Skip to content
All library documents

Choosing a Starting Point: Execution Algorithms or Trend Following

Article Quant Q&A · Author: k9ty
Read the original

Summary

The document distinguishes two meanings often given to algorithmic trading. Execution algorithms implement an existing trading view while seeking to manage execution costs or market impact. VWAP and TWAP are named as common benchmarks, with iceberg, sniper, and guerrilla approaches as further examples. Almgren and Chriss’s optimal execution framework is suggested as a foundation for studying this area.

For research focused on generating trading signals, the answers instead suggest simple moving average trend following and testing whether such strategies behave differently across markets. The discussion offers no performance evidence or specific implementation, and cautions that a short solo project may be difficult. It also notes that results depend on market, data, and infrastructure, so these suggestions are starting points rather than a recipe for a profitable system.

Key ideas

  • Execution algorithms seek to carry out an existing trading view while managing market impact.
  • VWAP and TWAP are basic benchmarks for scheduling trades against volume or time.
  • Optimal execution research provides a foundation for studying execution algorithms.
  • Moving average trend following is a simple starting point for research into signal generation.
  • Testing a strategy across markets can reveal where its behavior changes.

Tags

ExecutionTrend followingTechnical indicatorsMarket microstructure
Full text
# Which algorithm should I look into to kick off my research in algorithmic trading?


# Which algorithm should I look into to kick off my research in algorithmic trading?












I have recently undertaken a research into automated algorithmic trading algorithms.

The aim of the research is to focus on studying algorithmic trading and trying to improve a basic implementation of an algorithm that exists.

The algorithms usually analyze past market data, social media events and other factors in near real time to find patterns and trends using an auto-regressive learning model with gradient descent and thus they attempt to predict market movements.

I am currently reading through the book "Algorithmic trading and DMA by Barry Johnson" as this was suggested in one of the forums that I encountered.

I realize that there are various models that have been implemented. Given the time frame for the project, which is 8 months, I would like to know if there are any good suggestions on starting points for this project.

For example, any books, papers that you can recommend.

More specifically, it would be helpful if you could suggest a specific algorithm that I can look at to start with considering I am new to the field of algorithmic trading in particular.

Since, the question appears vague, I have added a description of the project below http://algorithmic-trading-research.tumblr.com/post/12885213702/project-description

## Answer by Shane (score 6)

https://quant.stackexchange.com/a/2390

The term "algorithmic trading", as its most often used, refers to strategies that are used to optimally execute an existing view (this usually means trying to reduce market impact). The most well-known benchmark strategy is VWAP, which targets the "volume weighted average price" (similar to TWAP, which instead tries to evenly space trades over time). Other examples include things like Iceberg, Sniper, and Guerrilla.

This is a very big area of study. The book you cite is a good source; I also really like "Optimal Trading Strategies". The classic paper on the subject is "Optimal Execution of Portfolio Transactions" (Almgren and Chriss 2000).

## Answer by SRKX (score 4)

https://quant.stackexchange.com/a/2388

Your question is very vague, which makes it almost off-topic.

I would suggest you to look into trend-following algorithms, as the most basic ones are very easy to understand; look for trading strategies involving moving averages.

However, if you want to improve an algorithm within 8 months, alone, I think it will be pretty difficult.

What is that for?

My educated guess: a master thesis.

If that's the case I'd suggest you to look to apply the trend-following strategies on different markets and see if they work everywhere (they don't) and explain why.

Algorithmic trading is not easy, it requires a lot of work, a good infrastructure and depending on the data required, a large amount of money.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.