Sports Analytics Weekly by kubeia.io - 4/2024

The DINO model is to toe the line between fantasy sports operator (legal in most places) and sports betting operator (illegal in many places, strictly regulated in others) as closely as your lawyers will allow. This is NOT a new move in the betting operator playbook – see historical horse racing, bingo slot machines and other shenanigans so barely-legal they would make the porn industry blush.

4500 words on the MLS Superdraft, which will see at most 10 players earn 1000 minutes in the league? The same week that teams report to preseason, a month after the draft?Call me crazy, but to me the draft is still interesting and exciting. In one sense, it’s a bizarre little corner of the sport copied directly from bigger leagues like the NBA or NFL. You have the whole production, breathlessly reporting each pick. There are interviews with selected players. The entire production team throws its weight behind the process.

In this sense of exploration and play, I have a plan for how the newsletter will be scheduled. Every month I will produce a new study on a topic in soccer analytics. The first study here in January 2024 will be on substitute effects. Following the Study will be the Sandbox. Every study produces table after table with new statistics for players and teams. I will look through the material produced for the study to find interesting stories, and I will take your requests to look at players, teams, and topics you want to know more about.

Every year the top high school basketball recruits get hyped up. How often do they pan out?

Machine learning is an amazing tool for many tasks. OpenCV is a great library for manipulating images. It would be great if we can put them together.In this 7-part crash course, you will learn from examples how to make use of machine learning and the image processing API from OpenCV to accomplish some goals. This mini-course is intended for practitioners who are already comfortable with programming in Python, know the basic concept of machine learning, and have some background in manipulating images. Let’s get started.

Moving object detection is used extensively for applications ranging from security surveillance to traffic monitoring. It is a crucial challenge in the ever-evolving field of computer vision. The open-source OpenCV library, known for its comprehensive set of tools for computer vision, provides robust solutions to the detection of moving objects. In this article, we examine a combination of Contour Detection and Background Subtraction that can be used to detect moving objects using OpenCV.

Value at Risk (VaR) serves as a crucial tool in the financial landscape. This statistical measure quantifies potential losses in portfolios over a specified time horizon, offering a tangible understanding of risk with a defined level of confidence. And this comprehensive guide not only provides an introduction to value at risk but a lot more that will help you dive into it.From optimising portfolios to regulatory compliance, VaR finds widespread application. However, the journey with VaR is not without challenges, including assumptions and oversimplified views

We are pleased to announce the release of scikit-learn 1.4! Many bug fixes and improvements were added, as well as some new key features. We detail below a few of the major features of this release.

The launch of fbref.com has changed public soccer analytics, and we sat down to interview Sean Forman, the president of Sports Reference, about the website, how it came to be, what he learned about soccer analytics from building it, and about their new Stathead feature (stathead.com).

The Continuous Ranked Probability Score, known as CRPS, is a score to measure how a proposed distribution approximates the data, without knowledge about the true distributions of the data.

Extreme Value Theory is the branch of probability theory and statistics that is concerned

with extremes of sequences of random variables and stochastic processes. The subject started in the

1920s with a seminal paper by Fisher and Tippett and an almost-parallel contribution by Fréchet, but

was first put on a firm mathematical ground by Gnedenko (1943). The original theory was for

independent and identically distributed (IID) random variables but it was quickly extended to other

cases such as independent non-identically distributed random variables, stationary (dependent)

sequences and continuous-parameter stationary processes. Beginning in the 1950s but more

prominently in the 1970s, the theory started to be extended to multivariate random variables and more

recently to general stochastic processes including spatial and spatio-temporal processes. Statistical

methods for extremes were originally developed in a classic but now outdated book by Gumbel (1958),

but were developed much more extensively in the 1980s with the increased availability of computers

and automated computational algorithms (e.g. fitting the Generalized Extreme Value distribution by

maximum likelihood). These days, the theory is widely applied in many fields but three of the main ones

are reliability and strength of materials, mathematical finance and insurance, and environmental

(especially, climate) extremes. A side interest of mine is applications to records in athletic events.