I work in model validation and I am familiar with classical investment strategies, economic theories and pricing models of options and I am curious about algorithmic trading. What is it driven by? What theory does it believe in? For example does it rely on economic theories like efficient market hypothesis, growth and inflation models etc.? Said differently, before you write down a trading algorithm, where do you get the ideas? From the economists of the bank/fund? From the views of the traders? Or is it a simple technical analysis type of approach?

I am also interested in any references.

EDIT: By algo trading I mean algorithms for investment to generate alpha.

  • 8
    $\begingroup$ "Algorithmic trading" could be interpreted in two ways. The first way is interested in what you might call optimal execution, which seeks to address questions such as "how do I structure large buy/sell orders such that I obtain the best buy/sell price?" or "what is the expected impact on the asset's price from my transaction?". The second one is about investing, that is generating returns by automating trading decisions. $\endgroup$ Commented Jan 22, 2021 at 13:57
  • $\begingroup$ Thank you. I should maybe have been more clear, I am asking about the investing part $\endgroup$
    – sigma1988
    Commented Jan 22, 2021 at 14:37
  • $\begingroup$ It's hard to put an exact number by the vast majority of algorithmic trading today is either statistical arbitrage (core belief - mean reversion) or latency arbitrage which exploits the information asymmetry. It's rule based, and hence automated, and because it's automated - it's HFT. $\endgroup$ Commented Jan 22, 2021 at 15:52
  • $\begingroup$ Even circumscribing to the investing side, there is a wide variety of underpinnings, generally differentiated by the holding period length. For example, high frequency trading (HFT) can be based on exploiting features from the market structure using speed $-$ see e.g. latency trading. $\endgroup$ Commented Jan 22, 2021 at 15:54
  • 1
    $\begingroup$ On the other hand, for longer holding periods, algorithmic trading is usually based on identifying and exploiting statistical patterns in security prices, e.g. reversion to the mean as wrote by @SergeiRodionov. Another one of those patterns is momentum, which is inconsistent with classical financial theory, but for which many papers have been published to try to explain it from a financial/economic perspective. $\endgroup$ Commented Jan 22, 2021 at 15:57


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.