flow. If they can't they can get a vendor who can help them.
Data rates will naturally manage themselves and will always grow towards available bandwidth. It's the nature of our market, especially as we move towards tighter spreads. The markets are becoming more and more continuous and that necessitates the need to quote at high rates.
The charts you guys put together are interesting but let's be clear: no human looks at data at these levels, only machines. And machines don't care about visual patterns in data flow. So, as a means to help illustrate how markets are evolving I'm all for the visualization. However, as a means to denigrate the very real needs of important market participants to remain competitive, I think you're doing a disservice to the uninformed reader of your reports.
Your data shows the increase in HFQ relative to much lesser increase in HFT well, but neither of the links I followed to why HFQ is bad gave me much of an idea, other than "it's extra network traffic".
Exactly. Nanex does not like high message rates because it increases the cost of running their service. like yummyfajitas and yourself no one can explain to me why high-rate quoting is bad. Producing pretty graphs showing quotes at regular intervals doing interesting things is not indicative of bad behavior, especially when many of those charts include after market data when the depth is extremely thin. And remember, every venue enforces message rate limits already.
The link in the original post labeled "high frequency quoting" led to a page [1] that makes a more concrete claim. Basically they say that the NBBO (national best bid and offer) are being manipulated to be small when there is no trading and larger when there is trading.
I think that's at least a little disingenuous though...
If there is no trading then it makes sense that bids will be increased and offers decreased in order to 'entice' trades. No trading means the bid/offer are too low/high - basic econ 101.
As soon as there is trading two things happen:
1) bids and offers are hit, meaning they are removed, leaving lower/higher bids/offers as the next best.
2) The HFT algorithms know that when there is a lot of trading they should lower bids and raise offers, because excess demand indicates the bid/offer are lower/higher than they need to be. Essentially, they are realizing that they are leaving money on the table.
This all might look like price manipulation to an outsider, but to anyone that knows what is going on it's just the way markets work. The difference is that it happens a lot slower in markets humans are used to.
In my mind what would indicate a problem is if bids and offers widen AHEAD of trading. This would mean the HFTs are finding out that somebody wants to trade and adjusting their quotes BEFOREHAND. That's front-running and illegal. But I don't see evidence of that here.
I'd be curious to know what you believe the harm is. Specifically, since you claim HFT harms long term investors, what are the mechanics by which it does so?
One thing the site seems to do is keep an occasional catalog of patterns they believe to be evidence of attempts to harm price discovery in the markets through various quote-system-gaming tricks, presumably intended to set up some other strategy. Here's one, an algorithm that seems (in their opinion) designed to target a particular stock and jack up its volatility: http://www.nxcoreapi.com/aqck/3271.html
Hi Eric - Firstly, thanks for the work that you do. I can always rely on nanex research to produce thoughtful and interesting discussions.
Many of the "bots" in your research seem to be adding quotes that are far away from NBBO. One theory is that these exist to provide market participants with more precise measures of their own latency - almost like echo location. They send out a pattern to the exchange, and then wait until they see it coming back on their market data feed. I think that its good for market participants to have an accurate means of measuring market data latency and that having that means ultimately enhances the efficiency of the market. Do you disagree with this stance? Is there some other better means of measuring latency that would be better for market participants to use?
The problem isn't trading fast, it's trading unfairly. Not following regulations. Heaping the costs of speed on others, while reaping the benefits. We have documented hundreds of examples of clear rule violations. http://www.nanex.net/aqck/aqckIndex.html
Did you know that in 1999, during the internet bull market, the system processed just 1000 quotes/second, and today it's 1.5 million/second? Guess who pays for that? Guess who benefits? http://www.nanex.net/aqck2/3528.html
Did you know that your orders to ETrade and other retail brokers are tagged as "dumb" and sold to the highest bidder like Knight, UBS, Citadel, who then internally match your order (it never goes to the exchange, except during the flash crash). They give you the price from the slower system, but buy/sell from the faster one. The incentive to skew the two is irresistible. http://www.nanex.net/aqck2/3519.html
When you get a price-improvement of a penny on a 100 share $60,000 apple trade, they steal 99 cents - and front run some other investor, who's order is left hanging (sometimes forever). It happens in Apple more than 2,000 times a day. http://www.nanex.net/aqck2/3520.html
When you offer to buy something at a price, then yank that price before the other party can see it (speed of light), is that fair? It happens millions of times a day. http://www.nanex.net/Research/bloodbot/bloodbot.html
Great to have nanex here. I am very interested in hearing why you (and surprising number of other vendors) support windows, but not unix. I don't mean that in an accusatory sense, I'm genuinely curious if there are that many people who run automated trading models on a windows machine.
Regarding the topic on hand. When e-trade sells their order flow, aren't they obligated to provide 'best execution?' You are speaking of rebate for liquidity, correct? If NITE and BATS are both quoting the same price, then e-trade has the choice of sending to either, but if NITE has better price, they e-trade has no choice but to route the order to NITE, correct?
Rebate for liquidity is entirely different. So is retail price improvement (aka sub-penny pricing).
Retail order flow is sold to a wholesaler like Nite who matches it internally. They are allowed (best execution) to give you the price at the NBBO during a 1 second window. For less active stocks, the NBBO price can change 1,000/second or more. For all stocks, the NBBO price will be slower than the direct feed price. This latency and differential makes it a risk free trade (so long as you don't accidentally release your test code into production: see http://www.nanex.net/aqck2/3525.html)
There is no way to prove what prices existed on all markets at the time your trade was executed, meaning it's on the honor system. So this depends on your trust in Wall Street to give you the best deal.
Even if it does narrow spreads, it doesn't matter at the margin. If the spread is 5 pennies on average without HFT, and 1 penny with it, investors are losing out. HFT is charging them a penny to remove a penny of randomness. If someone offered you a a game where you could either:
A) flip nickles: heads you keep get the nickle, tails your opponent gets it
B) flip pennies: heads you keep the penny 98% of the time, 2% of the time some other third party gets to keep it (an HFT outfit), your opponent gets it
I'm curious why you think I should have to "prove it" if willing buyers are trading with willing sellers on an exchange they chose to trade on. Seems pretty anti-market for a finance company.
Then why did BATS not reschedule their IPO again the next day? First impressions, regardless of how illogical they maybe, play a significant role in group behavior.
"Even if their had been a glitch?" Really? 3x the open was postponed, followed by 17 seconds of no quotes/trades on ANY stock from Nasdaq, followed by a crossed quote from same exchange, followed by 3+ hours of no quote in FB from the listed exchange.
