3. The downside to algorithmic trading

The major disadvantage of algorithmic trading is that with one mistake, traders can rapidly lose their money. The faulty algorithm can trigger hundreds of transactions within a few seconds and if something goes wrong, the investor’s capital can be lost in that same time frame. There have been many high-profile banks with traders who have encountered either an algo that goes haywire or the typical human fat finger mistake. An algorithmic trading platform needs operational hardware during the execution of trades. 

Dedicated computers, servers and connections are needed to ensure the system runs correctly. Some platforms do not operate entirely via internet connections and data servers and, instead, store impending trade orders on the client-side computer. In the event of a lost connection, that trade may not occur. Although the use of technology helps to eliminate human errors, algorithmic trading systems require to be constantly monitored due to the risk of errors, glitches, and power losses. Nasdaq recommends that traders set up monitoring and surveillance teams that work 24/7. However, monitoring systems 24/7 is beyond what most traders and small firms are looking for. The average retail trader thinking of trading whilst they sleep please recognise that automated trading is not unattended trading; what can go wrong will go wrong. Automated trading systems need constant optimising and, many times cannot compete with a strong human trader who is able to adapt to changing market conditions. 

Automated trading systems working purely of technical analysis will not be able to process news announcements and fundamental information which can easily move markets. Automated trading has also created illegal markets where algo traders spoof orders, this was highlighted to the public when the high-profile case of ‘The Hound of Hounslow” who was arrested and used as the scapegoat for the flash crash in 2010. Algo trading has been blamed for poor market price action which has lead to buying and selling based on no real underlying value expectations of stocks or the economy. If algo bots see momentum, they all fire orders creating many artificial market environments with no fundamental basis, this leads to erratic stocks and market prices, leading to very fast crashes.

4. How does algorithmic trading work?

Algorithmic trading bot identifies market inefficiencies and exploits them, potentially resulting in substantial profits. All algo trading strategies follow a set of defined rules; suppose a trader follows the below trade criteria:

• Purchase 50 shares of a stock as its 50-day moving average rises beyond the 200-day moving average
• Sell stock shares as its 50-day moving average dips under the 200-day moving average

These two simple instructions can be written to automatically monitor stock price (and the moving average indicators) and place the buy and sell orders when the defined conditions are met. This simple algorithm is much more efficient than a trader who would monitor the live prices and graphs or place orders himself. However, in practice, trading algorithms are much more complex and require a lot of knowledge and research. We do not recommend starting live trading without testing your strategy and performing various analysis. Quant Savvy offers 100% hands-free futures trading system to increase returns and minimise potential risk. Our automated trading systems invest autonomously, eliminating the psychology of investor from trading decisions.

Finding Inefficiencies in the Market

An inefficient market is one in which an asset’s prices do not accurately reflect its true value. Most markets display some level of inefficiencies. Quant Trading experts expose those inefficiencies in the market using sophisticated computerized algorithms that can generate returns for investors. Market inefficiencies exist due to information asymmetries, transaction costs, market psychology, and human emotion, among other reasons. Many believe in efficient market hypothesis, while others exploit deviations. However, an efficient market hypothesis is mainly academic and does not apply to live trading to generate real profits. In an inefficient market, some investors can make excess returns while others might lose more than expected. Quant trading experts say: “If you can’t measure risk, then you can’t manage risk”. Most of the human discretionary funds trade based on pure guesswork. All markets tend to trend beyond random, so there is no right or wrong method as long as exposed market inefficiencies show some repeatable behaviour – read more about quantifying market inefficiency. However, there is an important ‘Black Swan’ phenomenon to consider. Even though algorithmic trading is not affected by human emotions, there are multiple market dips, or black swans, with algo trading. There is a sudden dramatic move in the market before it returns to normalcy shortly afterwards.

Finding a Fundamental Truth to each Market

All-day trading automated trading strategies are based on finding a fundamental truth about the market. Quants aim to define rules which seek best to model repeatable non-random behaviour. Some Quants might use technical analysis and test if repeatable patterns exist in past market data, some may use arbitrage models and others may use order flow analysis. Whatever method is used the Quant will break down the premise into a series of basic rules and test them on various sets of market data. If the tested rules prove to show an edge well beyond random, then the rules can be pursued further, and a system can be designed around these core set of rules:

• Algo trading strategies can collect data on order flow and model the behaviour of other algorithms, once a participant tips one’s hand, your algo trading strategy can take advantage and expose and inefficiency time and time again.
• Investors love trend/momentum following but have to wait for months to profit. Moreover, long-term trades can be pure luck, whereas Quant tends to work on a day trading timeframe and can make even 100s of trades a day. On a day trading level, the same market trends exist, your algo trading strategy determines when is the best probability of catching a strong trend.

5. Algorithmic Trading Strategies

Creating your algo trading strategies you can keep to the KISS rules (keep it simple stupid) or go for complexity. What I find is that begin every strategy with a simple concept, something you may have observed the market doing. If you spot a repeatable pattern simply try to code the logic and test it on a very basic level. If you are curious to learn more, read our article: Finding a Trading Edge. The following was cribbed from Hedge Trade blog – read their blog post to learn more about each strategy

Trend-following strategies

Trends-following aims to take advantage of long, medium, and short term movement in the market. This strategy does not try to predict a new trend, it simply bets on the maintenance of a current trend. The simplest way to identify a trend is by following the moving average. Once moving average changes, traders determine when is the best time to buy or sell.

Arbitrage Opportunities

Arbitrage is when an investor buys a stock in a lower price market and then sells it on another market with a higher value. An investor can employ arbitrage by buying a stock on a foreign exchange where the price has not yet increased to another market’s rate. High-frequency trading (HFT) often employ arbitrage strategies. The strategies include different forms of arbitrage – index arbitrage, volatility arbitrage, statistical arbitrage and merger arbitrage along with global macro, long/short equity, passive market making, and so on. HFT relies on the ultra-fast speed of computer software, data access (NASDAQ TotalView-ITCH, NYSE OpenBook, etc) to important resources and connectivity with minimal latency (delay).

Rebalancing

Rebalancing strategy involves analyzing the investment holding within the portfolio and adjusting to the original allocation at the desired frequency to maintain a specific level of risks. When rebalancing takes place, there is an opportunity for algo traders to capitalize on these exchanges that offer 20 to 80 basis points profits.

Trading Range or Mean Reversion

Algorithms are used to make purchases when an asset’s price moves out of its defined range. This defined range is the ‘trading range’. It is also a ‘mean revision’ because it acts on the temporary price of an asset. That means that algorithms are looking for real averages and aims to exclude the outliers. Breakout, mean reversion, and trend following are three main themes the market can be in. If you can spot the current market theme you can apply an algorithm that exploits the market behaviour (the hard part is spotting the current theme).

Volume-weighted Average Price

Volume and price determine the average price of a security traded throughout the day. Volume traded average offers insight into the trend and value of a security. Passive investors often use the volume-weighted average price as a benchmark, which can include mutual funds and pensions.

Time-weighted average Price

This strategy aims to move discrete chunks of an asset at specific time intervals without causing the price of an asset to drop or rise dramatically. The idea is to maintain the average price when the order is being executed. If a large amount of stock is released to the market, traders and trading systems are alerted about a major change. The big movement in the market might be negatively interpreted and stocks’ prices can be dramatically affected.

Examples of algorithmic trading strategies can also use the following logic:

Pattern recognition

– Some daytraders feel that markets/stocks are fractal and there is some evidence to suggest this. Markets tend to work in waves/fractals or have other repeatable patterns. Daytraders that adhere to this style of trading often follow patterns such as head and shoulder, triangles, rectangles, wedges, etc.

Indicators

– Your basic moving averages, RSI, etc. (what we find is that most indicators are completely useless and lag the market. Moving averages can be used as a complementary filter to an otherwise decent algorithmic trading strategy, however.

Volatility

– Most daytraders know that most of their money will come from periods of high volatility; this is because on a day trading level this is when markets are at their most inefficient. Algo trading can perform superbly during these periods because they adhere to strict rules and do not suffer from fear and greed.

Technical Analysis

– Comprising of numerous features, from patterns to indicators.

Support/Resistance

– Can be categorized as technical analysis. Can be useful but very hard to code the logic of support and resistance into a quant trading system

Pairs trading

– The pairs trade is market-neutral, meaning the direction of the overall market does not affect its win or loss. The goal is to match two trading vehicles that are highly correlated, trading one long and the other short when the pair’s price ratio diverges “x” number of standard deviations – “x” is optimized using historical data. If the pair reverts to its mean trend, a profit is made on one or both of the positions.

6. Backtesting

Backtesting is the key ingredient to the success of an algorithmic trading strategy and trading bot. One of the best algorithmic trading advantages is that investors can backtest their trading strategy based on historical and real-time data. If you are interested in trying algorithmic trading for yourself it is important that you have advanced computer knowledge before setting up this type of trading strategy. If you don’t have any experience in trading or programming you should strongly consider leaving this algo trading to an expert with advanced knowledge.  

7. Learn Algorithmic Trading

The most difficult aspect of learning to become an Algo trader is where to start. Algorithmic trading requires knowledge in several fields: quantitative analysis/modelling, programming skills, trading/financial markets knowledge. The easiest way to begin any new pursuit is with the help and guidance of a mentor. Those who don’t have this option, often trust the help of books and the internet to find a starting point. However, you should consider that various resources might have misleading information, so we highly recommend to trust experts or find a mentor. Now the biggest caveat when it comes to finding books to learn algorithmic trading is that you are not going to be given automated strategies (not any that will work anyway) – these you must work to find on your own. What you can be given is outline how to begin the process of coding, software to use, correct backtesting methodology, and best metrics for evaluation of your quant trading strategy.

“First and foremost you need to learn how to code.”

This is probably the most daunting aspect for most daytraders or even ex floor traders. However, let me assure you that learning to code algorithmic trading strategies can be very simple and easy and most will pick it up within a couple of weeks of learning. Modern trading software has made it very easy for us to code, something like Multicharts is geared with simple coding language and most technical functions are pre-built into the software. For the average quant, you do not require anything more than top of the line retail software. Quant Savvy has compared different software and trading platforms to identify its pros and cons. There is no need to re-invent the wheel and start creating your own software (this takes too much time and effort and advanced computing and coding skills). Why take on the arduous task of creating your own software when there are companies out there working on this for you each and every day. Obviously, for very complex trading strategies a custom-built platform and software are required but I would say 9 out 10 daytraders needs are fully met by retail software. To begin with, I would simply recommend learning to code quant trading strategies in something simple like Multicharts using their own basic programming language (learn more). We have also prepared a list of learning resources, including books and blogs, for you to have where to start.  

Books to find ideas from:

Let me start out by saying most books on algorithmic trading are useless to the average trader, some are needlessly mathematically complex and others are simply flawed and basic. The only books I found of any use were ones where the backtesting process was rigorously discussed. Books where correct data mining practices and how not to overfit data or curve fit have been useful. Moreover, books wish to give you tools to evaluate your algorithmic trading strategy potential success and measure your edge have also proved to be useful. Do not expect to find the holy grail of day trading strategies in a book. A good book can set you on the right track but getting to the finish line will be based on your quantitative reasoning, logic and ability to find an edge. See article: Finding a Trading Edge. Books which give you a good story about algorithmic trading or day trading in general also can be a fun read as well are:

• When Genius Failed, by Lowenstein: popular recantation of the LTCM fiasco
• Market Wizards: Interviews with Top Traders (Books) Jack Schwager
• This Time is Different: Eight Centuries of Financial Folly (Books) – Carmen M. Reinhart

Other books which you might find algo trading strategy ideas – you can code up the rules and test them out (bear in mind doing this might be good practice but also be aware that hardly any of systems in the books work when tested rigorously; some work for certain periods of time and then fail under different market conditions):

• Studies in Tape Reading (Books) Richard D. Wyckoff
• Steidlmayer on Markets: Trading with Market Profile – J. Peter Steidlmayer
• Dynamic Trading: Dynamic Concepts in Time, Price and Pattern Analysis with Practical Strategies for Traders and Investors (Books) Miner, Robert C
• Technical Analysis of Stock Trends (Books) Magee, John

Advanced Quantitative books including complex mathematics and understanding behind quant trading:

• Advanced Trading Rules (Quantitative Finance) (Books) Acar, Emmanual
• Applied Quantitative Methods for Trading and Investment (The Wiley Finance Series) (Books) Dunis, Christian

Very important books on correct procedure on backtesting algo trading strategies:

• Evidence-Based Technical Analysis: Applying the Scientific Method and Statistical Inference to Trading Signals – David Aronson

9. Get Started