Talking Points:
-Algorithm Basics
-Algorithm Trading Styles
What Is an Algorithm?
In its simplest form, an algorithm is a list of steps needed to solve a problem. When referring to algorithmic trading, we refer to steps written in machine language so that a computer can understand what you want and execute trades on behalf of you and your goals. An algorithm spans multiple functions outside of trading but either way thealgorithmis used; it has a clear purpose to help compute large datasets in an efficient manner while abiding by key rules to help ensure the desired outcome. Algorithms accomplish this feat without having to worry about human biases or mental fatigue and high-level and high-frequency decision-making.
-Algorithm Trading Styles
The following list is not inclusive but does cover many commonly used strategies and styles in algorithmic trading:
Mean Reversion: Reverting to the mean takes the idea that an extended move away from a long-term average is likely short-term and duefor a reversion or retracement. Algorithms that quantify extended moves based on an oscillator will utilize the average price over a set time and use that level as a target. There many popular tools and calculations for quantifying an extension that is due to revert but risk management must also be included in the algorithm encasing new trend is developing.
Trend following: Trend following is the first, and still very popular technique of algorithmic-based momentum investing. Trends are easy to see, but can be hard to trade without the help of an algorithm. Because algorithms take over for the mind and the minds inherent biases, many of the fears that plague discretionary trend followers do not effect algorithms. A common fear when riding a strong trend is that it is about to turn or end, but that fear is often unfounded. One of the first widely followed trend following algorithms looked to buy a 20-day price breakout and hold that trade until a 20-day price low took them out of the trade. The traders who have and still do employ this algorithmic approach and other similar approaches are often amazed at how long the strongest trends extend that they would have likely exited had their algorithms not managed the trade and exit on their behalf.
News Trading: Another popular style of trading in the archaic world of discretionary trading that now belongs to the Quants is news trading. These strategies scan high important news events and calculate what type of print relative to prior news events and expectations would be needed to place a trade. As you can imagine, the efficiency of receiving the data and calculating whether a trade should be placed in entering that trade is of key focus. This form of algorithmic trading often gets the lion share of media’s attention.
Arbitrage:Arbitrage is a word that has multiple meetings and strategies built around the concept. Historically, you could have euros trading in London at a different price than in New York so that a trader could buy the lower and sell the higher until equilibrium had been established. Nowadays, arbitrage algorithm strategies are more geared to highly correlated assets whose underlying fundamental effects are very similar. When a wide spread in value between the highly correlated assets are recognized, the algorithm will either by the lower and or sell the higher until an equilibrium is met similar to the mean reversion strategy.
Trader sentiment: A cornerstone of DailyFX system desk is the algorithmic model based on trader sentiment through retail trader sentiment.
Other sentiment strategies will scan social networks like Twitter to identify popular trends that are developing and place trades accordingly. A high reading of anxiety or positivity towards a market can influence how these algorithms trade.
High-Frequency Trading and Scalping: For our purposes, will look at these as synonymous even though trading desks and hedge funds view them separately. True high-frequency trading attempts to beat out other traders to the thousand of a second and to do so some firms position their computers next door to an exchange to see in one millisecond faster than a competitor if something is rising by a penny.
Unless you’re looking to buy a house next to the New York Stock Exchange to compete with billion-dollar hedge funds, short-term trading or scalping is likely more up your alley. Even this term has evolved over time whereas traders use to look to make profits on the difference in the bid-ask spread but now has taken a wider meeting for very short-term traits.
Stealth and Iceberging: Similar to high-frequency trading, Iceburging often isn’t a concern as it is to large shops. The old saying that 90% of an iceberg is undersurface, with only 10% above paints the picture of this cost reduction strategy to break up larger orders over time into smaller orders to prevent the aforementioned high-frequency traders from recognizing and potentially front-running.
The stealth strategy is the high-frequency trading component that looks to uncover iceberg orders by tracking where it’s coming from and patterns of Iceburging relative to other large orders.
Summary:
The strategies above and the concepts mentioned hopefully kick the tires of your mind into thinking about what type of strategy you would be comfortable putting an algorithm behind. We often favor trend following strategies as their profit expectancy is high and costs are low, but there is a place for many of these strategies in the right environment.
