Sports Analytics Weekly by kubeia.io - 5/2024

The founding of the Betfair Exchange at the turn of the century changed the UK’s sports betting landscape forever.

In the realm of computer vision, YOLOv8 object tracking is revolutionizing the way we approach real-time tracking and analysis of moving objects. This article takes a close look at the fascinating world of YOLOv8 object tracking, offering a thorough understanding of its application in object tracking and counting. By reading this piece, you will gain insight into various practical implementations of object tracking and learn how these techniques can be effectively used in real-world scenarios. It also presents an in-depth exploration of the inference pipeline for object tracking and counting using YOLOv8.

crepes is a Python package that implements conformal classifiers, regressors, and predictive systems on top of any standard classifier and regressor, turning the original predictions into well-calibrated p-values and cumulative distribution functions, or prediction sets and intervals with coverage guarantees.The crepes package implements standard and Mondrian conformal classifiers as well as standard, normalized and Mondrian conformal regressors and predictive systems. While the package allows you to use your own functions to compute difficulty estimates, non-conformity scores and Mondrian categories, there is also a separate module, called crepes.extras, which provides some standard options for these.

There was a nice paper published a while ago by the excellent Taha Yasseri showing that soccer is becoming more predictable over time: from the early 90s to now, models trying to guess who would win a game had grown in accuracy. I got curious and asked myself: does this hold only for soccer, or is it a general phenomenon across different team sports? The result of this question was the paper: “Which sport is becoming more predictable? A cross-discipline analysis of predictability in team sports,” which just appeared on EPJ Data Science.

In recent years, excessive monetization of football and professionalism among the players have been argued to have affected the quality of the match in different ways. On the one hand, playing football has become a high-income profession and the players are highly motivated; on the other hand, stronger teams have higher incomes and therefore afford better players leading to an even stronger appearance in tournaments that can make the game more imbalanced and hence predictable. To quantify and document this observation, in this work, we take a minimalist network science approach to measure the predictability of football over 26 years in major European leagues. We show that over time, the games in major leagues have indeed become more predictable. We provide further support for this observation by showing that inequality between teams has increased and the home-field advantage has been vanishing ubiquitously. We do not include any direct analysis on the effects of monetization on football’s predictability or therefore, lack of excitement; however, we propose several hypotheses which could be tested in future analyses.

Let’s start with a hypothetical example: “What is the probability that a -7.5 favourite in a game with 44 total will cover -17.5?”There are several ways we could approach this. We could model the likelihood of each possible final score using some kind of probability distribution and/or regression. And that would work fine if done properly, but I want to think bigger. I want something that is flexible enough to answer not only this question but pretty much any other question you could ask about a game with a line of -7.5/44, no matter how complex. We could build a simulation, but that is much easier said than done and the results of any simulation will only be as good as the WORST assumption that goes into building it. We’re going to build the simplest possible type of model, an empirical one. Find a sufficient set of games that happened in the past and count how often the thing in question occurred in those.