In one of my recent stock trading bootcamps, I was asked: “What are trading bots?” I then realized it was something I don’t talk much about since on the retail side they’re most popular with Bitcoin and cryptocurrencies. But as that has started to leak over into the stock, ETF and futures markets, I decided it was time to cover it.
A trading bot or algorithmic trading system is a piece of software that connects to your trading platform(s). They’re designed to autonomously execute trades (sometimes thousands of round-turns per day) based on preprogrammed trading signals or indicators with the intent of turning a profit.
How Do Trading Bots or Algorithmic Trading Systems Work?
They work (for the most part) by combining three key components:
- The Signal: The trading bot will have a pre-programmed set of trading rules based on existing or proprietary indicators.
- Risk Allocation: The software determines the amount of capital to allocate once the conditions for the signal have been met.
- Execution: The trading bot must determine how to execute the order depending on the allocation of capital. It must be programmed to understand it can’t just dump a gigantic order onto the market or it will never get filled and/or cause the price to runaway without it.
All three components must work in perfect harmony to have a chance of consistently beating the market and turning a profit. If just one of the three components experiences an issue, the whole thing goes bust.
Are Trading Bots or Algorithmic Trading Systems Legal?
Yes, they’re completely legal and evergrowing on the institutional side of trading. Over recent years, trading bots and stock trading algorithms have started to become available to retail traders. As we’ll soon discuss, trading bots or trading algorithms account for the vast majority of trading volume on U.S. exchanges.
What Percentage of Trading Is Done Using Algorithmic Trading Systems or Trading Bots?
At this point, only 10% of stocks traded on U.S. exchanges are performed by actual humans. The other 90% is done by some form of algorithmic trading system or trading bot. It’s likely that over the coming decades, the remaining 10% will continue to be widdled away.
Do They Actually Work?
Yes, and no… The beauty with most of the systems available to the average retail trader is that they immediately eliminate one of the biggest enemies: emotions. But most of these trading systems are only as good as the set of rules laid out. On the institutional side of trading, these systems work most of the time without issue. As you’ll read below there’s been a few noteworthy incidents where the use of algorithms went completely wrong.
What Are High-Frequency Trading Systems (HFT)?
A High-Frequency Trading System, or HFT, is an algorithmic trading system capable of executing thousands of round turns (in and out of a trade) per day.
A Few Examples Of When Algorithmic Trading Went Terribly Wrong
Let’s take a trip down memory lane over the last decade and look at a few examples of how algorithmic trading systems have had a significant impact on the markets. We’ll open things up with a look at the ‘Flash Crash’ on May 6, 2010.
The Flash Crash of May 6, 2010
For those who don’t remember, this was where the Dow lost over 1000 points within about 15 minutes. One man was later largely attributed to the crash: Navinder Singh Sarao. Navinder built an algorithm that would create large E-mini sell orders at a multitude of different levels all just above the best asking price, spoofing the market. At times, he’d account for as much as 40% of all sell-side action. This would inevitably drive the prices down and that’s exactly how he’d capitalize. He’d turn the algo off, creating a pop in price already in a position to profit from the key timing of the volatility, making out like a bandit- sometimes to the tune of $1M in a single trading day.
If you take a look below you’ll see a chart of that day:
In approximately 36 minutes, the crash and recovery happened, leaving most traders scratching their heads with jaws still on the floor.
This moment in history forced regulators, to acknowledge the power that algorithms had in the market. The SEC soon followed with a list of new regulations and ‘safety nets’ for the market in hopes to prevent or at least reduce the number of moments like this in the future. But as you’ll soon find out, the holes in their safety net weren’t big enough for a 440+ million dollar glitch to fall through.
The Knight Capital Disaster
Fast forward to a very grim morning on August 1, 2012, for one of the leading market makers at the time, Knight Capital. The day already fueled with headlines of the growing instability of Greece at the time. That morning like any other day, Knight Capital started to receive an inflow of small retail orders, until things went haywire! Their system mistakenly streamed thousands of orders onto the market over a span of about 45 minutes. This system glitch resulted in their automated system executing over 4 million trades in over 150 stocks. The combined value of their positions exceeded $6 billion in total.
This influx of orders soon caught the attention of NYSE analysts, doubling the average volume they’d typically see. After a little investigating, they tracked the increase in volume to Knight Capital. Knight Capital’s CIO was reached and they quickly realized they were ill-prepared for a situation like this. There was no killswitch in place and it took twenty minutes before they could stop the bleeding.
After the dust settled, Knight Capital had lost over $440 million from the software bug. That was more than $10 million per minute! By the next day, their stock price had plummeted losing more than 70% of its value.
“So, Should I Get A Trading Bot?”
Ultimately, that’s a decision you’ll have to make for yourself. I realize we’re at a point where some cars and trucks are now fully capable of driving themselves. So it’s easy to understand how one might assume that enough things have improved with trading bots or algorithmic trading systems that we can feel confident putting our hard-earned money into their hands, right? Still, the answer for me is “No.”
The reason is, with the huge increase in the advertising of these systems, all promising the ability to outperform the average trader, there haven’t been any major studies or reports confirming those statements…it all sounds a little too good to be true.
In fact, the little information I was able to find showed the vast majority of people using some sort of retail trading bot ended up losing money. This is because they all still require some degree of human interaction and management. It’s not simply a piece of software that prints money or everyone with any sense would have one.
As you’ve probably experienced with a new software update to your phone, laptop, Smart TV, etc., at least once, a bug has stopped a certain element of the app or program from working. These bugs typically have little to no effect on our financial well-being. Perhaps I’m just old fashioned, but until I see some legitimate data or results using a retail system, I won’t be ready to turn over any amount of my portfolio to them just yet. Who knows, maybe in a few years from now that will all change. For now, I’m more confident leaving the trading to me and my trusted indicators.