No Cover Image

Journal article 38 views 6 downloads

Quantile regression analysis of in-play betting in a large online gambling dataset

Sebastian Whiteford, Alice Hoon Orcid Logo, Richard James, Richard Tunney Orcid Logo, Simon Dymond Orcid Logo

Computers in Human Behavior Reports, Volume: 6, Start page: 100194

Swansea University Authors: Sebastian Whiteford, Alice Hoon Orcid Logo, Simon Dymond Orcid Logo

  • 59770.pdf

    PDF | Version of Record

    © 2022 The Authors. This is an open access article under the CC BY license

    Download (2.01MB)

Abstract

In-play betting involves making multiple bets during a sporting event and is an increasingly popular form of gambling. Behavioural analysis of large datasets of in-play betting may aid in the prediction of at-risk patterns of gambling. However, datasets may contain significant skew and outliers nece...

Full description

Published in: Computers in Human Behavior Reports
ISSN: 2451-9588
Published: Elsevier BV 2022
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa59770
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: In-play betting involves making multiple bets during a sporting event and is an increasingly popular form of gambling. Behavioural analysis of large datasets of in-play betting may aid in the prediction of at-risk patterns of gambling. However, datasets may contain significant skew and outliers necessitating analytical approaches capable of examining behaviour across the spectrum of involvement with in-play betting. Here, we employ quantile regression analyses to investigate the relationships between in-play betting behaviours of frequency and duration of play, bets per day, net/percentage change, average stake, and average/percentage change across groups of users differing by betting involvement. The dataset consisted of 24,781 in-play sports bettors enrolled with an internet sports betting provider in February 2005. We examined trends in normally-involved and heavily-involved in-play bettor groups at the .1, .3, .5, .7 and .9 quantiles. The relationship between the total number of in-play bets and the remaining in-play betting measures was dependent on degree of involvement. The only variable to differ from this analytic path was the standard deviation in the daily average stake for most-involved bettors. The direction of some relationships, such as the frequency of play and bets per betting day, were reversed for most-involved bettors. Crucially, this highlights the importance of determining how these relationships vary across the spectrum of involvement with in-play betting. In conclusion, quantile regression provides a comprehensive account of the relationship between in-play betting behaviours capable of quantifying changes in magnitude and direction that vary by involvement.
Keywords: In-Play; Live-action; Gambling; Quantile regression; Internet betting
College: College of Human and Health Sciences
Funders: International Center for Responsible Gaming
Start Page: 100194