Successful trading is all about finding an edge. That means having a strategy that gives you a better chance of making money than not using one at all. So how do you go about finding an edge? In this article, we'll discuss the concept of edge and teach you how to use the Tradewell platform to identify whether or not your trading strategy provides one.
In order to be a successful trader, you need to have an edge. This means that you have some sort of advantage over other market participants that results in long-term profitability. Temporary edges in the market often exist as side effects of regulation or appear due to cycles, whereas sustainable advantages tend to come at the expense of less capable market participants, whose mistakes are usually reliably occurring.
As systematic trader Laurent Bernut has recognized, while edges themselves may be numerous and appear differentiated, they generally fall into a handful categories. Below is a list of four types of edges, accompanied by specific examples.
If an edge is said to be systemic, it means that it cannot be easily captured by standard measures market participants. Examples of this include edges that arise from country-specific exchange regulations and securities laws that benefit native citizens over foreigners.
A cyclical edge refers to an advantage that materializes periodically because the market has momentarily changed in a way that rewards a particular trading style. This topic is addressed at length by Nicolas Nassim Taleb in his seminal work Fooled by Randomness: The Hidden Role of Chance in the Markets and in Life. Cyclical edges are by nature unpredictable events: their timing and appearance cannot be foreseen.
Arbitrage represents one of the better known examples of a cyclical edge. The technique takes advantage of the price difference between the two securities. A trader who performs arbitrage will simultaneously purchase and sell identical or similar assets in different markets.
One of the more famous examples of arbitrage was known as the “kimchi premium”, a phenomenon that saw Korean exchanges list Bitcoin at a higher price than their American counterparts.
Arbitrageurs like FTC founder Sam Bankman-Fried could then buy bitcoin in the American exchange and later sell it in the Korean exchange at a higher price. Arbitrage opportunities like this can be highly profitable but are notoriously short-lived.
A sustainable edge is the opposite of a cyclical edge. It persists for extended periods of time and can often be practiced in different markets. Sustainable edges generally reflect the skill of the trader and less the good fortune of encountering a favorable market. Below are some examples of well-known sustainable edges.
Mean reversion trading is based on the belief that a particular security will revert to its previous state after an extreme change in pricing. This theory allows traders to short unexpected upswings and go long abnormal lows. Markets that have a historical propensity to mean-revert, such as agricultural commodities and natural gas, have historically provided the most reliable edge for this trading style.
Trend following is a strategy for investing or trading which takes advantage of long, medium, and short-term moves in various markets.There are a number of reasons that explain why this edge exists. Trends often snowball after traders jump on the bandwagon. People tend to pursue information that confirms their views, which can result in people buying assets that have recently made money and selling assets that have declined, causing trends to continue. Moreover, some risk-management models are designed to sell in down markets to balance their budgets, and buy in up markets if new opportunities present themselves. This behavior also tends to reinforce trends.
Traders can find their edge in the options market by exploiting mis-priced premiums. The opportunity exists because implied volatility often overstates how much a security will move.
Factor investing whose edge is grounded in targeting quantifiable firm characteristics that have empirically shown to maximize returns. A factor-base approach typically includes considerations of size, low volatility, value, momentum, growth, but evidence points to a wide range of factors capable of influencing outperformance.
Not all edges are considered legal to exploit, and in fact, some, if acted upon, carry the risk of prison time. Trading on material nonpublic information constitutes one such example. The term is used to describe corporate news that has not yet been made public and could have an impact on the share price.
Trading on material nonpublic information takes place whenever a person with fiduciary responsibility at a company trades shares in the business based on confidential information before it becomes public.
We’ve covered some of the reasons why mean reversion and trend following strategies have been empirically shown to provide an edge in specific securities markets. In fact, knowing the source of an edge is critical information for any serious trader, especially those pursuing strategies with structural and cycle edges. The reason of course is that if you can identify the conditions that give rise to your edge, you are more likely to identify when those conditions are no longer present in the market. And you therefore are more likely to successfully anticipate when your edge will expire.
Once a trader believes he or she has an edge, backtesting becomes a critical tool. The process can help traders confirm edge without the need for months of trading. Backtesting is perhaps an even more useful tool in determining what trading strategies do not in fact provide an edge. As long as you have reliable data and a good backtesting framework, it is recommended that you run both in- and out-of-sample tests for any strategy that shows promise of delivering an edge.
Backtest results from a strategy that provides an edge will look meaningfully different from the results generated by random strategies. Let's compare two strategies to see just how great the difference is.
The following trading rules are an example of a counter-trend strategy that looks to short the exchange-traded security UNG whenever its closing price deviates far above its 14-day moving average.
There are two key arguments that explain why this trade is not only likely to work, but also likely to be sustainable.
The first is the propensity of natural gas – the core asset held by UNG – to mean-revert. A study by Henrik Andersson in 2007 looked at nearly 300 different commodities over a time period of 36 years. It found clear evidence that commodity prices are mean-reverting instruments. This stands in stark contrast to other financial assets, which are generally likely to trend.
The second argument is that UNG primarily holds front-month futures contracts, which are often in a state of natural depreciation due to the effects of contango. So when the price of UNG deviates well above its mean, it will in theory face additional headwinds once it begins to revert.
Below is a list of five qualities that the backtest results of a good trading strategy are likely to exhibit. We can check the results of our trading strategy to confirm whether or not it possesses each quality, and thereby determine if an edge is likely to be present. The output featured in the article is taken directly from Tradewell.
Average profitability per trade (APPT) is the average amount a trader can expect to win or lose per trade. Traders who focus too much on one aspect of trading may lose sight of the bigger picture: trading performance depends largely on maintaining a high APPT. The value of APPT must also be large enough to cover slippage. The results of backtest show a healthy 20-day APPT for traders entering with a short bias, This data point provides evidence that an edge may in fact be present.
Strong edges, particularly those that originate from swing strategies, often exhibit profitability across multiple time frames. Notice how in the screen capture below the trading rules produce positive returns across every time interval, from 1 trading day to 60. When the results of a backtest exhibit this quality, we can be more confident that an edge truly exists. Similarly, when backtest results feature positive returns that are limited to a single timeframe or appear uneven across timeframes, they deserve comparatively more skepticism.
What truly separates a cyclical edge from a sustainable edge is the ability of the latter to generate profits across bull and bear market regimes alike. If a strategy is robust enough to perform in a manner independent from market regimes, it signals that a sustainable edge may in fact be present. In the example we present here, it can be argued that our strategy may not live up to this test, primarily because the backtest consists almost entirely of bear market data.
By comparing the average values of the two data sets in order to determine if they came from the same population, a t-test helps a trader identify if the trading signals from his strategy produce returns that are materially different than random returns. In the screenshot below, we see that the large t-value present in our backtest suggests that returns are in fact significantly different from random returns. This is yet another indication that the strategy may possess an edge.
A positive skew is the name we give a risk profile that has lower volatility when it loses money and higher volatility when making money. Simply put, positive skew means that a strategy produces winning trades that are on average larger than the size of its losing trades. Negative skew is the opposite of positive skewness; it is when stock's volatility increases when losing money and decreases while gaining. In the image below, we see two curves – one displaying the distribution of return for our signals and the other displaying the distribution of returns for all the data points included in our backtest. We can see that the signal results exhibit a risk profile that has a deep negative skew. If one were a long-only trader, this would be a signal to immediately abandon the strategy, but the trading rules we are backtesting call for shorting UNG. As such, the skew suggests that there is likely an edge.
It's easy to think that if your backtest results meet all of the criteria outlined in preceding paragraphs, then your trading strategy must have an edge. But this isn't always the case. In fact, your backtest might not really have one at all. Before you finally trade your strategy, you’ll want to answer a number of questions.
Overfitting is when a backtesting strategy performs well on historical data but poorly on other, unseen data. It occurs when a strategy is too specific to the data it was trained on, and results in no real edge since the strategy is unable to make accurate predictions with future examples.
It’s fairly normal for a trading strategy to perform better in backtest than in live markets. Therefore, whether an edge that is validated through backtesting can truly be exploited in real life often comes down to the scale of slippage experienced during live market trading. Slippage is the difference between what you expect to pay for an order and the price at which it executes. When markets are volatile, there is typically a greater chance of slippage. It can also occur when a large order has been placed and cannot be filled because there is not enough liquidity at the price.
In the event that the edge is cyclical or systemic as opposed to sustainable, there is a chance that it has become obsolete.
Good trading strategies can be the difference between a profitable and unprofitable experience. Conversely, bad trading strategies can lead to large losses in a very short period of time. So when it comes to trading, having an edge is key.
Knowing how to identify the presence of an edge by observing backtest results, and arriving at an understanding of why the edge exists in the first place are two crucial skills a professional trader must master.
Backtesting is of course a key tool that traders can use to confirm an edge, but practitioners will still need to be careful to make sure they’ve avoided common pitfalls in their research before they actually put capital at risk.
Start with the free version and upgrade when you need a larger metric library and longer lookback periods.