martes, 26 octubre, 2021

By Mike Bellafiore
I was leaving work late on a Friday night when my Blackberry buzzed with this message: “How come you all are using people to trade instead of computers? Aren’t you kinda stuck in the Stone Age here?”

High-frequency trading (HFT) is all the rage in my business and beyond. It is the execution of stock transactions by powerful computers and algorithms at speeds faster than you can blink your eyes. But as blogger Kid Dynamite lampooned, “Was HFT responsible for Obama winning the Nobel Peace Prize? Can the dominance of the Yankees be explained by HFT? No.”

New York Senator Chuck Schumer, with information fed to him by the head blogger at Zero Hedge, generated enough noise to move the SEC to propose a rule change. Even the mainstream media got in on the act. The New York Times and The Wall Street Journal both penned lengthy articles on the topic.
With all the talk about whether HFT adds anything of societal value other than the ubiquitous liquidity, where does this leave us as discretionary traders? With more and more computer programs being designed to co-opt our money one nickel at a time, how do we adapt?


For background, I have been actively trading my own equities account, mostly intraday, since 1997. Every one to two years since then, there has been a drastic series of events: the Asian financial crisis, the dot-com bubble, decimalization, 9/11, the integration of NYSE floor trading with electronic execution, subprime and the recent near collapse of the financial system.

Out of necessity, I had to adapt my trading style every time the nature of the market changed. I started as a relative strength tech trader, picked up momentum trading during the Internet boom, mastered bounce trading in 2001, taught myself how to fade stocks during the slow-paced 2003-’06 years, shifted to uptrending plays for the next year, and then spent 2008 whacking bids and playing the downside momentum in overleveraged financials.

At each point of my career, I witnessed veteran traders drop out of the business, unable to adapt. With each new technological development, members of our community gave new excuses for not making money. With more than half of U.S. equities volume now controlled by algorithmic trading, HFT might be my biggest challenge yet. I stand ready to overcome it.

I empathize with frustrated traders in the trading ring, ducking the thrown elbows of HFT. On the second trading day of November, I lost my cool momentarily and called out some HFT programs that had out-traded me in American International Group, Inc. (AIG). “[Expletive] chiseler,” I exclaimed. This felt good temporarily but did not improve my profits. I went to bid or offer for 1,000 shares in AIG, and I got cut by a machine. This is all the byproduct of high-frequency trading. And I was getting cut for half a penny!

Below are seven effects, so far as I can see, that HFT has had on U.S. equity markets. Each is then followed with a suggestion or two about how we can adapt.

Average daily volume has skyrocketed 164 percent since 2005, according to NYSE data. Up to 73 percent of this volume is handled by algorithmic trading programs. One unpleasant side effect is that stocks touch more price points and may cause you to be stopped out of excellent positions.
Adapt: Use better stops. The best way to describe this solution is to talk about a trade I had in Mohawk Industries (MHK) Oct. 30 (see Figure 1).
The stock was weak, and the intraday downtrend was heading for a collision with its long-term support of $44. That level held the bid and then finally dropped. Naturally, I took a short position. Overall, the market was weak—check. MHK was now below the $44 level—check. Intraday, the stock was in a downtrend—check. Most of the intraday volume occurred at $44.25, and the sellers won—check.
MHK kept trading lower, and I added to my short position. Price at $43.70 offered some support but then dropped. I expected a serious down move to follow. But out of nowhere, while watching the prints, I saw significant buying at $43.68. What? The stock was broken, and now even this dopey $43.70 level had cracked. Who the heck was buying $43.68?
This was just another example of what I call the buy-the-new-low program. Algo programs buy almost every new low because they bet that short-term shorts will get trapped and have to cover higher. I saw this and had the option to cover but did not. I loved all the checks in my favor for my longer-term trade. I just hung in until the quick upward move subsided.
So I set my stop at $44.11, confident that it would not be triggered. I was wrong. The buy-the-new-low program drove MHK above $44; I was stopped out. MHK continued up to $44.25. But not two hours later, the stock was trading almost three points lower.
These are the challenges that we face as traders. High-frequency trading causes stocks to hit more prices, even prices they probably should not, and hunt our stops.
So when these bots trigger stops, refocus on the tape (see next paragraph) and trade with both hands on the wheel. Re-enter your trade if you suspect you were artificially stopped out.
Adapt: Learn how to read the tape. Few retail and independent traders have developed the trading skills to read the tape. Reading the tape is the craft of watching Level II quotes and being able to determine a stock’s direction based on order flow. Although use of charts and technical analysis are important, they will only take you so far. It is harder to read the tape because of HFT, but it is a critical skill to develop nonetheless.
Dr. Brett N. Steenbarger, a noted trading psychologist, writes that active traders must combat HFT with a “keen reading” of order flow. I agree. Let me offer an example (see Figure 2).
At 10:45ish, my eyes were starting to twitch from more than an hour of staring at constantly changing red and green across four trading screens. My stomach growled as my early morning protein bar had left me hungry. I thought that a stop to the men’s room might be a good idea.
But this was a key market moment, so I placed my physical desires aside. At $42.30, I spotted a huge battle on the tape in Capital One Financial (COF). My charts confuse each other, like a Democrat talking to a Republican about health care reform, because COF was right at the trend line. This is where I had to stay focused and catch the next move.
So what happened? There was an unusual hold on the offer at $42.30. I entered a short position in multiple lots. The price lifted; I flipped and was now long multiple lots. COF did not explode, so I sold.
The tape signaled that programs were creating a false breakout because the $42.37 offer was not cleared as quickly as it should have been. So the play was to sell and get flat. Less knowledgeable traders did not see the signals from the tape and ended up hitting a large position much lower than $42. Reading the tape saved me thousands of dollars.

Certain trading plays are less effective now that algorithms exist. For example, it was once a high-percentage play to buy a stock when it broke out, making a significant new high. You would get long and play the momentum to the upside. For me, this trade was like a Friday night with my best girl, a bucket of popcorn and a thriller about to start. Today, it rarely works. Meet the sell-the-new-high program turned on by high-frequency traders with more money and bullying power than you.
Lately, when a stock makes an important new high, it will inch even higher until a seller finally refuses to lift for higher ground. Immediately, the sell-the-new-high program will sell much lower than you bought. The longs are trapped and squeezed out of their positions.
Adapt: Change your entries. Shorting a new intraday low for a stock in a downtrend almost never works now, perhaps only if the Volatility Index (VIX) clears 45 again. So you must make the adjustment to short such a stock into an up move. But these stocks also tend to trade just a little higher than you would expect because so many more market participants are playing games.
So today, not only does an old-school, following-the-trend play not work, but the adjustment you have to make must account for the increase in high-frequency trading.
It is certainly challenging not to hit a stock that touches a new significant low when this pattern has worked for the previous decade. You have to fight that instinct. You may have to short, cover, stay with the trade and then short again before catching the move.
As such, your win rate may decrease, your risk-to-reward ratio may be less favorable and your losses before you catch the resumption of the downtrend might grow. Then when the stock starts to head lower again, it will take more time to do so. Like Starbucks, more bots are around every corner.
In this case, HFTs are using a passive sell algorithm, so your stock may not look like it is turning as it had in the past. This may cause you to doubt whether this is just a natural up move, which offers an excellent shorting opportunity, or whether the stock is ready to reverse. The key is to recognize the trading pattern and not get shaken because a few shorts did not work out initially and then hold for the move to the new low.

When I began trading and there was a big buy order, say 50,000 shares, the buyer just cleared the offers until the order was filled. Today, we see much better buying.
Almost all big hedge funds or mutual funds possess passive algorithmic programs that can slowly fill an order, 100 shares at a time, without making any noise. The result is a slower up move.
And after an uptrend has been created, these buyers may just turn off their algorithms, let a stock drop and then restart the passive algorithms. Because of this, momentum trading has become less profitable at times and more difficult. (Again, a high VIX could change this.)
Adapt: Focus on buying into pullbacks. Wait for strong stocks—those that are trending up on an intraday basis—to pull back before you enter positions.

False buy and sell signals may lead you to place trades that are nothing more than programs gone wild. On select days, HFT programs can go at each other like punks in a bar fight. A stock may triple its volume on a given day, close above previous resistance and signal a long to you. But remember, this may just be the bots doing their thing.
Adapt: Watch the Level II quotes to confirm information. Before entering a position, make sure you see real buyers and sellers on the tape.

As high-frequency trading programs chisel your bids and offers, you have to recognize the situation. See my AIG rant at the beginning of the article.
Adapt: Place your orders at the prices you want and go make a sandwich. Don’t get lured into a match you cannot win against HFTs. Play your game on your terms or sit on your hands.

Slippage is greater due to HFT, which seems contradictory to all that supposed new liquidity in the market.
Say you are long 1,000 shares and seek to exit a stock if it trades below an arbitrary price of $50.50 with a stop order at $50.49. The stock touches $50.49, and you expect this to be your fill price with some slight slippage to $50.48 and $50.47. Only it isn’t. It’s 100 shares at $50.48, 100 at $50.47, 200 at $50.45, 100 at $50.43, 100 at $50.41, 200 at $50.38 and, finally, 200 shares at $50.36.
Doing the math, that is an extra $63 out of your pocket that you were not expecting. And that was just one trade!
The programs are so fast and sensitive that they drop the bids for lower prices when they sense a sell order. Where you may have exited the whole thing for no lower than $50.48 before, HFT has now made you take another trip to the ATM after its latest pickpocket.
Adapt: Reduce your manual stops and enter limit orders. Limit orders give you the control over your exit price, minimizing the likelihood of slippage. Check out Dendreon Corporation (DNDN) on April 28, 2009, for further motivation.

One trader at my firm always chirps the following: “Beware of the V move.”
Trading 101 teaches us that when stocks break from an important level that they should continue in the direction of the break. For example, I was trading Visa (V) in October, and $77 was an important intraday level. Visa consolidated near $77 but then broke to the downside quickly trading to $76.
The fundamental trading play is to short: If V pops after it has failed at $77, then you short. Generally, if Visa trades down to $76 and jumps to $76.50, this is an opportunity to add to a short position. I saw what I needed, and the sellers beat the buyers at $77. Visa should have trended lower.
Well not so much in this market with these programs and in this trade. As Figure 3 demonstrates, Visa went right back to the battle area of $77. The V pattern can happen from time to time, but the frequency with which it occurs in today’s market has grown. When some of the HFT strategies are done cutting each other after this technical breakdown, then the stock just goes back to where it started. It makes no sense, but it’s reality.
Adapt: Be mentally agile while you trade. When your stock is trending, consider the possibility that a V move may be about to make an unwelcome visit.

High-frequency trading is here to stay, though flash trading and dark pools are on a fast track to being regulated. I have never made a dime as a trader complaining or mocking another market participant. Although some changes to HFT are welcomed, the interested parties—Congressional leaders and regulators—are in the room. Offer your comments but then move on to your job—trading.
Trading is a choice, not a right. HFT is challenging but nothing to fear. Adapt to overcome these new obstacles and become a better trader. Hey, maybe one day your trading system will become so proficient that you will become a high-frequency trader.
Mike Bellafiore is co-founder and managing partner of SMB Capital, a New York-based proprietary trading firm. His book, One Good Trade: Inside the Highly Competitive World of Proprietary Trading, will be published by Wiley this year. Chris Gillick, a freelance financial journalist, who writes about hedge funds and traders, contributed to this article.

juan F. Villegas


Leave a Comment