2011•Bristol Research (University of Bristol)Open access

Studies of Interaction Between Human Traders and Algorithmic Trading Systems

Marco De Luca, CS Szostek, John Cartlidge, Dave Cliff

Open full text 16 citations

Abstract

In recent years there has been a very significant increase in the percentage of trades in the global financial markets that are initiated and executed by automated “robot” algorithmic trading software systems, autonomously performing trading roles that a decade or more ago would have been performed by human traders. The anonymity of many current electronic trading systems, operated by major exchanges and multilateral trading facilities, mean that an individual trader, whether human or robot, never knows2 if the counterparty to a particular trade is a human or not. There are commonly-quoted estimates that the proportion of robot-executed trades is approaching 30%-70% on major European and US equity exchanges. In foreign-exchange markets, where there are no central exchanges, the proportion of spot (immediate-execution) transactions that are executed by robots is widely believed to be even higher. From this, it is clear that the current global financial markets involve a very significant degree of interaction between human and robot traders. The interactions between human traders in electronic markets has long been studied in the field known as Experimental Economics, and more recently the interactions between software-agent traders in electronic markets has been the topic of various abstract research studies in so-called Agent-based Computational Economics (ACE). These two research fields are largely distinct: the first studies markets populated entirely by human traders; the second studies markets populated entirely by algorithmic software-agent traders. There is a surprising lack of studies of the interactions between human traders and robot traders. That is, there is very little scientific literature that explores heterogeneous markets, populated by both humans and robots. In this document we review the very small amount of published literature that does describe scientific studies of interactions between human and robot traders under experimental conditions. We contend that the relative lack of such studies is a serious omission from the literature. We propose that De Luca’s (2010) Open Exchange (OpEx) open-source design for studying human-robot interactions in electronic marketplaces should be used as a free de facto standard for future work in this area. We illustrate the use of OpEx by summarising recently published peer-reviewed accounts of early experiments with OpEx, and then present a detailed description and analysis of results from some new experiments, conducted specifically for this review document, where we relax some of the artificial experimental constraints that have been used in earlier work. Experiments with the OpEx system indicate that the previously reported outperformance of the algorithmic trading systems over humans may well be related to the artificial nature of the experiment design that was employed in the earlier research: a design essentially unchanged since the first experimental economics results were published in the early 1960’s. When the flow of orders in the market was trickled in gradually (rather than all orders being released simultaneously, which was an artificial constraint in the designs of the earlier experiments) the performance of the software agents was no longer so clearly superior. Furthermore, when the software agents were slowed to operate on the same sort of timescales that human traders act on, the data we have thus far indicates that the market dynamics alter, but further experiments would be required to establish the significance of this with appropriate levels of certainty.

Open-access reader

About this research paper

What this paper is about

In recent years there has been a very significant increase in the percentage of trades in the global financial markets that are initiated and executed by automated “robot” algorithmic trading software systems, autonomously performing trading roles that a decade or more ago would have been performed by human traders. The anonymity of many current electronic trading systems, operated by major exchanges and multilateral trading facilities, mean that an individual trader, whether human or robot, never knows2 if the counterparty to a particular trade is a human or not. There are commonly-quoted estimates that the proportion of robot-executed trades is approaching 30%-70% on major European and US equity exchanges. In foreign-exchange markets, where there are no central exchanges, the proportion of spot (immediate-execution) transactions that are executed by robots is widely believed to be even higher. From this, it is clear that the current global financial markets involve a very significant degree of interaction between human and robot traders. The interactions between human traders in electronic markets has long been studied in the field known as Experimental Economics, and more recently the interactions between software-agent traders in electronic markets has been the topic of various abstract research studies in so-called Agent-based Computational Economics (ACE). These two research fields are largely distinct: the first studies markets populated entirely by human traders; the second studies markets populated entirely by algorithmic software-agent traders. There is a surprising lack of studies of the interactions between human traders and robot traders. That is, there is very little scientific literature that explores heterogeneous markets, populated by both humans and robots. In this document we review the very small amount of published literature that does describe scientific studies of interactions between human and robot traders under experimental conditions. We contend that the relative lack of such studies is a serious omission from the literature. We propose that De Luca’s (2010) Open Exchange (OpEx) open-source design for studying human-robot interactions in electronic marketplaces should be used as a free de facto standard for future work in this area. We illustrate the use of OpEx by summarising recently published peer-reviewed accounts of early experiments with OpEx, and then present a detailed description and analysis of results from some new experiments, conducted specifically for this review document, where we relax some of the artificial experimental constraints that have been used in earlier work. Experiments with the OpEx system indicate that the previously reported outperformance of the algorithmic trading systems over humans may well be related to the artificial nature of the experiment design that was employed in the earlier research: a design essentially unchanged since the first experimental economics results were published in the early 1960’s. When the flow of orders in the market was trickled in gradually (rather than all orders being released simultaneously, which was an artificial constraint in the designs of the earlier experiments) the performance of the software agents was no longer so clearly superior. Furthermore, when the software agents were slowed to operate on the same sort of timescales that human traders act on, the data we have thus far indicates that the market dynamics alter, but further experiments would be required to establish the significance of this with appropriate levels of certainty.

Why it matters

OpenAlex reports 16 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

In recent years there has been a very significant increase in the percentage of trades in the global financial markets that are initiated and executed by automated “robot” algorithmic trading software systems, autonomously performing trading roles that a decade or more ago would have been performed by human traders. The anonymity of many current electronic trading systems, operated by major exchanges and multilateral trading facilities, mean that an individual trader, whether human or robot, never knows2 if the counterparty to a particular trade is a human or not. There are commonly-quoted estimates that the proportion of robot-executed trades is approaching 30%-70% on major European and US equity exchanges. In foreign-exchange markets, where there are no central exchanges, the proportion of spot (immediate-execution) transactions that are executed by robots is widely believed to be even higher. From this, it is clear that the current global financial markets involve a very significant degree of interaction between human and robot traders. The interactions between human traders in electronic markets has long been studied in the field known as Experimental Economics, and more recently the interactions between software-agent traders in electronic markets has been the topic of various abstract research studies in so-called Agent-based Computational Economics (ACE). These two research fields are largely distinct: the first studies markets populated entirely by human traders; the second studies markets populated entirely by algorithmic software-agent traders. There is a surprising lack of studies of the interactions between human traders and robot traders. That is, there is very little scientific literature that explores heterogeneous markets, populated by both humans and robots. In this document we review the very small amount of published literature that does describe scientific studies of interactions between human and robot traders under experimental conditions. We contend that the relative lack of such studies is a serious omission from the literature. We propose that De Luca’s (2010) Open Exchange (OpEx) open-source design for studying human-robot interactions in electronic marketplaces should be used as a free de facto standard for future work in this area. We illustrate the use of OpEx by summarising recently published peer-reviewed accounts of early experiments with OpEx, and then present a detailed description and analysis of results from some new experiments, conducted specifically for this review document, where we relax some of the artificial experimental constraints that have been used in earlier work. Experiments with the OpEx system indicate that the previously reported outperformance of the algorithmic trading systems over humans may well be related to the artificial nature of the experiment design that was employed in the earlier research: a design essentially unchanged since the first experimental economics results were published in the early 1960’s. When the flow of orders in the market was trickled in gradually (rather than all orders being released simultaneously, which was an artificial constraint in the designs of the earlier experiments) the performance of the software agents was no longer so clearly superior. Furthermore, when the software agents were slowed to operate on the same sort of timescales that human traders act on, the data we have thus far indicates that the market dynamics alter, but further experiments would be required to establish the significance of this with appropriate levels of certainty.

Key concepts: Computer science, Business, Economics

Related papers

Back to paper searchBrowse research topicsOriginal source
Studies of Interaction Between Human Traders and Algorithmic Trading Systems — Research Paper | ScholarLens