An assumption with a solid  financial foundation that may yield long-term profits. Algorithm hypotheses require a thorough analysis of historical data.

The practice of using a computer system to carry out fully automated trades according to pre-programmed algorithms.

An investor or trader that uses an algorithmic trading system.

An automated system with data collection processes and data queries. The system uses computer algorithms to make trading decisions and reports, to manage real-time financial portfolios without any human intervention.

A set of definitions and protocols that allow communication between two software applications.

A strategy to take advantage of temporary price differences for the same asset in two different markets. This trading strategy makes profits without much risk.

The use of past data to evaluate an algorithm’s performance.

A trading strategy that takes advantage of market inefficiencies from corporate events like mergers and acquisitions, corporate restructures, share buybacks, extraordinary dividends, etc. to trade in the short term.

The use of future data from the present to a certain point in the future to evaluate an algorithm’s performance.

Investors anticipate the action of exchange-traded funds (ETFs) according to the public prospectus and simulate part of the action prior to the fund’s rebalancing.

Unseen data from the market used for testing.

A strategy that sets up a price grid around a predefined value to make profits from market fluctuations independent of market trends.

An algorithmic trading strategy with an extremely large number of transactions at high speed. It opens and closes positions in a very short period of time.

A discrepancy between paper profits and actual profits. Paper profit is a theoretical profit under the assumption all orders are executed at the desired price and volume without any fees.

The past historical data used to train and optimize trading algorithms.

A signal to determine the ticker symbol and the appropriate time to trade. Indicators are divided into technical indicators such as MA, and RSI, and base indicators such as PE, and ROE. Depending on the algorithms, an indicator can be custom defined based on its intended use.

Determines the optimal capital per trade to maximize the long-term performance of a trading algorithm.

A group of strategies when investors open both long and short positions to minimize the effects of market risk on portfolio profitability.

A strategy of simultaneously placing orders at the best bid price and the best ask price. This strategy makes profits from the bid-ask spread.

The maximum loss from the peak to the bottom for an asset or portfolio.

A strategy to buy stocks when the price is lower than their intrinsic or average value; and vice versa sell when the price is higher.

Investors buy rising stocks in terms of price (or short-sell falling stocks) with the reasoning that stocks will follow their momentum trends in price.

A process of finding the values of parameters for a trading algorithm to produce the best performance in the target market.

The past historical data used to validate trading algorithms after the training phase. It is to assess the profitability of an algorithm in the future. In-sample and out-of-sample historical data do not have any overlap.

The phenomenon that the post-optimization algorithm gives good results on in-sample historical data but performs poorly on unseen out-of-sample data.

A strategy that finds two stocks with high correlation to match a long position with a short position. When their correlation deviates above the long-term average, investors match a long position on the underperforming stock, and a short position on the outperforming stock, expecting the deviation to be temporary. When the price correlation converges to the average, the investor closes the position to realize the profit.

A stage in testing future data. It uses real-time data in a simulated trading environment to evaluate an algorithm’s performance.

A predefined arithmetic value like price, volume, or parametric constants of mathematical models.

Breaks down the order with volume based on the percentage of volume according to the market liquidity. As the market trading volume increases, it will trade more stocks and vice versa.

A special strategy that focuses on an ultra-short time frame to open and close positions in order to make very small profits.

A system that uses computers combined with human intervention to optimize the decision-making process.

The rate of expected return minus the risk-free rate, divided by the standard deviation of the rate of return. It measures the expected performance per unit of risk.

Also known as a factor-based strategy, a strategy to build portfolios according to rule-based processes. It uses business factors like liquidity, value, and quality as criteria for making trading decisions.

A group of market-neutral strategies developed from the pair trading strategy. It uses statistical mathematical models with computer systems to identify trading opportunities. They are mostly from unusual relative price changes from one stock compared to another.

A set of statements and trading logic to execute trading decisions, including but not limited to buy or sell orders and price, volume, and order types. Within the scope of our book, “algorithms” without further explanation can be understood as “trading algorithms”.

A rule-based system used to make trading decisions based on indicators and parameters.

 The equally weighted average price of all orders made during the execution period.

Breaks down the order into equal parts and consistently places orders, separated by an equal amount of time.

The weighted average price by the volume of all orders made during the execution period.

Breaks down and places orders at different times and different volumes based on historical data of trading volume.