Sports Analytics Weekly by kubeia.io - 8/2024

BB has been an advantage play community since 2015 and in the decade before this we were involved with various betting syndicates. Through our experiences with the more professional teams we identified four dark arts that each of them worked on to maintain their profitability.

Some bettors are more proficient at one skillset than others, but without putting effort into each of these arts then it you may end up quitting – landing on that dumpster pile where too many failed professional bettors have ended up before.

The Dark Arts are:

Edge Identification

Getting On

Staking Optimally

Psychological Management

The unseen evolution steering MLS through adolescence is a function of technology, money and necessity.The league’s childhood was one of volatility and austerity. Its present, at age 29, is full of incongruity and contrasts. While Lionel Messi dazzles, and owners boast of a future among soccer’s elite, self-imposed roster restrictions stunt their collective growth. Arcane rules alienate fans. Spending limits handicap the quality of play. They promote parity within MLS, but impede competitiveness with foreign clubs — for continental titles, and for TV audiences.Off the screen, though, an elemental reason for the restrictions is disappearing.

Positional flexibility is all the rage in the NFL. Teams are looking for players who can do multiple things and are difficult to gameplan against. I once worked for an offensive coordinator who loved to say, “the more you can do, the more you can do.” It sounds trite, but players with the ability to perform different roles on the field can be used in a wider variety of situations and utilized in different ways.

This year at the StatsBomb Conference I learned that my review blog posts no longer need introducing. I was pleasantly surprised to hear from multiple participants in the conference’s research-paper competition that my annual blog posts had helped them enter this exciting field and shape their research ideas. When I started this series of blog posts, I had precisely that goal in mind. Needless to say, I hope this overview of the soccer analytics content that I liked the most in 2023 will inspire another generation of soccer analytics enthusiasts in the new year!

How can we measure something as fluid as a football match in charts and numbers?People have wrestled with that question for as long as we’ve kept records of the sport. As far back as the 1960s, Charles Reep was hand-drawing elaborate colour-coded charts to visualise England’s play. In the last few decades, the number of familiar football stats doubled when FIFA decided to count assists as well as goals, then exploded when new data companies made it possible to assemble a standard menu of shot counts, possession percentages, corner kicks, yellow cards and so on.

Socceraction is a Python package for objectively quantifying the impact of the individual actions performed by soccer players using event stream data. The general idea is to assign a value to each on-the-ball action based on the action's impact on the game outcome, while accounting for the context in which the action happened. The video below gives a quick two-minute introduction to action values.

Abstract:Huge amounts of money are invested every year by football clubs on transfers. For both growth and survival, it is crucial for recruiting departments to make smart choices when targeting players. Therefore, it is very important to identify the right parameters to monitor to predict market value. The following paper aims at determining the relevant features that successfully forecast future value for football players. Success is measured against their market value from TransferMarkt. To select prominent features, we use Lasso regressions and Random Forest algorithms. Some obvious variables are selected but we also observe some subtle dependencies between features and future market value. Finally, we rank the Golden Boy nominees using our forecasts and show our methodology can successfully compare football players based on their quality.

Modeling features as expected returns can be a useful way to develop trading strategies, but it requires some care.The main advantage is that it directly aligns with the objective of predicting and capitalising on future returns. This can make optimisation and implementation more intuitive. It also facilitates direct comparison between features and provides a common framework for incorporating new signals or reassessing existing ones.

This article is meant as an ad-hoc response to Ben Recht’s recent blog series on whether we need conformal prediction intervals. I have been thinking a lot about the use of conformal prediction myself and this seems like a good opportunity to share some thoughts and learnings from working on conformal prediction the past few years.

In the inaugural episode of The StatsBomb Football Podcast, join host Seth Partnow alongside StatsBomb CEO Ted Knutson and Head of Football Analysis Matt Edwards as they delve into the groundbreaking world of football analytics. Discover the current landscape of data utilization in the NFL and college football, drawing insightful comparisons with other sports.

What does a performance analyst actually do? (w. Carlon Carpenter, Head Video Analyst at Houston Dynamo) Jon Mackenzie is joined by Carlon Carpenter to discuss what a performance analyst does.How did he get his foot in the door at University of Virginia?What are the biggest differences between the analysis processes at the different levels?Are preconceived biases a thing in video/performance analysis?How much of the analysis is of what the coaching staff want to see vs. highlighting things they should be seeing?This episode of the Tifo Football Podcast is presented by Jon Mackenzie, produced by Mike Zimmermann.

In addition to allowing us to find any kind of relationship between two sequences without making any assumptions about their distributions, and being robust to outliers, Hoeffding's D also has a few other benefits: it is symmetric, so that hoeffd(X, Y) is the same as hoeffd(Y, X), it is always bounded (the result will never be less than -0.5 or larger than 1.0), and when one of the sequences is a constant, the Hoeffding's D goes to zero. This is not true of some other powerful measures of association,

I read my first Bill James book in 1984, took my first statistics class in 1985, and began graduate study in statistics the next year. Besides giving me the opportunity to study with the best applied statistician of the late 20th century (Don Rubin) and the best theoretical statistician of the early 21st (Xiao-Li Meng), going to graduate school at Harvard in 1986 gave me the opportunity to sit in a basement room one evening that October with about 20 other students, screaming at the TV, "Put Stapleton in!" Unfortunately, John McNamara didn't hear us, and the rest was history.