Algorithmic trading is a defined set of instructions based on timing, price, quantity, or any mathematical model. The goal of algorithmic trading is to provide a way of identifying trades based on the same methods a human uses. But in doing this, an algorithm removes both human error and emotional interference.
A quick example of an algorithmic instruction could be to monitor a market or stock for a particular trading indicator. For example, “wait for the relative strength index to reach 35, then buy. Of course, there would be other metrics to also satisfy to avoid ‘no trade zones’.
The instruction to then sell might be “wait for the relative strength index to reach 65, then sell”. This would be a very basic algorithm for buying and selling an asset.
This is a comprehensive list of types of algorithmic trading. All trading algorithms need to identify trades with a high probability of being profitable. Here are a few trading strategies that trading algorithms might look to take advantage of.
Trend trading algorithms are quite common because of their simplicity. The algorithm will track highs and lows in price movement and look for a trend in chart pattern either up or down.
The simplicity of these algorithms come from not needing to forecast or predict future price movement. The goal is to simply find trend lines of support and resistance without complicated calculations.
Trend trading algorithms may also use moving averages such as 50 and 200 day moving averages to identify trends forming.
In a similar way to trend trading, range trading will involve identifying a common support level and common resistance level. The strategy would be to buy off of support and sell off of resistance or vice versa.
In this instance the algorithm will look to identify common areas of price where the direction changes (a reversal). As with trend trading there is not advanced calculations for predicting future price.
Alternatively range trading may be used in the form of mean reversion. The idea is that highs and lows are temporary and that price always moves back to the mean price. Buying and selling is triggered based on the breaking of the defined price range.
Volume-weighted Average Price (VWAP)
The VWAP measures the average price of an asset for the day, this is based on volume and price. This strategy will look to break up large orders and release smaller portions based a volume profile. The strategy will attempt to execute as close to the VWAP as possible.
Time-weighted Average Price (TWAP)
The time-weighted average price strategy sends smaller orders to the market over time instead of all at once. The goal is to trade near the average price between start and end times, in order to limit market impact.
Relative Strength Index (RSI)
I mentioned this in the example earlier, it’s a rather simple idea. The relative strength index is a momentum indicator that shows how over-bought or over-sold an asset is. The indicator is an oscillator between 0 and 100. When the RSI reaches above 70 the asset is becoming over-bought. When the RSI drops below 30, it is becoming oversold. Price tends to correct its self by changing direction around these areas.
The implementation shortfall strategy is a way to save money on the cost of an order by trading off the real-time market. This means that the order is not executed right away, but instead is done at a later time when the opportunity cost is lower.