Research
Betwise and Betwise members produce research, publications and software for use by other members.
Past research papers, presentations, documentation and software for a few of our favourite Betwise projects are included on this page. There is a description with further background and other useful links for each project, and you can download the PDF publications relating to each by clicking on the thumbnail image. Links to further noteworthy Betwise projects and software without PDF publications are listed at the bottom of this page.
 |
Feasibility of producing a profitable data driven model for betting on horseracingThis paper, written as a Masters thesis by Betwise member Mikko Juola (and reproduced here with his permission) uses Betwise’s Smartform database and Betfair prices to research feasibility for building a profitable, data driven betting model. It’s a great paper with useful background information that will benefit any reader thinking of creating predictive models for sports betting. More specifically, since the data used for the models is provided by Betwise, it gives pointers for the creation of useful predictive features that subscribers can reproduce in their own work. See also the Betwise daily oddsline which presents the output of our current machine learning model for probabilities and odds available on horseracing on a daily basis. |
 |
R is for Racing (London R user group presentation)This presentation, first delivered at the London R user group, provides background to using the R language for the data analysis and modelling of horseracing data using Betwise’s Smartform database. Wrriten with Professor Jay Emerson, it’s a teaser for a new book we have been working on for some time, R is for Racing. R is for Racing uses an extract from the Betwise data and R scripts to provide many ideas for exploratory data analysis as well as original software to implement models that turfistes and statisticians alike will find of immediate utility and can be applied on a daily basis. In the final stages, the completed book will be released in Summer 2024. |
 |
The BetfaiR package for RA Betwise software project with Bryan W. Lewis, we implemented access to the Betfair Sports Exchange API in R, with lots of R functions for visualising and manipulating betting market data. Bryan presented the work at the annual R finance conference in Chicago in 2011 – you can find the full slides from that presentation here. The guide to the package on this page also provides further material and background to automating betting activity with the Betfair API. Whilst the guide is useful for working with the data and defining an automated betting framework, the API implementation itself is now out of date, so the R package that we recommend to use for the current Betfair API (written by Phill Clarke, another Betwise member), is the excellent abettor, Phill has also written a series of Betwise blog posts that show how to use Smartform with R as a programming interface to produce derived variables and visualisations... |
There are lots more software packages and useful scripts available from Betwise and its members, too many for one page, but we would also give mention here to the following resources and example utilities available:
The Betwise Blog: Lots of example software as well as discourse on analytics for sports betting, especially horseracing. Too many threads to mention, but in line with the comments about creation of machine learning models on this page, we cite this post from Steve Tilley about how to use Smartfrom insights for one of the most important steps in the machine learning process, feature selection.
The Betwise Smartform Github which includes Betwise and member contributed code for working with Smartform. In particular, see the Excel VBA code for the race card tool from Henry Young and written up in a series of Betwise blog posts. This is a great utility for those members who prefer to visualise and work interactively with daily data.
Betwise Q&A: Many pointers to ffurther research material, troubleshooting and specific answers from Betwise members to commonly sought analytical solutions. See one of our favourite threads from Patrick Gillen about creation of speed ratings using Smartform including prototype software in R.