What are the most popular strategies used in high-frequency trading? Vikram Rangnekar Pair Trading - Trade two stocks which naturally track each other an example could be Coke and Pepsi, make money when they fall out of line on the idea that they will have to revert back to tracking each other. This is a common mean revision strategy used by hedge funds and might not exactly fit high-frequency trading however it still fall under algorithmic trading. Volume-Weighted Average Price - VWAP is used to execute large orders at a better average price. It is the ratio of the value traded to the total volume traded over a time period Time-Weighted Average Price - TWAP like VWAP is another sophisticated strategy for buying or selling large blocks of shares without affecting the price. Percentage of Volume - POV is used where the traders want to define the percentage, trading intervals and price when there is a need to trade in large blocks of stock without affecting price. Iceberg and Sniffer - Are algorithms used to detect and react to other traders trying to hide large block trades using the above algorithms. Flash Orders - Markets expose their order books in advance to algorithms subscribed to receive flash orders. This creates a two-tiered market for most passive investors where algorithms can front run them. A flash order received to sell a stock at a price allows algorithms to clear their own deal books of that stock at a higher price. A lot of HF algorithms and the minimum latency network infrastructure is to ensure that you can collect the liquidity-rebate that markets pay to ensure a highly liquid environment. When a lot of actors are rushing in to provide this liquidity you have to be the fastest and smartest to catch the rebate. While VWAP, TWAP, POV are technicals, they are also benchmarks that algorithms use while making their trading decisions. For example in theory if the price of a buy trade is lower than the VWAP, it is a good trade and its not a good trade if the price is higher than the VWAP. It is obviously way more complicated than this and today algo trading firms probably use vastly more complex derivatives of these mentioned strategies. Source: https://www.quora.com/What-are-the-most-popular-strategies-used-in-high-frequency-trading
As far as I know, there are three kind of strategies: Pair Trading (as mentioned in your post), based on relative price to a correlated stock. Trend following, based on relative price to the previous price. Mean reversion, based on relative price to the (recent) mean. And derivatives and combinations of these three strategies within various frequency of triggers, from near real time response to markets, e.g. nano seconds, to batch mode, from milliseconds to hours. This batch time is depended on the time required to digest new data, for simple static algorithms, it will be in nano seconds or milliseconds for complex adaptive algorithms, that is the algorithms change as new data is processed, this sort of process may be completed from seconds to hours. Did I miss anything? I will be very happy to understand if there are other kind of strategies.