Sports analytics

Data analytics in sports is crucial to evaluate the performance of single players and the whole team. The literature proposes a number of tools for both offence and defence scenarios. Data coming from tracking location of players, in this respect, may be used to enrich the amount of useful information. In basketball, however, actions are interleaved with inactive periods. This paper describes a methodological approach to automatically identify active periods during a game and to classify them as offensive or defensive. The method is based on the application of thresholds to players kinematic parameters, whose values undergo a tuning strategy similar to Receiver Operating Characteristic curves, using a ground truth extracted from the video of the games.

In this paper, we explore some of the applications of computer vision to sports analytics. Sport analytics deals with understanding and discovering patterns from a corpus of sports data. Analysing such data provides important performance metrics for the players, for instance in soccer matches, that could be useful for estimating their fitness and strengths. Team level statistics can also be estimated from such analysis. This paper mainly focuses on some the challenges and opportunities presented by sport video analysis in computer vision. Specifically, we use our multi-camera setup as a framework to discuss some of the real-life challenges for machine learning algorithms.

Major League Baseball (MLB) has a storied history of using statistics to better understand and discuss the game of baseball, with an entire discipline of statistics dedicated to the craft, known as sabermetrics. At their core, all sabermetrics seek to quantify some aspect of the game, often a specific aspect of a player's skill set - such as a batter's ability to drive in runs (RBI) or a pitcher's ability to keep batters from reaching base (WHIP). While useful, such statistics are fundamentally limited by the fact that they are derived from an account of what happened on the field, not how it happened. As a first step towards alleviating this shortcoming, we present a novel, contrastive learning-based framework for describing player form in the MLB. We use form to refer to the way in which a player has impacted the course of play in their recent appearances. Concretely, a player's form is described by a 72-dimensional vector. By comparing clusters of players resulting from our form representations and those resulting from traditional sabermetrics, we demonstrate that our form representations contain information about how players impact the course of play, not present in traditional, publicly available statistics. We believe these embeddings could be utilized to predict both in-game and game-level events, such as the result of an at-bat or the winner of a game.

Nowadays, analytics is playing a key role in any field. Analytics is used in our day to day life. People don’t want to purchase even without analytics. Before few months IPL (Indian Premier League for Cricket) auction was taking placed. Even franchises are interested to pick a cricket player based on their past performances. For all this, they need accurate data and analytics with the comparative study to pick a particular player for their team. Here, we are comparing the last 10 years cricket sports data of IPL with interesting facts. We can perform much analytics but here we are focusing on the comparison between popular players, team, and fielders. Performance of top-10 players in each category like batting, bowling and fielding. Valuable player for the specific team.

The Pythagorean Win-Loss formula can be effectively used to estimate winning percentages for sporting events. This formula was initially developed by baseball statistician Bill James and later was extended by other researchers to sports such as football, basketball, and ice hockey. Although one can calculate actual winning percentages based on the outcomes of played games, that approach does not take into account the margin of victory. The key benefit of the Pythagorean formula is its utilization of actual average runs scored and actual average runs allowed. This article presents the application of the Pythagorean Win-Loss formula to two different types of limited-overs cricket formats, namely One Day International cricket (ODI) and Twenty20 cricket. The data for the application was used from the matches played by the top 10 International Cricket Council (ICC) members who participated in the 2019 ICC Cricket World Cup. For matches for which the second batting team won, runs scored w...

The aim of this paper is to study and develop Bayesian models for the analysis of volleyball match outcomes as recorded by the set-difference. Due to the peculiarity of the outcome variable (set-difference) which takes discrete values from $-3$ to $3$, we cannot consider standard models based on the usual Poisson or binomial assumptions used for other sports such as football/soccer. Hence, the first and foremost challenge was to build models appropriate for the set-differences of each volleyball match. Here we consider two major approaches: a) an ordered multinomial logistic regression model and b) a model based on a truncated version of the Skellam distribution. For the first model, we consider the set-difference as an ordinal response variable within the framework of multinomial logistic regression models. Concerning the second model, we adjust the Skellam distribution in order to account for the volleyball rules. We fit and compare both models with the same covariate structure as...

Bowling effectiveness is a key factor in winning cricket matches. The team captain should decide when to use the right bowler at the right moment so that the team can optimize the outcome of the game. In this study, we investigate the effectiveness of different types of bowlers at different stages of the game, based on the conceded percentage of runs from the innings total, for each over. Bowlers are generally categorized into three types: fast bowlers, medium-fast bowlers, and spinners. In this article, the authors divided the twenty over spell of a T20I match into four stages; namely, Stage 1: overs 1-6 (PowerPlay), Stage 2: overs 7-10, Stage 3: overs 11-15, and Stage 4: overs 16-20. To understand the broad spectrum of the behavior of game variables, a Quantile Regression methodology is used for statistical analysis. Following that, a Bayesian approach to Quantile Regression is undertaken, and it confirms the initial results.

In this thesis we propose three different models for the serve skill in modern volleyball: the ball is idealized as spherical in the simplest model, while it has
prolate ellipsoidal shape in the other two. The models are very adaptable,
in fact they simulate (jump) float serves, but also overhand serves in which
the ball has arbitrary spin direction. The most elaborate model takes into
account a viscoelastic deformation that the ball may undergo, due to its
compressibility, after the impact with the server’s hand.
The theory underlying the models does use results of fluid dynamics and
aerodynamics, which are treated in some detail. In particular, it is found
a computationally convenient representation for the lift force on a prolate ellipsoid in a rotational flow.
Once the models are translated into MATLAB code, they are tested by
numerical analysis: they are compared firstly among them, taking into account
trajectories and energy balances, secondly the paths of some videotaped serves, performed by professional athletes, are used to assess the validity of the models.
The results of these tests show promising compliance.

In cricket, all-rounders play an important role. A good all-rounder should be able to contribute to the team by both bat and ball as needed. However, these players still have their dominant role by which we categorize them as batting all-rounders or bowling all-rounders. Current practice is to do so by mostly subjective methods. In this study, the authors have explored different machine learning techniques to classify all-rounders into bowling all-rounders or batting all-rounders based on their observed performance statistics. In particular, logistic regression, linear discriminant function, quadratic discriminant function, naïve Bayes, support vector machine, and random forest classification methods were explored. Evaluation of the performance of the classification methods was done using the metrics accuracy and area under the ROC curve. While all the six methods performed well, logistic regression, linear discriminant function, quadratic discriminant function, and support vector m...

In this episode, Derek and Nathan speak with Andrew Stoeten, Toronto Blue Jays (Major League Baseball) columnist at The Athletic and co-host of the Birds All Day podcast (and formerly of Drunk Jays Fans (DJF) and Blue Jays Nation).

Andrew explains how he got into baseball writing and then the three delve into the tensions of having personal investments in fandom while working in a profession connected to sport and examine the difficulties of having political inclinations when "Republicans buy sneakers too." In the latter half of the episode, they turn to a discussion of how advanced analytics and fantasy sports can have a dehumanizing impact on the way players are viewed and then consider the future of baseball in the pandemic.

The purpose of this study is to examine players' physical and performance measures taken at the NFL Scouting Combine and compare these to their future performance in the NFL. From 2002-2016, three types of player data (N=5,506) were collected from secondary data sources. Results players earned on various NFL Scouting Combine drills and measurements (e.g., height, weight, 40-yard dash time, vertical jump, bench press repetitions, shuttle run time, and 3-cone drill time), the position players play on the field (e.g., quarterback, running back, wide receiver, tight end, offensive line, defensive line, linebacker, and defensive back), and if players received elite performance awards (e.g., Pro Bowl and All-Pro selections) in the future were collected. After analyzing the data, the results indicate that (1) NFL quarterbacks that received All-Pro and Pro Bowl awards tend to be taller, weigh more, run faster in the 40-yard dash, jump higher, complete more bench presses, and are slower for the shuttle run and 3-cone drill; (2) All-Pro and Pro Bowl NFL running backs tend to weight more, run the 40-yard dash faster, do not jump as high, complete more bench presses, and complete the shuttle run and 3-cone drills slower; (3) NFL wide receivers that were selected for the Pro Bowl or as All-Pros tend to be taller, weigh more, run the 40-yard dash faster, have a higher vertical jump, and run the 3-cone drill faster; (4) NFL tight ends that received All-Pro and Pro Bowl awards tend to be taller, weight more, run the 40-yard dash faster, jump higher, complete more bench presses, run the shuttle run slower, and complete the 3-cone drill faster; (5) All-Pro and Pro Bowl NFL offensive linemen tend to run the 40-yard dash faster, jump higher, are able to complete more bench presses, and run both the shuttle run and 3-cone drill faster; (6) NFL defensive linemen that were selected as All-Pro and Pro Bowl players tend to be taller, weight more, run the 40-yard dash faster, jump higher, complete more bench presses, and run the 3-cone drill faster; (7) NFL linebackers that were named to the All-Pro and Pro Bowl teams tend to be taller, weight more, run the 40-yard dash faster, jump higher, can complete more bench presses, and run both the shuttle run and 3-cone drill faster; (8) NFL wide receivers that were selected for the Pro Bowl or as an All-Pro tend to weigh more, run the 40-yard dash faster, jump higher, can complete fewer bench presses, and run the shuttle run slower. Certified strength and conditioning specialists for college and professional teams will be able to use these results to help train and set performance goals for American football athletes with whom they work and train.

Cricket is a worldwide popular game where different technologies are being used to help the match umpires to make decisions. Often due to the human perception, deciding whether a bowled delivery is a no-ball or legal ball which causes controversy. As only a single ball can change the fate of the game, so it is obvious to make accurate decision regarding no ball. In this paper, we have divided the bowling crease into two regions and applied image subtraction method on both regions to find the change in pixel values. Later, we have applied our proposed method on real world video frames. Since our proposed method to detect the overstep no ball is grounded on pixel by pixel image subtraction, it eliminates the inadequacy of human perception.

We study and develop Bayesian models for the analysis of volleyball match outcomes as recorded by the set-difference. Due to the peculiarity of the outcome variable (set-difference) which takes discrete values from −3 to 3⁠, we cannot consider standard models based on the usual Poisson or binomial assumptions used for other sports such as football/soccer. Hence, the first and foremost challenge was to build models appropriate for the set-difference of each volleyball match. Here we consider two major approaches: (a) an ordered multinomial logistic regression model and (b) a model based on a truncated version of the Skellam distribution. For the first model, we consider the set-difference as an ordinal response variable within the framework of multinomial logistic regression models. Concerning the second model, we adjust the Skellam distribution to account for the volleyball rules. We fit and compare both models with the same covariate structure as in Karlis & Ntzoufras (2003). Both ...

International Journal of VLSI design & Communication Systems (VLSICS) is a bi monthly open access peer-reviewed journal that publishes articles which contribute new results in all areas of VLSI Design & Communications. The goal of this journal is to bring together researchers and practitioners from academia and industry to focus on advanced VLSI Design & communication concepts and establishing new collaborations in these areas.

Authors are solicited to contribute to this journal by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the VLSI design & Communications.

The introduction of DRS and rapidly-degrading tires in 2011 boosted on-track overtaking levels in Formula 1 to unprecedented highs. Since then, overtaking has steadily decreased again, culminating in a 60-percent reduction in 2017. In this paper, using a Poisson model on individual-level overtaking data from 2011 to 2018, it was found that about half the decrease can be attributed to the cars, 20 to 30 percent to the reduction in field size and about 20 percent to more uniform race strategies.

When two people are on the phone, although they cannot observe the other person's facial expression and physiological state, it is possible to estimate the speaker's emotional state by voice roughly. In medical care, if the emotional state of a patient, especially a patient with an expression disorder, can be known, different care measures can be made according to the patient's mood to increase the amount of care. The system that capable for recognize the emotional states of human being from his speech is known as Speech emotion recognition system (SER). Deep learning is one of most technique that has been widely used in emotion recognition studies, in this paper we implement CNN model for Arabic speech emotion recognition. We propose ASERS-CNN model for Arabic Speech Emotion Recognition based on CNN model. We evaluated our model using Arabic speech dataset named Basic Arabic Expressive Speech corpus (BAES-DB). In addition of that we compare the accuracy between our previous ASERS-LSTM and new ASERS-CNN model proposed in this paper and we comes out that our new proposed mode is outperformed ASERS-LSTM model where it get 98.18% accuracy.

This paper aims to develop an interpretable machine learning model to predict plays (pass versus rush) in the National Football League that will be useful for players and coaches in real time. Using data from the 2013–2014 to 2016– 2017 NFL regular seasons, which included 1034 games and 130,344 pass/rush plays, we first develop and compare several machine learning models to determine the maximum possible prediction accuracy. The best performing model, a neural network, achieves a prediction accuracy of 75.3%, which is competitive with the state-of-the-art methods applied to other datasets. Then, we search over a family of simple decision tree models to identify one that captures 86% of the prediction accuracy of the neural network yet can be easily memorized and implemented in an actual game. We extend the analysis to building decision tree models tailored for each of the 32 NFL teams, obtaining accuracies ranging from 64.7% to 82.5%. Overall, our decision tree models can be a usefu...

Online streaming media services and growing internet access have made live television events interactive, social, and global. Football (soccer) broadcasts are a perfect example. As business models underpinning football media evolve, this thesis explores the implications of varying levels of fan engagement throughout matches. Specifically, data science techniques including supervised learning (multiple linear regression, ARIMA modeling) are applied to quantify the aggregate impact of different match events (goals, yellow cards, red cards) on global, real-time search volumes during individual matches. Findings include ~6.2%, 4.1%, and 12.5% bumps in scaled search volume levels in the immediate aftermath of goals, home yellows, and away reds, with individual observations ~5.6%, 8.3% lower when one side has a man down and when matches have goal differential greater than or equal to 3.

Brigham Young University's highly-visible football program stands at the nexus of The LDS Church and the beloved American cultural institution of football. This study examines the program in the context of the proliferation of " analytics, " or advanced statistics, that is revolutionizing decision-making in sports, business, and other areas of life. New analytic thinking has enabled the evaluation of the program's performance under different head coaches. Its overachievement for its aggregate talent level in the past may be credited to its best talent being concentrated at the two most important positions. Currently, the program is acquiring the best talent at the most important position but not at the second and third most important, though that will likely change with recent events. Helpful to the program's recruiting efforts would be membership in a top-tier conference. The program's outgoing staff has demonstrated conscientiousness of some analytic findings. The new staff has used new metrics and standard analytic practices, though receptiveness to analytics among its high-level members has not been uniform. This study has implications for how Church-affiliated institutions respond to changes in the intellectual atmosphere.

Abstract or Description
The standard plus/minus model for rating NBA players combines the offensive and defensive capabilities of each player into a single metric. While this is con- venient for the sake of summary, it makes it difficult to isolate the particular contributions that a player makes to either effort. Although adjusted plus/minus and other methods are proposed to address this, given a relatively large number of players against the number of events observed in one season, estimates are sub- jected to high variance. To correct for this, we construct a penalized regression model that identifies the specific offensive and defensive contributions of each player on each possession, and tune the model using L2-regularization method to optimize its predictive power. It overcomes the limitations of simple and adjusted plus/minus by incorporating offensive and defensive effects separately, and the shrinkage term controls for high variance of the estimates. Furthermore, our model captures net home court advantage on offense, and estimate players' contributions in offense and defense. Finally, we demonstrate application of our method by simulating unseen matches to correctly predict their outcomes.

During the National Basketball Association (NBA) playoffs, teams are required to frequently travel to different venues to play opponents in series of up to seven games. Despite playoff schedules allowing for some rest between games, it is still possible for teams to face circadian misalignment when playing. Thus, the current study serves as a replication and extension of previous research, which has indicated that there is an advantage for teams playing closer to their circadian peak and when they are traveling east. This study specifically investigates the effects of travel, as well as time of game on various performance indicators in professional basketball. We examined a series of box-score statistics (e.g., game outcomes, points scored, shooting percentages, rebounds, assists, steals, blocks, turnovers, and personal fouls) from a total of 499 postseason games played between the 2013–14 and 2018–19 NBA seasons. Findings from our study indicate that teams traveling eastward scored...

In sports, including Test cricket, athletes from years past serve as performance role models and set benchmarks for subsequent generations of players. Sports fans often wonder: are players of today as good as greats from the past? Alternatively, how do today’s athletes compare with greats from yesteryears? This paper attempts to answer that question for Test match cricket. We applied data mining to batting performance of eighty, now retired, Test Cricket Greats (TCG from hereon) from eight major Test cricket countries. Batting performance attributes included batting average, strike rate, numbers of fifties and hundreds scored, among others. Using k-Means cluster analysis, TCG performance records were classified into three clusters which was our Training Model. Two clusters were populated by established batsmen and the third cluster included bowlers, all-rounders with significant bowling, and some batsmen. The Learning Model was applied to predict classifications of thirty two Test C...

The purpose of this study is to examine players’ physical
and performance measures taken at the National Football
League (NFL) Scouting Combine and compare these with their
future performance in the NFL. From 2002 to 2016, 3 types of
player data (N = 5,506) were collected from secondary data
sources. Results of players earned on various NFL Scouting
Combine drills and measurements (e.g., height, body mass, 40-
yard dash time, vertical jump, bench press repetitions, shuttle
run time, and 3-cone drill time), the position players play on the
field (e.g., quarterback, running back, wide receiver, tight end,
offensive line, defensive line, linebacker, and defensive back),
and if players received elite performance awards (e.g., Pro
Bowl and All-Pro selections) in the future were collected. After
analyzing the data, the results indicate that (a) NFL
quarterbacks that received All-Pro and Pro Bowl awards tend
to be taller, weigh more, run faster in the 40-yard dash, jump
higher, complete more bench press reps, and are slower for
the shuttle run and 3-cone drill; (b) All-Pro and Pro Bowl NFL
running backs tend to weigh more, run the 40-yard dash faster,
do not jump as high, complete more bench press reps, and
complete the shuttle run and 3-cone drills slower; (c) NFL wide
receivers who were selected for the Pro Bowl or as All-Pros
tend to be taller, weigh more, run the 40-yard dash faster, have
a higher vertical jump, and run the 3-cone drill faster; (d) NFL
tight ends that received All-Pro and Pro Bowl awards tend to
be taller, weigh more, run the 40-yard dash faster, jump higher,
complete more bench press reps, run the shuttle run slower,
and complete the 3-cone drill faster; (e) All-Pro and Pro Bowl
NFL offensive linemen tend to run the 40-yard dash faster,
jump higher, are able to complete more bench press reps,
and run both the shuttle run and 3-cone drill faster; (f) NFL
defensive linemen who were selected as All-Pro and Pro Bowl
players tend to be taller, weigh more, run the 40-yard dash
faster, jump higher, complete more bench press reps, and
run the 3-cone drill faster; (g) NFL linebackers who were
named to the All-Pro and Pro Bowl teams tend to be taller,
weigh more, run the 40-yard dash faster, jump higher, can
complete more bench press reps, and run both the shuttle
run and 3-cone drill faster; and (h) NFL wide receivers that
were selected for the Pro Bowl or as an All-Pro tend to weigh
more, run the 40-yard dash faster, jump higher, can complete
fewer bench press reps, and run the shuttle run slower. Certified
strength and conditioning specialists for college and professional
teams will be able to use these results to help train
and set performance goals for American football athletes with
whom they work and train.

OBJECTIVES This study aimed to: 1) examine recent seasonal changes in performance indicators for different National Rugby League (NRL) playing positions; and 2) determine the accuracy of performance indicators to classify and discriminate positional groups in the NRL. DESIGN Retrospective, longitudinal analysis of individual performance metrics. METHODS 48 performance indicators (e.g. passes, tackles) from all NRL games during the 2015-2019 seasons were collated for each player´s match-related performance. The following analyses were conducted with all data: (i) one-way ANOVA to identify seasonal changes in performance indicators; (ii) principal component analysis (PCA) to group performance indicators into factors; (iii) two-step cluster analysis to classify playing positions using the identified factors; and (iv) discriminant analysis to discriminate the identified playing positions. RESULTS ANOVA showed significant differences in performance indicators across seasons (F=2.3-687.7; p=0-0.05; partial η2=0.00-0.075). PCA pooled all performance indicators and identified 14 factors that were included in the two-step cluster analysis (average silhouette=0.5) that identified six positional groups: forwards, 26.7%, adjustables, 17.2%, interchange, 23.2%, backs, 20.9%, interchange forwards, 5.5% and utility backs, 6.5%. Lastly, discriminant analysis revealed five discriminant functions that differentiated playing positions. CONCLUSIONS Results indicated that player's performance demands across different playing positions did significantly change over recent seasons (2015-2019). Cluster analysis yielded a high-level of accuracy relative to playing position, identifying six clusters that best discriminated positional groups. Unsupervised analytical approaches may provide sports scientists and coaches with meaningful tools to evaluate player performance and future positional suitability in RL.

Linked data have made significant progress over the last few years and many kinds of datasets are transformed into this format at a highly increasing rate, contributing to the openness, connectivity and re-use of web data. However, this progress is not the case for a popular sport like basketball, at least as far as the raw statistics is concerned. This kind of data contains valuable information that can be used by fans, teams and coaches, statisticians and other scientists. In this work, statistical data from Euroleague are transformed into linked data, thereby filling the relevant gap in the LOD Cloud, while ways of exploitation of them are presented, from fascinating applications for the fans, like the Euroleague Timeline, to cases of complex processing, analysis and visualization of data through software like R.