Daniel G Calugar Foundation – Merit Scholarship Program

The power of data is undeniable, and it's affecting nearly every industry in the world. Investor Daniel Calugar points out that data and machine learning are dramatically impacting the financial industry.

In the last ten years alone, the amount of assets under management for algorithmic trading has doubled, up to more than $1 trillion. This wave of change is influencing an increasing number of traditional investment management firms to employ data scientists and experts in machine learning so they can adopt a more rigorous scientific approach to investing that will help them improve returns.

At its core, algorithmic trading is designed to remove one very crucial human element from trading -- emotion. Machines have an unbelievable ability to make thousands of decisions in the blink of an eye and do so without weighing emotional consequences.

So, even though no algorithmic system will ever be able to reach the human interpretation of the market environment, humans will never be able to properly manage and track the sheer amount of information that algorithmic systems can. That's why algorithmic trading is not just here to stay; it's set to grow at even more rapid rates.

Below, investor Dan Calugar describes in detail what algorithmic trading is, how it works, the benefits of using it, the types of trading strategies that can be employed, and what is needed to use it.

What is Algorithmic Trading?

Algorithmic trading goes by several different names. You may have heard it called automated trading or black box trading. But most people will refer to it by an abbreviation, algo trading.

In its simplest form, algorithmic trading is an approach that financial investment companies and individuals can take using computers to the best of their abilities. It takes advantage of the immense processing power these computers have to make split-section trading decisions automatically -- without the need for human intervention.

A computer program will be set to follow very defined instructions, known as an algorithm, to execute a trade. Data will be inputted into the computer, which will run the data through the algorithm, identify anything that fits into the defined set of instructions, and then execute a trade at any time that occurs.

Algorithmic trading takes human emotion out of the trading process and removes the need for intense manual work. For example, instead of a human constantly monitoring financial markets for various buy and sell indicators, the computer does it for them using the algorithm's parameters as its guide.

What Does Algorithmic Trading Take into Account?

Algorithmic trading will take into account whatever the person, team, or company asks it to. However, humans still have complete control over algo trading since they are the ones who will code the algorithm to meet their trading strategies.

The algorithms can monitor anything from price, timing, quantity, comparative analysis, and much more. Moreover, there is almost no limit to what can be included in an algorithm for trading purposes.

Traders will set parameters for the different metrics they want to track and then tell the computer to execute the trade once those parameters have been met.

Example of Algorithmic Trading

The best way to understand how algorithmic trading works is to look at an example. Remember that algo trading can set parameters for both buys and sales.

The trader may set a criterion that tells the computer to purchase 250 shares any time a stock's current price is lower than its average price over a period of time. So, for example, if Google's stock price is a certain percentage below its moving average for the last 50 days, the computer will execute a trade to buy 250 shares.

On the flip side, if Google's stock price is a certain percentage above its moving average for the last 50 days, the computer might be told -- through the algorithm -- to sell a certain percentage of shares the trader owns.

This is just one algorithmic trading strategy that can be employed. Below, we'll dive into some additional ones.

What Are Some Other Strategies That Can Be Used with Algo Trading?

The strategy described in the above example is based on the moving average of a stock price. Therefore, the algorithm's parameters are set around the moving averages of the stock.

Some of the variables that can be input are:

• The time period for the moving average
• The percentage above or below that moving average of the current price should be before executing a trade
• How many shares to buy/sell (either a set number, total dollar amount, or percentage of something)
• The stocks to track

Moving average is one of the most common algo trading strategies traders employ and one of the easiest to explain as well. Dan Calugar provides some additional algo trading strategies below.

Following a Trend

The moving average can be considered an indicator of finding a trend in the market. The algorithm is set to make trading decisions based on a specific market trend: the moving average and a stock's current price relative to it.

This is just one way that algorithmic trading can be used to follow a trend. Other trends that can be followed include movements in the price level, channel breakouts, and other technical indicators that are related to them.

The algo trading strategy of following a trend is considered the simplest since it doesn't involve making any price forecast or prediction.

Rebalancing Index Funds

Index funds are often considered some of the most stable investments in the stock market. As such, investors typically dump money into them as long-term investments rather than for short-term gains.

However, the latter can be done through algorithmic trading. All index funds will have a defined period by which they will rebalance their holdings so everything can be on par with its benchmark index. For instance, an S&P 500 Index Fund will regularly rebalance to ensure it's aligned with the S&P 500.

Through algorithmic trading, traders can capitalize on trades they expect to occur that could offer as much as 80 basis points of profit, depending on how many stocks are included in the fund right before the rebalancing.

Algorithmic trading is very successful here because it can perfectly time the trades so the investor can get the best price possible. It's something that would be very challenging for human analysts to time well.

Arbitrage

Arbitrage trading takes a lot of experience, perfect timing, and the ability to browse the price of the same stock at multiple markets. Computers can simultaneously scan various markets and compare prices to find any discrepancies.

Taking advantage of these discrepancies is what's known as arbitrage trading. Arbitrage trading involves buying a stock that's dual listed at a low price from one market and then selling it at a higher price from another market simultaneously, according to Daniel Calugar.

The price differential is considered to be profit that's free from risk or arbitrage. Traders can take the same approach for comparisons of stocks versus futures.

Differentials in price do exist at times and taking advantage of these differences can be quite profitable. Algorithmic trading can be very effective at instantly identifying arbitrage and capitalizing on it.

Mean Reversion

One of the most common models used in all statistical analyses is the mean, which represents the average value of something. The generally accepted mathematical theory is that everything will eventually revert to its mean.

In financial markets, algo trading can track diversions from the mean and make trades based on them. The theory here is that assets may experience a temporary high or low in price, but eventually, it will periodically come back to its mean.

Traders can define the price range for the mean and then implement an algorithm based around it to execute trades automatically whenever an asset's price divers from that price range.

If an asset's price goes above the mean range, the algorithm will tell the computers to execute a sale. If it goes below, it will complete a purchase.

What Are Some of the Benefits Algorithmic Trading Provides?

There are many benefits that algorithmic trading can provide investors. Here are some of the most significant advantages.

Objective Decision Making

The most commonly cited benefit of algorithmic trading is that it allows for objective decision-making. For example, Dan Calugar says that the decisions on buying or selling are based on specific rules and data rather than "feel" and emotions.

Even the most experienced and savvy traders can be influenced by public sentiment, excitement, and other emotions. When they are, they might be swayed in a direction that diverts from their overall trading strategy.

Algorithms help traders avoid acting on these emotions and ensure that they always stick to the planned strategy.

Optimized Trading

Whether a trade is successful or not depends on many different factors. Sometimes, the difference between success and failure can be a dollar or a minute.

Successful traders aren't defined by what assets they buy and sell as much as they are defined by when they buy and sell assets and at what price. In other words, the timing of the trade is just as important as the trade itself.

The biggest challenge for traders is timing trades perfectly. With so many different markets and assets to monitor simultaneously, it can be hard to make split-second trading decisions in real-time. This could cause the trader to miss out on the best price or sell too late, resulting in significant losses or missed profits.

Algorithmic trading allows for trades to be executed at the best prices possible. In addition, they are executed instantly, which allows the trade to be locked in before significant price swings.

Testing Protocols

One of the downsides to algorithms (as we'll discuss in a bit) is that if the parameters aren't set correctly or if the theory behind them isn't solid, trades could result in significant losses. One of the best parts about using computers is they can run millions of simulations quickly to test theories.

Investor Daniel Calugar points out that this allows traders to test their algorithms before risking any real money in the market. Through back-testing, algorithmic trading can use any available real-time and historical data to analyze whether a trading strategy is viable.

If the strategy proves unsuccessful or less successful than desired, the algorithm can be tweaked and back-tested again. Finally, when the trader is comfortable with the results, they can implement it in the real market.

Eliminates Mistakes

Human error is a genuine threat to successful trading. In addition to incorrect decisions being made based on emotions, humans can make mistakes when trying to execute a trade manually. For example, if they incorrectly type in details of the trade, they can miss out on an opportunity or incur exorbitant fees to correct the mistake.

Computers can eliminate these mistakes. Once the algorithm is set, the computer will execute the exact trade the investor tells it to, in the exact way they tell it to. There's no risk of the computer making a manual error when executing the trade.

What, if any, Are the Downsides of Algorithmic Trading?

Like anything in life, there are some downsides to algorithmic trading. Dan Calugar outlines a few of them below.

No Human Control

Algorithms are designed to remove human elements from trading. In many cases, this is a positive, but there are some downsides to it as well.

The main one is that all trades will be made automatically. If the algorithm's parameters are met, the computer will automatically execute the trade it's supposed to.

Humans cannot change their minds, cross-check other data sources, or decide they'd rather pass on a trade for one reason or another. The algorithm is either on and operating or off and not operating.

Requires Monitoring

While traders won't need to monitor the markets as closely as they did without algorithms, they will need to monitor the algorithms themselves consistently. If market conditions change, for instance, the algorithm will need to be updated to reflect any new trading strategy.

If the algorithms aren't updated at all -- or just aren't updated in a timely fashion -- the computers could end up executing ill-advised trades that could result in considerable losses. Humans won't be able to remain completely hands-off because of this, and they need to act quickly if they feel the current algorithm is no longer the best approach.

To effectively monitor the algorithm, traders need to know the details of it inside and out -- and understand the potential effects of the changing market conditions. Time can't be lost reading up on the algorithm's parameters while it's still running, as it could result in poor trades.

If the traders decide to deactivate the algorithm while they make tweaks and run additional back-testing, they could be missing out on excellent trading opportunities while the computer is "sidelined."

Programming Skills

Traders may be able to properly identify trading strategies that consistently work and the specific levers that need to be pulled to make it work. However, if they don't have the proper programming skills -- or have people on the team who have those skills -- then there will be no way to create the algorithm to put it into action.

If the trader isn't part of a larger financial institution that has trained IT professionals on staff, then they must rely on hiring an outside vendor to code the algorithm for them. There are obvious risks to this, Daniel Calugar says, including simple mistakes, no way to double-check their work, and the potential for fraud and theft of the strategy.

Will Algorithmic Trading Transform the Financial Industry?

The question probably shouldn't be will algorithmic trading transform the financial industry -- since it already has -- but just how much will it transform the industry.

Dan Calugar says algo trading is not just a fad; it's here to stay and will grow exponentially in the next few years.

While algorithmic trading was initially reserved just for institutional investors and large financial firms, even the casual trader today can use algo trading to help them make investment choices. In addition, various programs and mobile apps put the power of algorithmic trading literally in the hands of individual investors.

While these tools may not be as robust or personalized as those used by institutional investors, they still help individuals automatically make trading decisions based on data, science, and statistics rather than pure emotion and manual research.

Algorithmic trading can be highly beneficial to many different types of investors, and investor Daniel Calugar says more changes and advancements in technology and working knowledge of AI and machine learning will continue to refine the process and improve it.