Last Updated on February 17, 2021 by Oddmund Groette
Quant trading, an abbreviation for quantitative/quantified trading, is growing in popularity because of the possibility of having the computer do all the work and reduce emotions and behavioral mistakes. This type of trading was previously only available for big trading firms, but today this is a viable option for most retail traders. We believe most retail traders stand a better chance if they are systematic.
However, quant trading has both pros and cons. In this article, we look at eight advantages and disadvantages of quantitative trading.
What is quant trading?
Trading is challenging and needs to be treated as a profession. It is difficult to find profitable strategies, and yet another hurdle is overcoming behavioral mistakes. Emotions, biases, greed, and fear are just a few of the issues you have to address. Trading is all about making decisions about an uncertain future.
Fortunately, you can dramatically reduce human errors with the advent of technology such as VPS (a virtual private server), software, coding, and programming.
Quant is an abbreviation for quantitative or quantified – pick your definition. Quant trading, also called algo or algorithmic trading, involves making 100% testable strategies run alone on a trading platform without human interference. The strategy can, for example, run on your VPS where you start the systems and portfolios before trading hours and stop them after trading hours.
Quant trading lets you focus on developing strategies and portfolios while all the trading and execution run automatically on their own. This way, you reduce the possibilities of human errors drastically. Furthermore, it gives you time to pursue other things, perhaps even having a separate source of income.
How does quant trading work?
An algorithm has predefined rules for buy and sell signals and scans the market at preset intervals, often with some seconds “rest” between each scan, so your computer or VPS doesn’t get overloaded, depending on how many strategies you are running. You can develop strategies by testing various hypotheses, or you can let your computer look for potential inefficiencies while you are sleeping. The latter is more and more common, but it involves rigorous out-of-sample testing. As a quant trader, you can never rest on your laurels. Trading requires constant work to make sure you have a plan B (and replace strategies when they stop working).
What steps are required for quant trading?
The following bullet points need to be done to develop a quant strategy:
- First, you need to formulate a plan or hypothesis. The strategy might be mean-reversion, momentum, trend-following, or perhaps a day trade.
- When you have an idea, you must be able to formulate the design down to pinpoint accuracy.
- When you have an automated strategy, you must implement it into your existing arsenal of strategies. Does it add value and diversification to the existing strategies?
- If the strategy passes both the out-of-sample test and the diversification criteria, you add it to your portfolio and software, for example, Amibroker.
- When the four previous tasks are done, implement the strategy in your software and let it run without interference.
The pros of quant and automated trading:
Quant trading has many benefits over traditional discretionary trading. Here are some of them:
No human intervention
The only intervention from your side is in the development process. As soon as the strategy is ready, it’s all about hands-off. Your computer is not liable to behavioral mistakes.
Quantitative trading reduces cognitive errors
Quantitative trading reduces human errors, and accuracy improves, which means that the possibility of errors is reduced drastically. The quant strategies are double-checked and rechecked again. Any technical indicator is defined to the smallest detail and no judgment error is made.
Quantitative strategies give leverage
Increased speed. An algorithm can analyze 100 different strategies, each with several criteria, in a split second. No human can accomplish this feat. Depending on your time frame, speed is of the essence. Automation makes you disciplined and gives you leverage and a wide range of capabilities.
Quantified strategies allow backtesting
Quant systems let you backtest on historical data with no room for interpretation. There is no discretionary judgment, which is impossible to backtest and test scientifically. Having strategies that are backtested should give a boost to your confidence.
Infinite numbers of strategies can be implemented
Computing power lets you trade almost infinite numbers of strategies, systems, and portfolios. The magic behind Jim Simons’ Medallion Fund’s spectacular performance is their enormous number of different strategies spread across different markets and time frames. Strategies that perform differently from each other lead to a smooth equity curve and high profit factors and Sharpe ratios.
Automated trading frees up time
Quant trading lowers your costs and increases efficiency. There is no need to sit in front of the screen all day, and it frees time you can spend on getting additional income. Moreover, there is reason to believe the friction costs of getting in and out of position go down.
Quant systems resemble your backtests
The order entry speed is much higher than what you can do physically. When the scan is done, an order gets sent out in one-hundredth of a second. The speed of entry and exit helps you resemble your backtest to the exact detail.
Systems let you follow your plan
Quant systems make you consistent. A plan requires you to follow the plan, and predefined rules for buying and selling make you follow the plan down to the smallest detail. Plan the trade and trade the plan. Despite ups and downs in the market, your quant systems help you stay consistent and disciplined.
The cons of quant and automated trading:
Unfortunately, there are not only advantages with quant trading. Pitfalls are plentiful, and you are liable to black swans. You are dependent on your technology, and your code might turn your portfolio into a disaster. You can’t prepare for all scenarios.
Quantitative trading is prone to curve fitting
The greatest pitfall is curve fitting and optimization. The reliance on historical data inevitably leads to many strategies that are a result of chance and randomness. Out-of-sample tests reduce the element of curve fitting, but even that can’t eliminate it. As Nassim Taleb says: if you want to bankrupt a fool, give him lots of information. Randomness is much more prevalent than most traders believe. It’s easy to fool yourself by trusting the data 100%. We recommend testing your strategies in a paper account for at least six months before you go live, depending on the time frame and number of observations. The more observations, the shorter the testing period is required.
Being a quant requires coding skills
You need to learn to code and program. While this might scare off many potential traders, especially those over 40 years of age, it doesn’t require much effort to learn. Besides, learning has diminishing returns the more you learn. You make a great leap by learning little, but the marginal utility falls as you learn more. Just knowing a little helps you a long way.
If you are dependent on hiring someone to do most of your tasks, you are at a considerable disadvantage. Do yourself a favor and self-study for a couple of months, and you’ll make giant leaps in your trading.
Quants are liable to black swans
A lot can go wrong when you automate. Errors in code can lead to multiple trades, potentially ruining your account, which happened to Knight in 2012. Moreover, just a tiny mistake of where you put parenthesis can alter the strategy substantially. “Black swans” happen. You might think you have it all planned, but you can’t prepare for the things you don’t know.
Even though computers let you automate all the tasks, we don’t recommend leaving it completely alone. Connectivity and power outages are both potential black swans. For example, if you are using a VPS, a sudden reboot on their end can create havoc in your systems. However, that’s why we have smartphones (?).
Quant trading involves fixed costs
Quant trading requires costs and subscriptions. You need quality historic share/future price data, live data fees, VPS, and software. You can expect a price tag of a few hundred USD a month, depending on your level of service.
Some strategies are impossible to code – keep it simple
Some strategies can be difficult to code, if not impossible. They might be complex and require programming skills you don’t have, or you somehow can’t quantify the criteria. Nevertheless, complex strategies are not necessarily better than simpler ones, quite the opposite. An underappreciated trading skill is making “simple” strategies with few variables. Our own anecdotal experience indicates many programmers turned traders make things too complicated.
A computer is “dumb”
A computer is “dumb” – you need to code the computer what to do. Artificial intelligence has come a long way, but the human mind is still superior to your trading software. A human mind can understand irrational behavior. A computer can’t.
Quants need to develop strategies continually
Quant trading requires constant development of existing and new strategies. The reality of automated trading is that most, if not all, strategies eventually die or get “arbed” away. The only constant in the market is change.
Loss of control
Ironically, the loss of control might make you more uncertain. For example, many quant traders turned off their programs in March 2020 when the Covid-19 created enormous volatility. Opposite, if you already “know” the strategy will not work, it can be problematic just to ignore one or two trades. There are always temptations to fiddle with the systems.