☰ Menu eFinancialCareers

Right now, human traders have the edge over quants and algos

traders, trading, algorithmic trading, e-trading, electronic trading, trading algos, trading algorithms, human traders, pit traders, Wall Street, banks, bankers, hedge funds, sell side, buy side

The sun may be setting on the era of humans traders, but the current market environment gives them a chance to shine.

It’s not easy being a trader in this day and age. Old-school trading pits have been cleared out. Electronic trading platforms powered by various types of artificial intelligence, including machine learning and artificial neural networks (ANN, also known as connectionist systems) are in ascendance, causing both buy-side and sell-side firms to employ fewer human traders.

It’s stressful for traders when markets are volatile, but choppy seas with swings in both directions often present opportunities. The current prolonged period of market calm is the flipside of that coin, where traders are twiddling their thumbs waiting for some action. However, there’s a silver lining. Many quants and algorithmic trading strategies struggle mightily in the absence of market volatility, presenting human traders with an opportunity to prove their worth.

“Some of these algorithms will focus on market-making, trying to capture the difference between the bid and the offer, which are typically categorized as high-frequency trading,” says Daniel Gramza, the founder and president of Gramza Capital Management and DMG Advisors and a trading coach to brokerages, banks and hedge funds. “Sideways or calm markets can be a challenge for the machine-learning trend-following algorithm – primarily because it needs a number of steps to recognize that it may be in a sideways market and how it will trade that environment.”

Advantage: human traders. The trader can quickly recognize the potential for a sideways move and easily adjust their trading strategy, Gramza says.

“The trader can assess what is the typical magnitude of a sideways move, how long it typically last, is there a particular time of day, week or month when sideways moves occur and how does it typically break out of the sideways move,” Gramza says. “Although these parameters are simple to assess for the trader, it can be challenging for the machine-learning algorithms,” he says.

“My experience with trading firms around the world is that the human trader still has an edge over the machine-learning algorithms in the calm sideways market environment.”

“Human traders may have the instincts and intuition to add statistically significant value, because the best ones do a good job of picking price levels and sourcing liquidity,” adds Dave Weisberger, the head of equities at strategic advisory firm Viable Markets and the president of boutique consultancy Exquam whose also worked at Two Sigma, Citi, Salomon Smith Barney and Morgan Stanley.

“Good traders have the intuiton to say, ‘The price has gone down too much, I want to stand firm, I don’t want to trade here – particularly in this sort of market where it’s not moving too much, that’s relevant,” he says. “Also, you need a human trader for negotiating a block trade – on the sell-side, you need sales-traders who have lots of experience and know who they can call that won’t leak information when looking for a counterparty on the other side of the trade.”

In addition, human traders have the advantage when it comes to deciding on the appropriate benchmark for the trading strategy.

“On the buy-side, the portfolio manager is also a human, and while they might be better off implementing their trading goals mathematically, they don’t, so traders interact with PMs and ask, ‘Is this price a good price? Do you want to get it done?’” Weisberger says. “Humans are good for that sort of thing.”

On the other hand, when it comes to routing, placing and modifying orders, trading algorithms are necessary.

“It’s inarguable that a human pushing buttons to break an order up is not how to do optimal routing – in a situation where you have five different offers on the screen, by the time you get to the second or third the others will be gone,” Weisberger says. “When it comes to order placement and probing for liquidity, which means making tens of thousands of actions over the course of the day in the way the market is structured, you need machines for that.

“There are certain things that machines have to do – the last mile of the trade,” he says. “Now algorithms are good at slicing up orders.”

However, be warned: As AI and machine learning get better, they will start taking over more trading processes such as the determination of benchmarks and setting price levels.

“Right now placing and modifying orders on exchanges should be done by algorithms, whereas humans 100% need to negotiate liquidity,” Weisberger says. “The battleground is in the middle: picking price levels and choosing optimal strategies based on the requirements of the PMs.”


Photo credit: vvvita/GettyImages

Comments (0)

Comments

The comment is under moderation. It will appear shortly.

React

Screen Name

Email

Consult our community guidelines here